diff --git a/R/acquisition.R b/R/acquisition.R
index 5235ab43..c25c5952 100644
--- a/R/acquisition.R
+++ b/R/acquisition.R
@@ -47,7 +47,7 @@
 #' @export
 prob_improve <- function(trade_off = 0, eps = .Machine$double.eps) {
   if (!is.numeric(trade_off) & !is_function(trade_off)) {
-    rlang::abort("`trade_off` should be a number or a function.")
+    cli::cli_abort("{.arg trade_off} should be a number or a function.")
   }
 
   lab <- "the probability of improvement"
@@ -55,7 +55,7 @@ prob_improve <- function(trade_off = 0, eps = .Machine$double.eps) {
   if (rlang::is_function(trade_off)) {
     farg <- names(formals(trade_off))
     if (length(farg) == 0) {
-      rlang::abort("The `trade_off` function should have at least one argument.")
+      cli::cli_abort("The {.fn trade_off} function should have at least one argument.")
     }
     lab <- paste(lab, "with variable trade-off values.")
   }
@@ -107,7 +107,7 @@ predict.prob_improve <-
 #' @rdname prob_improve
 exp_improve <- function(trade_off = 0, eps = .Machine$double.eps) {
   if (!is.numeric(trade_off) & !is_function(trade_off)) {
-    rlang::abort("`trade_off` should be a number or a function.")
+    cli::cli_abort("{.arg trade_off} should be a number or a function.")
   }
 
   lab <- "the expected improvement"
@@ -115,7 +115,7 @@ exp_improve <- function(trade_off = 0, eps = .Machine$double.eps) {
   if (rlang::is_function(trade_off)) {
     farg <- names(formals(trade_off))
     if (length(farg) == 0) {
-      rlang::abort("The `trade_off` function should have at least one argument.")
+      cli::cli_abort("The {.fn trade_off} function should have at least one argument.")
     }
     lab <- paste(lab, "with variable trade-off values.")
   }
@@ -161,13 +161,13 @@ predict.exp_improve <- function(object, new_data, maximize, iter, best, ...) {
 #' @rdname prob_improve
 conf_bound <- function(kappa = 0.1) {
   if (!is.numeric(kappa) & !is_function(kappa)) {
-    rlang::abort("`kappa` should be a number or a function.")
+    cli::cli_abort("{.arg kappa} should be a number or a function.")
   }
   lab <- "the confidence bound"
   if (rlang::is_function(kappa)) {
     farg <- names(formals(kappa))
     if (length(farg) == 0) {
-      rlang::abort("The `trade_off` function should have at least one argument.")
+      cli::cli_abort("The {.fn trade_off} function should have at least one argument.")
     }
     lab <- paste(lab, "with variable kappa values.")
   }
diff --git a/R/augment.R b/R/augment.R
index 2ca56fc0..d0ab84d5 100644
--- a/R/augment.R
+++ b/R/augment.R
@@ -49,7 +49,7 @@ augment.tune_results <- function(x, ..., parameters = NULL) {
     parameters <- select_best(x, metric = obj_fun, eval_time = obj_eval_time)
   } else {
     if (!is.data.frame(parameters) || nrow(parameters) > 1) {
-      rlang::abort("'parameters' should be a single row data frame")
+      cli::cli_abort("{.arg parameters} should be a single row data frame.")
     }
   }
 
@@ -86,11 +86,9 @@ merge_pred <- function(dat, pred, y) {
   pred <- pred[order(pred$.row), ]
   pred <- pred[, c(".row", pred_cols)]
   if (nrow(pred) != nrow(dat)) {
-    rlang::warn(
-      paste(
-        "The original data had", nrow(dat), "rows but there were",
-        nrow(pred), "hold-out predictions."
-      )
+    cli::cli_warn(
+      "The original data had {nrow(dat)} rows but there were {nrow(pred)}
+       hold-out predictions."
     )
   }
 
diff --git a/R/case_weights.R b/R/case_weights.R
index b53309f6..a17147f2 100644
--- a/R/case_weights.R
+++ b/R/case_weights.R
@@ -26,15 +26,17 @@
 #' @export
 .use_case_weights_with_yardstick.default <- function(x) {
   message <- c(
-    paste0("Unknown case weights object with class <", class(x)[[1]], ">. "),
+    "An object with {.cls hardhat_case_weights} was expected, not
+    {.obj_type_friendly {x}}.",
     i = paste0(
-      "Define a `.use_case_weights_with_yardstick()` method for this type to ",
-      "declare whether or not these case weights should be passed on to yardstick."
+      "Define a {.fn .use_case_weights_with_yardstick} method for this type to ",
+      "declare whether or not these case weights should be passed on to
+       {.pkg yardstick}."
     ),
-    i = "See `?.use_case_weights_with_yardstick` for more information."
+    i = "See {.help .use_case_weights_with_yardstick} for more information."
   )
 
-  rlang::abort(message)
+  cli::cli_abort(message)
 }
 
 #' @rdname dot-use_case_weights_with_yardstick
@@ -55,7 +57,7 @@ extract_case_weights <- function(data, workflow) {
   col <- extract_case_weights_col(workflow)
 
   if (!rlang::is_quosure(col)) {
-    rlang::abort("`col` must exist and be a quosure at this point.", .internal = TRUE)
+    cli::cli_abort("{.arg col} must exist and be a quosure at this point.", .internal = TRUE)
   }
 
   loc <- eval_select_case_weights(col, data)
@@ -63,10 +65,10 @@ extract_case_weights <- function(data, workflow) {
   case_weights <- data[[loc]]
 
   if (!hardhat::is_case_weights(case_weights)) {
-    rlang::abort(paste0(
-      "Case weights must be a supported case weights type, as determined by ",
-      "`hardhat::is_case_weights()`."
-    ))
+    cli::cli_abort(
+      "Case weights must be a supported case weights type, as determined by
+      {.fn hardhat::is_case_weights}."
+    )
   }
 
   case_weights
@@ -96,12 +98,11 @@ eval_select_case_weights <- function(col,
   )
 
   if (length(loc) != 1L) {
-    message <- paste0(
-      "`col` must specify exactly one column from ",
-      "`data` to extract case weights from."
+    cli::cli_abort(
+      "{.arg col} must specify exactly one column from {.arg data} to
+       extract case weights from.",
+      call = call
     )
-
-    rlang::abort(message, call = call)
   }
 
   loc
diff --git a/R/checks.R b/R/checks.R
index f2e209a9..b7111a34 100644
--- a/R/checks.R
+++ b/R/checks.R
@@ -3,38 +3,39 @@
 #' @rdname empty_ellipses
 check_rset <- function(x) {
   if (!inherits(x, "rset")) {
-    rlang::abort(paste0(
-      "The `resamples` argument should be an 'rset' object, such as the type ",
-      "produced by `vfold_cv()` or other 'rsample' functions."
-    ))
+    cli::cli_abort(
+      c(
+        "The {.arg resamples} argument should be an {.cls rset} object,",
+        "i" = "Such objects are produced by {.fn vfold_cv} or other
+           {.pkg rsample} functions."
+      )
+    )
   }
   if (inherits(x, "loo_cv")) {
-    rlang::abort(
-      "Leave-one-out cross-validation is not currently supported with tune."
-    )
+    cli::cli_abort("Leave-one-out cross-validation is not currently supported with {.pkg tune}.")
   }
   if (inherits(x, "nested_cv")) {
-    rlang::abort("Nested resampling is not currently supported with tune.")
+    cli::cli_abort("Nested resampling is not currently supported with tune.")
   }
   if (inherits(x, "permutations")) {
-    rlang::abort("Permutation samples are not suitable for tuning.")
+    cli::cli_abort("Permutation samples are not suitable for tuning.")
   }
   invisible(NULL)
 }
 
-backend_options_msg <- "`backend_options` should be created by `tune::new_backend_options()`."
+backend_options_msg <- "{.arg backend_options} should be created by {.fn tune::new_backend_options}."
 
 check_backend_options <- function(backend_options) {
   if (!is.null(backend_options) &&
       !inherits(backend_options, "tune_backend_options")) {
-    rlang::abort(backend_options_msg)
+    cli::cli_abort(backend_options_msg)
   }
 
   invisible(NULL)
 }
 
 
-grid_msg <- "`grid` should be a positive integer or a data frame."
+grid_msg <- "{.arg grid} should be a positive integer or a data frame."
 
 check_grid <- function(grid, workflow, pset = NULL, call = caller_env()) {
   # `NULL` grid is the signal that we are using `fit_resamples()`
@@ -47,12 +48,11 @@ check_grid <- function(grid, workflow, pset = NULL, call = caller_env()) {
   }
 
   if (nrow(pset) == 0L) {
-    msg <- paste0(
-      "No tuning parameters have been detected, ",
-      "performance will be evaluated using the resamples with no tuning. ",
-      "Did you want to [tune()] parameters?"
-    )
-    rlang::warn(msg)
+    cli::cli_warn(c(
+      "No tuning parameters have been detected, performance will be
+       evaluated using the resamples with no tuning.",
+      "Did you want to assign any parameters with a value of {.fn tune}?"
+    ))
 
     # Return `NULL` as the new `grid`, like what is used in `fit_resamples()`
     return(NULL)
@@ -60,7 +60,7 @@ check_grid <- function(grid, workflow, pset = NULL, call = caller_env()) {
 
   if (!is.numeric(grid)) {
     if (!is.data.frame(grid)) {
-      rlang::abort(grid_msg)
+      cli::cli_abort(grid_msg)
     }
 
     grid_distinct <- distinct(grid)
@@ -69,7 +69,7 @@ check_grid <- function(grid, workflow, pset = NULL, call = caller_env()) {
     grid_distinct <- vctrs::new_data_frame(grid_distinct, n = nrow(grid_distinct))
 
     if (!identical(nrow(grid_distinct), nrow(grid))) {
-      rlang::warn(
+      cli::cli_warn(
         "Duplicate rows in grid of tuning combinations found and removed."
       )
     }
@@ -87,32 +87,28 @@ check_grid <- function(grid, workflow, pset = NULL, call = caller_env()) {
     extra_tune_params <- setdiff(tune_params, grid_params)
 
     if (length(extra_grid_params) != 0L) {
-      extra_grid_params <- glue::single_quote(extra_grid_params)
-      extra_grid_params <- glue::glue_collapse(extra_grid_params, sep = ", ")
+      n_extra <- length(extra_grid_params)
 
-      msg <- glue::glue(
-        "The provided `grid` has the following parameter columns that have ",
-        "not been marked for tuning by `tune()`: {extra_grid_params}."
+      cli::cli_abort(
+        "The provided grid has {n_extra} parameter column{?s}
+         ({.var {extra_grid_params}}) that {?has/have} not been marked for tuning
+         by {.fn tune}."
       )
-
-      rlang::abort(msg)
     }
 
     if (length(extra_tune_params) != 0L) {
-      extra_tune_params <- glue::single_quote(extra_tune_params)
-      extra_tune_params <- glue::glue_collapse(extra_tune_params, sep = ", ")
+      n_extra <- length(extra_tune_params)
 
-      msg <- glue::glue(
-        "The provided `grid` is missing the following parameter columns that ",
-        "have been marked for tuning by `tune()`: {extra_tune_params}."
+      cli::cli_abort(
+        "The provided grid is missing the following {n_extra} parameter
+         column{?s} that {?has/have} been marked for tuning by {.fn tune}:
+         {.val {extra_tune_params}}."
       )
-
-      rlang::abort(msg)
     }
   } else {
     grid <- as.integer(grid[1])
     if (grid < 1) {
-      rlang::abort(grid_msg)
+      cli::cli_abort(grid_msg)
     }
     check_workflow(workflow, pset = pset, check_dials = TRUE, call = call)
 
@@ -158,22 +154,21 @@ check_parameters <- function(wflow, pset = NULL, data, grid_names = character(0)
 
   if (needs_finalization(pset, grid_names)) {
     if (tune_recipe) {
-      rlang::abort(
-        paste(
-          "Some model parameters require finalization but there are recipe",
-          "parameters that require tuning. Please use ",
-          "`extract_parameter_set_dials()` to set parameter ranges ",
-          "manually and supply the output to the `param_info` argument."
+      cli::cli_abort(
+        c(
+          "Some model parameters require finalization but there are recipe
+           parameters that require tuning.",
+          "i" = "Please use {.fn extract_parameter_set_dials} to set parameter
+           ranges manually and supply the output to the {.arg param_info}
+           argument."
         )
       )
     }
-    msg <- "Creating pre-processing data to finalize unknown parameter"
     unk_names <- pset$id[unk]
-    if (length(unk_names) == 1) {
-      msg <- paste0(msg, ": ", unk_names)
-    } else {
-      msg <- paste0(msg, "s: ", paste0("'", unk_names, "'", collapse = ", "))
-    }
+    num_unk <- length(unk_names)
+    msg <-
+      cli::format_inline(
+        "Creating pre-processing data to finalize {num_unk} unknown parameter{?s}: {.val {unk_names}}")
 
     tune_log(list(verbose = TRUE), split_labels = NULL, msg, type = "info")
 
@@ -194,7 +189,7 @@ is_installed <- function(pkg) {
 
 check_installs <- function(x, call = caller_env()) {
   if (x$engine == "unknown") {
-    rlang::abort("Please declare an engine for the model")
+    cli::cli_abort("Please declare an engine for the model")
   } else {
     m_type <- class(x)[1]
     deps <- parsnip::get_dependency(m_type)
@@ -207,7 +202,7 @@ check_installs <- function(x, call = caller_env()) {
     if (any(!is_inst)) {
       needs_installed <- unique(deps[!is_inst])
       cli::cli_abort(
-        "{cli::qty(needs_installed)} Package install{?s} {?is/are} \\
+        "{cli::qty(needs_installed)} Package install{?s} {?is/are}
          required for {.pkg {needs_installed}}.",
         call = call
       )
@@ -218,44 +213,39 @@ check_installs <- function(x, call = caller_env()) {
 check_bayes_initial_size <- function(num_param, num_grid, race = FALSE) {
   msg <-
     cli::pluralize(
-      "There {cli::qty(num_param)}{?is/are} {num_param} tuning parameter{?s} \\
+      "There {cli::qty(num_param)}{?is/are} {num_param} tuning parameter{?s}
       and {num_grid} grid point{?s} {?was/were} requested."
     )
 
+  msg_list <- c(
+    "{msg}",
+    "i" = "The GP model requires 2+ initial points. For best performance,
+               supply more initial points than there are tuning parameters."
+  )
+  bullet_msg <-
+    c(
+      `!` = "{msg}",
+      `*` = cli::pluralize(
+        "There are {cli::qty(diff)}{?as many/more} tuning parameters
+          {cli::qty(diff)}{?as/than} there are initial points.
+          This is likely to cause numerical issues in the first few
+          search iterations.")
+    )
+
 
   if (race) {
     race_msg <- "With racing, only completely resampled parameters are used."
-  } else {
-    race_msg <- NULL
+    msg_list <- c(msg_list, "i" = race_msg)
+    bullet_msg <- c(bullet_msg, `*` = race_msg)
   }
 
   if (num_grid == 1) {
-    rlang::abort(
-      c(
-        msg,
-        glue::glue(
-          "The GP model requires 2+ initial points. For best performance, \\
-          supply more initial points than there are tuning parameters."
-        ),
-        race_msg
-      ),
-      call = NULL
-    )
+    cli::cli_abort(msg_list)
   }
 
   if (num_grid < num_param + 1) {
     diff <- num_param - num_grid + 1
-    cli::cli_bullets(
-      c(
-        `!` = msg,
-        `*` = cli::pluralize(
-          "There are {cli::qty(diff)}{?as many/more} tuning parameters \\
-          {cli::qty(diff)}{?as/than} there are initial points. \\
-          This is likely to cause numerical issues in the first few \\
-          search iterations."),
-        `*` = race_msg
-      )
-    )
+    cli::cli_bullets(bullet_msg)
   }
 
   invisible(NULL)
@@ -265,11 +255,10 @@ check_param_objects <- function(pset) {
   params <- purrr::map_lgl(pset$object, inherits, "param")
 
   if (!all(params)) {
-    rlang::abort(paste0(
-      "The workflow has arguments to be tuned that are missing some ",
-      "parameter objects: ",
-      paste0("'", pset$id[!params], "'", collapse = ", ")
-    ))
+    cli::cli_abort(
+      "The workflow has arguments to be tuned that are missing some parameter
+       objects: {.val {pset$id[!params]}}"
+    )
   }
   invisible(pset)
 }
@@ -280,15 +269,15 @@ check_param_objects <- function(pset) {
 #' @param check_dials A logical for check for a NULL parameter object.
 check_workflow <- function(x, ..., pset = NULL, check_dials = FALSE, call = caller_env()) {
   if (!inherits(x, "workflow")) {
-    rlang::abort("The `object` argument should be a 'workflow' object.")
+    cli::cli_abort("The {.arg object} argument should be a {.cls workflow} object.")
   }
 
   if (!has_preprocessor(x)) {
-    rlang::abort("A formula, recipe, or variables preprocessor is required.")
+    cli::cli_abort("A formula, recipe, or variables preprocessor is required.")
   }
 
   if (!has_spec(x)) {
-    rlang::abort("A parsnip model is required.")
+    cli::cli_abort("A parsnip model is required.")
   }
 
   rlang::check_dots_empty(call = call)
@@ -303,10 +292,10 @@ check_workflow <- function(x, ..., pset = NULL, check_dials = FALSE, call = call
     incompl <- dials::has_unknowns(pset$object)
 
     if (any(incompl)) {
-      rlang::abort(paste0(
-        "The workflow has arguments whose ranges are not finalized: ",
-        paste0("'", pset$id[incompl], "'", collapse = ", ")
-      ))
+      cli::cli_abort(
+        "The workflow has arguments whose ranges are not finalized:
+         {.val {pset$id[incompl]}}."
+      )
     }
   }
 
@@ -359,9 +348,10 @@ check_metrics <- function(x, object) {
              x <- yardstick::metric_set(brier_survival)
            },
            unknown = {
-             rlang::abort("Internal error: `check_installs()` should have caught an `unknown` mode.")
+             cli::cli_abort("Internal error: {.fn check_installs} should have
+                            caught an {.val unknown} mode.")
            },
-           rlang::abort("Unknown `mode` for parsnip model.")
+           cli::cli_abort("Unknown {.val mode} for parsnip model.")
     )
 
     return(x)
@@ -372,37 +362,37 @@ check_metrics <- function(x, object) {
   is_surv_metric_set <- inherits(x, c("survival_metric_set"))
 
   if (!is_numeric_metric_set && !is_class_prob_metric_set && !is_surv_metric_set) {
-    rlang::abort("The `metrics` argument should be the results of [yardstick::metric_set()].")
+    cli::cli_abort("The {.arg metrics} argument should be the results of
+                   {.fn yardstick::metric_set}.")
   }
 
   if (mode == "regression" && !is_numeric_metric_set) {
-    msg <- paste0(
-      "The parsnip model has `mode = 'regression'`, ",
-      "but `metrics` is a metric set for a different model mode."
+    cli::cli_abort(
+      c(
+        "The parsnip model has {.code mode = 'regression'}, but {.arg metrics}
+        is a metric set for a different model mode."
+      )
     )
-    rlang::abort(msg)
   }
 
   if (mode == "classification" && !is_class_prob_metric_set) {
-    msg <- paste0(
-      "The parsnip model has `mode = 'classification'`, ",
-      "but `metrics` is a metric set for a different model mode."
+    cli::cli_abort(
+      c(
+        "The parsnip model has {.code mode = 'classification'}, but {.arg metrics}
+         is a metric set for a different model mode."
+      )
     )
-    rlang::abort(msg)
   }
 
   if (mode == "censored regression" && !is_surv_metric_set) {
-    msg <- paste0(
-      "The parsnip model has `mode = 'censored regression'`, ",
-      "but `metrics` is a metric set for a different model mode."
-    )
-    rlang::abort(msg)
+    cli::cli_abort(c(
+      "The parsnip model has {.code mode = 'censored regression'},
+       but {.arg metrics} is a metric set for a different model mode."
+    ))
   }
   x
 }
 
-bayes_msg <- "`initial` should be a positive integer or the results of [tune_grid()]"
-
 #' @export
 #' @keywords internal
 #' @rdname empty_ellipses
@@ -418,13 +408,14 @@ check_initial <- function(x,
                           ctrl,
                           checks = "grid") {
   if (is.null(x)) {
-    rlang::abort(bayes_msg)
+    cli::cli_abort("{.arg initial} should be a positive integer or the results
+                   of {.fn tune_grid}")
   }
   if (is.numeric(x)) {
     x <- create_initial_set(pset, n = x, checks = checks)
     if (ctrl$verbose) {
       message()
-      msg <- paste0(" Generating a set of ", nrow(x), " initial parameter results")
+      msg <- cli::format_inline(" Generating a set of {nrow(x)} initial parameter results")
       tune_log(ctrl, split_labels = NULL, msg, type = "go")
     }
 
@@ -446,13 +437,16 @@ check_initial <- function(x,
     }
   } else {
     if (!inherits(x, "tune_results")) {
-      rlang::abort(bayes_msg)
+      cli::cli_abort("{.arg initial} should be a positive integer or the results
+                      of {.fn tune_grid}")
     }
     if (ctrl$save_pred & !any(names(x) == ".predictions")) {
-      rlang::abort("`save_pred` can only be used if the initial results saved predictions.")
+      cli::cli_abort("{.arg save_pred} can only be used if the initial results
+                      saved predictions.")
     }
     if (!is.null(ctrl$extract) & !any(names(x) == ".extracts")) {
-      rlang::abort("`extract` can only be used if the initial results has extractions.")
+      cli::cli_abort("The {.fn extract} function can only be used if the initial
+                     results have extractions.")
     }
     param_nms <- .get_tune_parameter_names(x)
     if (inherits(x, "tune_race")) {
@@ -565,9 +559,9 @@ check_gp_data <- function(x) {
 
   n_uni <- length(unique(x$mean))
   if (n_uni == 1) {
-    msg <- glue::glue(
+    msg <- cli::pluralize(
       "All of the {met} values were identical. The Gaussian process model cannot
-     be fit to the data. Try expanding the range of the tuning parameters."
+       be fit to the data. Try expanding the range of the tuning parameters."
     )
     message_wrap(msg, prefix = "!", color_text = get_tune_colors()$message$danger)
   }
@@ -583,7 +577,7 @@ check_gp_failure <- function(current, prev) {
 
   # first model failed or all previous models failed
   if (is.null(prev) || inherits(prev, "try-error")) {
-    rlang::abort("Gaussian process model was not fit.")
+    cli::cli_abort("Gaussian process model was not fit.")
   }
 
   # return prev model
@@ -597,12 +591,12 @@ check_no_tuning <- function(x) {
     return(invisible(FALSE))
   }
   srcs <- unique(tune_param$source)
-  num_srcs <- length(srcs)
-  srcs <- paste0(srcs, collapse = " and ")
-  msg_1 <- cli::pluralize("{num_param} argument{?s} {?has/have} been tagged for tuning in {?this component/these components}: {srcs}. ")
-  msg_2 <- "Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them."
-  msg <- paste(msg_1, msg_2, sep = "\n")
-  rlang::abort(msg, call = NULL)
+  cli::cli_abort(c(
+    "{num_param} argument{?s} {?has/have} been tagged for tuning in
+     {?this component/these components}: {srcs}.",
+    "i" = "Please use one of the tuning functions (e.g. {.fn tune_grid})
+           to optimize them."
+  ))
 }
 
 dyn_inputs <- c("integrated_survival_metric", "dynamic_survival_metric")
@@ -617,8 +611,9 @@ check_eval_time <- function(eval_time, metrics) {
     )
   }
   if (is.null(eval_time) & needs_eval_time) {
-    rlang::abort(
-      "One or more metric requires the specification of time points in the `eval_time` argument.",
+    cli::cli_abort(
+      "One or more metric requires the specification of time points in the
+       {.arg eval_time} argument.",
       call = NULL
     )
   }
diff --git a/R/collect.R b/R/collect.R
index c3442160..b61ca8e9 100644
--- a/R/collect.R
+++ b/R/collect.R
@@ -152,12 +152,10 @@ collect_predictions.tune_results <- function(x, ..., summarize = FALSE, paramete
   has_coll_col <- coll_col %in% names
 
   if (!has_coll_col) {
-    msg <- paste0(
-      "The `.predictions` column does not exist. ",
-      "Refit with the control argument `save_pred = TRUE` to save predictions."
+    cli::cli_abort(
+      "The {.field .predictions} column does not exist. Please refit with the
+      control argument {.code save_pred = TRUE} to save predictions."
     )
-
-    rlang::abort(msg)
   }
 
   x <- filter_predictions(x, parameters)
@@ -183,17 +181,9 @@ filter_predictions <- function(x, parameters) {
   }
   params <- attr(x, "parameters")
   if (is.null(params)) {
-    rlang::warn(
-      paste(
-        strwrap(
-          paste(
-            "The object is missing some attributes; it is probably from",
-            "an earlier version of `tune`. The predictions can't be filtered."
-          ),
-          prefix = ""
-        ),
-        collapse = "\n"
-      )
+    cli::cli_warn(
+      "The object is missing some attributes; it is probably from an earlier 
+       version of {.pkg tune}. The predictions can't be filtered."
     )
 
     return(x)
@@ -202,12 +192,8 @@ filter_predictions <- function(x, parameters) {
   param_names <- params$id
   parameters <- dplyr::select(parameters, dplyr::any_of(param_names))
   if (ncol(parameters) != length(param_names)) {
-    rlang::abort(
-      paste0(
-        "`parameters` should only have columns: ",
-        paste0("'", param_names, "'", collapse = ", ")
-      )
-    )
+    cli::cli_abort("{.arg parameters} should only have columns:
+                    {.val {param_names}}.")
   }
   x$.predictions <-
     purrr::map(x$.predictions, dplyr::inner_join, parameters, by = param_names)
@@ -400,12 +386,7 @@ average_predictions <- function(x, grid = NULL) {
   if (!is.null(grid)) {
     grid <- dplyr::select(grid, dplyr::all_of(param_names))
     if (ncol(grid) != length(param_names)) {
-      rlang::abort(
-        paste0(
-          "`grid` should only have columns: ",
-          paste0("'", param_names, "'", collapse = ", ")
-        )
-      )
+      cli::cli_abort("{.arg grid} should only have columns: {.val {param_names}}.")
     }
     x$.predictions <-
       purrr::map(x$.predictions, dplyr::inner_join, grid, by = param_names)
@@ -429,11 +410,9 @@ average_predictions <- function(x, grid = NULL) {
   } else if (any(metric_types %in% c("survival", "time"))) {
     x <- surv_summarize(x, param_names, y_nms)
   } else {
-    msg <- paste(
-      "We don't know about metrics of type:",
-      paste(unique(metric_types), collapse = ", ")
+    cli::cli_abort(
+      "We don't know about metrics of type: {.val {unique(metric_types)}}."
     )
-    rlang::abort(msg)
   }
 
   if (dplyr::is_grouped_df(x)) {
@@ -566,7 +545,10 @@ estimate_tune_results <- function(x, ..., col_name = ".metrics") {
 
   all_bad <- is_cataclysmic(x)
   if (all_bad) {
-    rlang::abort("All models failed. Run `show_notes(.Last.tune.result)` for more information.")
+    cli::cli_abort(
+      "All models failed. Run {.code show_notes(.Last.tune.result)} for more
+       information."
+    )
   }
 
   # The mapping of tuning parameters and .config.
diff --git a/R/compat-vctrs.R b/R/compat-vctrs.R
index 5db8591d..a49c5651 100644
--- a/R/compat-vctrs.R
+++ b/R/compat-vctrs.R
@@ -175,5 +175,5 @@ stop_incompatible_cast_iteration_results <- function(x, to, ..., x_arg = "", to_
 }
 
 stop_never_called <- function(fn) {
-  rlang::abort(paste0("Internal error: `", fn, "()` should never be called."))
+  cli::cli_abort("Internal error: {.fn {fn}} should never be called.")
 }
diff --git a/R/compute_metrics.R b/R/compute_metrics.R
index 94690e90..b62727f6 100644
--- a/R/compute_metrics.R
+++ b/R/compute_metrics.R
@@ -76,7 +76,8 @@ compute_metrics.default <- function(x,
                                     summarize = TRUE,
                                     event_level = "first",
                                     ...) {
-  rlang::abort("No `compute_metrics()` method exists for this type of object.")
+  cli::cli_abort("No {.fn compute_metrics} method exists for
+                 {.obj_type_friendly {x}}.")
 }
 
 #' @export
@@ -89,14 +90,14 @@ compute_metrics.tune_results <- function(x,
   rlang::check_dots_empty()
   check_bool(summarize)
   if (!".predictions" %in% names(x)) {
-    rlang::abort(paste0(
-      "`x` must have been generated with the ",
-      "control argument `save_pred = TRUE`."
-    ))
+    cli::cli_abort(
+      "{.arg x} must have been generated with the control argument
+      {.code save_pred = TRUE}."
+    )
   }
 
   if (!inherits(metrics, "metric_set")) {
-    rlang::abort("`metrics` must be a metric set.")
+    cli::cli_abort("{.arg metrics} must be a metric set.")
   }
 
   new_metrics_info <- metrics_info(metrics)
diff --git a/R/conf_mat_resampled.R b/R/conf_mat_resampled.R
index 49b1e7ef..f6e9fc5f 100644
--- a/R/conf_mat_resampled.R
+++ b/R/conf_mat_resampled.R
@@ -36,22 +36,21 @@
 conf_mat_resampled <- function(x, ..., parameters = NULL, tidy = TRUE) {
   rlang::check_dots_empty()
   if (!inherits(x, "tune_results")) {
-    rlang::abort(
-      "The first argument needs to be an object with class 'tune_results'."
-    )
+    cli::cli_abort("The first argument needs to be {.cls tune_results} object,
+                   not {.obj_type_friendly {mtcars}}.")
   }
   if (!any(names(x) == ".predictions")) {
-    rlang::abort(
-      paste0(
-        "The function was not run with the `save_pred = TRUE` option. ",
-        "Please re-run with that option."
-      )
+    cli::cli_abort(
+      "The function was not run with the {.code save_pred = TRUE} option.
+       Please re-run with that option."
     )
   }
   preds <- collect_predictions(x, summarize = FALSE, parameters = parameters)
 
   if (!any(names(preds) == ".pred_class")) {
-    rlang::abort("Cannot find the predicted classes. Was this a classification model?")
+    cli::cli_abort(
+      "Cannot find the predicted classes. Was this a classification model?"
+    )
   }
   # check for multiple parameter combinations
   params <- .get_tune_parameter_names(x)
@@ -61,19 +60,17 @@ conf_mat_resampled <- function(x, ..., parameters = NULL, tidy = TRUE) {
       dplyr::select(!!!params) %>%
       distinct()
     if (nrow(param_combos) > 1) {
-      rlang::abort(
-        paste0(
-          "It looks like there are ", nrow(param_combos), " tuning parameter ",
-          "combination(s) in the data. Please use the `parameters` ",
-          "argument to select one combination of parameters."
-        )
+      cli::cli_abort(
+        "It looks like there are {nrow(param_combos)} tuning parameter
+        combination in the data. Please use the {.arg parameters}
+        argument to select one combination of parameters."
       )
     }
   }
 
   truth <- .get_tune_outcome_names(x)
   if (length(truth) != 1) {
-    rlang::abort("Cannot determine the proper outcome name")
+    cli::cli_abort("Cannot determine the proper outcome name")
   }
 
   id_cols <- grep("(^id$)|($id[1-9]$)", names(preds), value = TRUE)
diff --git a/R/control.R b/R/control.R
index 9dc69d6c..fded0799 100644
--- a/R/control.R
+++ b/R/control.R
@@ -312,7 +312,7 @@ new_backend_options <- function(..., class = character()) {
   out <- rlang::list2(...)
 
   if (any(rlang::names2(out) == "")) {
-    rlang::abort("All backend options must be named.")
+    cli::cli_abort("All backend options must be named.")
   }
 
   structure(out, class = c(class, "tune_backend_options"))
diff --git a/R/extract.R b/R/extract.R
index 0a167f1e..8c6e51c2 100644
--- a/R/extract.R
+++ b/R/extract.R
@@ -115,10 +115,13 @@ extract_recipe.tune_results <- function(x, ..., estimated = TRUE) {
 check_empty_dots <- function(...) {
   opts <- list(...)
   if (any(names(opts) == "estimated")) {
-    rlang::abort("'estimated' should be a named argument.")
+    cli::cli_abort("'{.arg estimated}' should be a named argument.")
   }
   if (length(opts) > 0) {
-    rlang::abort("'...' are not used in this function.")
+    cli::cli_warn(
+      "The {.code ...} are not used in this function but {length(opts)}
+       object{?s} {?was/were} passed: {.val {names(opts)}}"
+    )
   }
   invisible(NULL)
 }
diff --git a/R/filter_parameters.R b/R/filter_parameters.R
index afff0fc2..0658c068 100644
--- a/R/filter_parameters.R
+++ b/R/filter_parameters.R
@@ -57,7 +57,8 @@ filter_parameters <- function(x, ..., parameters = NULL) {
   check_filter_dots(rlang::enquos(...))
   # check for type
   if (!inherits(x, "tune_results")) {
-    rlang::abort(paste0(cl_x, " should have class 'tune_results'."))
+    cli::cli_abort("{.arg {caller_arg(x)}} should have class {.cls tune_results};
+                    {.obj_type_friendly {x}} was passed.")
   }
   x <- filter_by_join(x, parameters, nm = cl_x)
   x <- filter_by_filter(x, ...)
@@ -73,11 +74,9 @@ filter_by_join <- function(x, parameters = NULL, nm = "") {
   filter_names <- names(parameters)
   filter_names <- filter_names[filter_names != ".config"]
   if (length(intersect(filter_names, param_names)) == 0) {
-    msg <- paste0(
-      "There are no columns in 'parameters' that match with ",
-      as.character(nm)
+    cli::cli_abort(
+      "There are no columns in {.arg parameters} that match with {.val {nm}}."
     )
-    rlang::abort(msg)
   }
   extra_names <- setdiff(filter_names, param_names)
   if (length(extra_names) > 0) {
@@ -93,7 +92,7 @@ filter_by_join <- function(x, parameters = NULL, nm = "") {
   tst_orig <- bind_rows(x$.metrics)
   tst_filtered <- filter_join_iter(tst_orig, parameters)
   if (nrow(tst_filtered) == 0) {
-    rlang::abort("No parameter combinations were selected using your subset.")
+    cli::cli_abort("No parameter combinations were selected using your subset.")
   }
 
   x$.metrics <- purrr::map(x$.metrics, filter_join_iter, .subset = parameters)
@@ -133,9 +132,12 @@ check_filter_dots <- function(dots, call = rlang::caller_env()) {
   res <- purrr::map(dots, ~try(rlang::eval_tidy(.x), silent = TRUE))
 
   if (any(purrr::map_lgl(res, inherits, "data.frame"))) {
-    rlang::abort(
-      c("An element passed to `...` is a data frame rather than a filter expression.",
-        "i" = "Did you forget to name the `parameters` argument?"),
+    cli::cli_abort(
+      c(
+        "An element passed to {.arg ...} is a data frame rather than a filter
+         expression.",
+        "i" = "Did you forget to name the {.arg parameters} argument?"
+      ),
       call = call
     )
   }
diff --git a/R/finalize.R b/R/finalize.R
index a3f9b073..a0c365b4 100644
--- a/R/finalize.R
+++ b/R/finalize.R
@@ -30,7 +30,8 @@
 #' finalize_model(knn_model, lowest_rmse)
 finalize_model <- function(x, parameters) {
   if (!inherits(x, "model_spec")) {
-    rlang::abort("`x` should be a parsnip model specification.")
+    cli::cli_abort("{.arg x} should be a parsnip model specification, not
+                    {.obj_type_friendly {x}}.")
   }
   check_final_param(parameters)
   pset <- hardhat::extract_parameter_set_dials(x)
@@ -54,7 +55,7 @@ finalize_model <- function(x, parameters) {
 #' @rdname finalize_model
 finalize_recipe <- function(x, parameters) {
   if (!inherits(x, "recipe")) {
-    rlang::abort("`x` should be a recipe.")
+    cli::cli_abort("{.arg x} should be a recipe, not {.obj_type_friendly {x}}.")
   }
   check_final_param(parameters)
   pset <-
@@ -80,7 +81,7 @@ finalize_recipe <- function(x, parameters) {
 #' @rdname finalize_model
 finalize_workflow <- function(x, parameters) {
   if (!inherits(x, "workflow")) {
-    rlang::abort("`x` should be a workflow")
+    cli::cli_abort("{.arg x} should be a workflow, not {.obj_type_friendly {x}}.")
   }
   check_final_param(parameters)
 
@@ -101,10 +102,11 @@ finalize_workflow <- function(x, parameters) {
 
 check_final_param <- function(x) {
   if (!is.list(x) & !tibble::is_tibble(x)) {
-    rlang::abort("The parameter object should be a list or tibble")
+    cli::cli_abort("The parameter object should be a {.cls list} or
+                   {.cls tibble}, not {.obj_type_friendly {x}}.")
   }
   if (tibble::is_tibble(x) && nrow(x) > 1) {
-    rlang::abort("The parameter tibble should have a single row.")
+    cli::cli_abort("The parameter {.arg tibble} should have a single row.")
   }
   invisible(x)
 }
diff --git a/R/fit_best.R b/R/fit_best.R
index 0590a303..d91b7380 100644
--- a/R/fit_best.R
+++ b/R/fit_best.R
@@ -78,10 +78,8 @@ fit_best <- function(x, ...) {
 #' @export
 #' @rdname fit_best
 fit_best.default <- function(x, ...) {
-  cls <- class(x)
   cli::cli_abort(
-    "There is no `fit_best()` method for an object with \\
-     {cli::qty(cls)} class{?es} {.var {cls}}."
+    "No {.fn fit_best} exists for {.obj_type_friendly {x}}."
   )
 }
 
@@ -162,8 +160,9 @@ fit_best.tune_results <- function(x,
     }
   } else {
     if (!is.null(add_validation_set)) {
-      rlang::warn(
-        "The option `add_validation_set` is being ignored because the resampling object does not include a validation set."
+      cli::cli_warn(
+        "The option {.arg add_validation_set} is being ignored because the
+         resampling object does not include a validation set."
       )
     }
     dat <- x$splits[[1]]$data
diff --git a/R/grid_code_paths.R b/R/grid_code_paths.R
index 505d6ccc..51d4f88e 100644
--- a/R/grid_code_paths.R
+++ b/R/grid_code_paths.R
@@ -96,7 +96,8 @@ tune_grid_loop_agua <- function(resamples,
                                 eval_time = NULL,
                                 rng) {
   if (!rlang::is_installed("agua")) {
-    rlang::abort("`agua` must be installed to use an h2o parsnip engine.")
+    cli::cli_abort("The {.pkg agua} package must be installed to use an h2o
+                   parsnip engine.")
   }
 
   parallel_over <- control$parallel_over
@@ -126,7 +127,8 @@ parallel_over_finalize_agua <- function(parallel_over) {
   }
 
   if (!identical(parallel_over, "resamples")) {
-    rlang::abort("`agua` only supports `parallel_over = \"resamples\".")
+    cli::cli_abort("The {.arg agua} package only supports
+                   {.code parallel_over = 'resamples'}.")
   }
 
   parallel_over
@@ -298,7 +300,7 @@ tune_grid_loop_impl <- function(fn_tune_grid_loop_iter,
     return(results)
   }
 
-  rlang::abort("Invalid `parallel_over`.", .internal = TRUE)
+  cli::cli_abort("Invalid {.arg parallel_over}.", .internal = TRUE)
 }
 
 # ------------------------------------------------------------------------------
diff --git a/R/grid_helpers.R b/R/grid_helpers.R
index 365386d5..91b6131e 100644
--- a/R/grid_helpers.R
+++ b/R/grid_helpers.R
@@ -15,25 +15,22 @@ predict_model <- function(new_data, orig_rows, workflow, grid, metrics,
   }
 
   if (length(orig_rows) != nrow(x_vals)) {
-    msg <- paste0(
-      "Some assessment set rows are not available at ",
-      "prediction time. "
-    )
+    msg <- "Some assessment set rows are not available at prediction time."
 
     if (has_preprocessor_recipe(workflow)) {
-      msg <- paste0(
-        msg,
-        "Consider using `skip = TRUE` on any recipe steps that remove rows ",
-        "to avoid calling them on the assessment set."
-      )
+      msg <-
+        c(
+          msg,
+          i = 
+            "Consider using {.code skip = TRUE} on any recipe steps that
+             remove rows to avoid calling them on the assessment set."
+          
+        )
     } else {
-      msg <- paste0(
-        msg,
-        "Did your preprocessing steps filter or remove rows?"
-      )
+      msg <- c(msg, i = "Did your preprocessing steps filter or remove rows?")
     }
 
-    rlang::abort(msg)
+    cli::cli_abort(msg)
   }
 
   # Determine the type of prediction that is required
diff --git a/R/grid_performance.R b/R/grid_performance.R
index f9983124..fcfa024a 100644
--- a/R/grid_performance.R
+++ b/R/grid_performance.R
@@ -76,10 +76,10 @@ metrics_info <- function(x) {
   types <- unique(metrics_info$type)
 
   if (length(outcome_name) > 1L) {
-    rlang::abort(paste0(
-      "Internal error: Multiple outcomes are not ",
-      "supported in `.estimate_metrics()`."
-    ))
+    cli::cli_abort(
+      "Multiple outcomes are not supported in {.fn .estimate_metrics}.",
+      .internal = TRUE
+    )
   }
 
   if (case_weights_column_name() %in% names(dat)) {
@@ -95,7 +95,7 @@ metrics_info <- function(x) {
   } else if (all(types == "time" | types == "survival")) {
     estimate_surv(dat, metric, param_names, outcome_name, case_weights, types)
   } else {
-    rlang::abort("Metric type not yet supported by tune.")
+    cli::cli_abort("Metric type not yet supported by {.pkg tune}.")
   }
 }
 
@@ -127,7 +127,8 @@ estimate_class_prob <- function(dat, metric, param_names, outcome_name,
       } else if (identical(event_level, "second")) {
         probs <- probs[[2]]
       } else {
-        rlang::abort("`event_level` must be either 'first' or 'second'.")
+        cli::cli_abort("{.arg event_level} must be either {.val first} or
+                        {.val second}.")
       }
     }
   }
@@ -162,18 +163,18 @@ unnest_parameters <- function(x, params = NULL) {
    if (is.null(params)) {
     return(x)
    }
-   
+
   # When multi_predict() is used, .pred will have the tuning parameter values.
   # Other (non-submodel) parameters will be outside of 'x'.
   # If this happens, pull the submodel parameters out of .pred and put them at
   # the out level ('x') with the rest (if any)
-  
+
   outer_nms <- names(x)
   has_pred <- any(outer_nms == ".pred")
    if (!has_pred) {
     return(x)
   }
-  
+
   inner_nms <- names(x$.pred[[1]])
   has_inner_params <- any(params %in% inner_nms)
   if (!has_inner_params) {
diff --git a/R/last_fit.R b/R/last_fit.R
index 611d0eae..d080b752 100644
--- a/R/last_fit.R
+++ b/R/last_fit.R
@@ -98,11 +98,10 @@ last_fit <- function(object, ...) {
 
 #' @export
 last_fit.default <- function(object, ...) {
-  msg <- paste0(
-    "The first argument to [last_fit()] should be either ",
-    "a model or workflow."
+  cli::cli_abort(
+    "The first argument to {.fn last_fit} should be either a model or workflow,
+    not {.obj_type_friendly {object}}."
   )
-  rlang::abort(msg)
 }
 
 #' @export
@@ -117,16 +116,17 @@ last_fit.model_fit <- function(object, ...) {
   ))
 }
 
+tune_pp_msg <- "To tune a model specification, you must preprocess with a
+                formula, recipe, or {.pkg dplyr} selectors, not
+                {.obj_type_friendly {preprocessor}}."
+
 #' @export
 #' @rdname last_fit
 last_fit.model_spec <- function(object, preprocessor, split, ..., metrics = NULL,
                                 eval_time = NULL, control = control_last_fit(),
                                 add_validation_set = FALSE) {
   if (rlang::is_missing(preprocessor) || !is_preprocessor(preprocessor)) {
-    rlang::abort(paste(
-      "To tune a model spec, you must preprocess",
-      "with a formula or recipe"
-    ))
+    cli::cli_abort(tune_pp_msg)
   }
 
   control <- parsnip::condense_control(control, control_last_fit())
diff --git a/R/load_ns.R b/R/load_ns.R
index 23d2cb73..c036d3d4 100644
--- a/R/load_ns.R
+++ b/R/load_ns.R
@@ -53,8 +53,7 @@ full_load <- c("kknn", "earth")
     did_load <- purrr::map_lgl(x, requireNamespace, quietly = TRUE)
     if (any(!did_load)) {
       bad <- x[!did_load]
-      msg <- paste0("'", bad, "'", collapse = ", ")
-      rlang::abort(paste("These packages could not be loaded:", msg))
+      cli::cli_abort("The package{?s} {.pkg {bad}} could not be loaded.")
     }
   }
 
diff --git a/R/logging.R b/R/logging.R
index 335c66e1..448a90b3 100644
--- a/R/logging.R
+++ b/R/logging.R
@@ -390,7 +390,8 @@ format_msg <- function(loc, msg) {
   }
 
   if (!is_workflow(result)) {
-    rlang::abort("Internal error: Model result is not a workflow!")
+    cli::cli_abort("Internal error: Model result is not a workflow, but
+                   not {.obj_type_friendly {object}.")
   }
 
   # Extract the parsnip model from the fitted workflow
diff --git a/R/merge.R b/R/merge.R
index 7c6758c1..3b27b03e 100644
--- a/R/merge.R
+++ b/R/merge.R
@@ -79,8 +79,7 @@ update_model <- function(grid, object, pset, step_id, nms, ...) {
   for (i in nms) {
     param_info <- pset %>% dplyr::filter(id == i & source == "model_spec")
     if (nrow(param_info) > 1) {
-      # TODO figure this out and write a better message
-      rlang::abort("There are too many things.")
+      cli::cli_abort("Cannot update; there are too many parameters.")
     }
     if (nrow(param_info) == 1) {
       if (param_info$component_id == "main") {
@@ -111,7 +110,7 @@ update_recipe <- function(grid, object, pset, step_id, nms, ...) {
 
 merger <- function(x, y, ...) {
   if (!is.data.frame(y)) {
-    rlang::abort("The second argument should be a data frame.")
+    cli::cli_abort("The second argument should be a data frame.")
   }
   pset <- hardhat::extract_parameter_set_dials(x)
 
diff --git a/R/min_grid.R b/R/min_grid.R
index 406c31ff..4e8ef71a 100644
--- a/R/min_grid.R
+++ b/R/min_grid.R
@@ -102,7 +102,7 @@ get_fixed_args <- function(info) {
 
 get_submodel_info <- function(spec) {
   if (is.null(spec$engine)) {
-    rlang::abort("Please set the model's engine.")
+    cli::cli_abort("Please set the model's engine.")
   }
   param_info <-
     get_from_env(paste0(class(spec)[1], "_args")) %>%
@@ -255,7 +255,7 @@ min_grid.linear_reg <- function(x, grid, ...) {
 
 no_penalty <- function(x, nm) {
   if (length(nm) == 0 || all(colnames(x) != nm)) {
-    rlang::abort("At least one penalty value is required for glmnet.")
+    cli::cli_abort("At least one penalty value is required for {.pkg glmnet}.")
   }
   invisible(NULL)
 }
diff --git a/R/plots.R b/R/plots.R
index 81ecc5b7..66d7c8c9 100644
--- a/R/plots.R
+++ b/R/plots.R
@@ -85,18 +85,16 @@ autoplot.tune_results <-
     type <- match.arg(type)
     has_iter <- any(names(object) == ".iter")
     if (!has_iter && type != "marginals") {
-      rlang::abort(paste0("`type = ", type, "` is only used with iterative search results."))
+      cli::cli_abort("{.code type = {glue::double_quote(type)}} is only used with
+                      iterative search results.")
     }
     pset <- .get_tune_parameters(object)
     if (any(is.na(pset$object))) {
       p_names <- pset$id[is.na(pset$object)]
-      msg <-
-        paste0(
-          "Some parameters do not have corresponding parameter objects ",
-          "and cannot be used with `autoplot()`: ",
-          paste0("'", p_names, "'", collapse = ", ")
-        )
-      rlang::abort(msg)
+      cli::cli_abort(
+        "Some parameters do not have corresponding parameter objects and
+         cannot be used with {.fn autoplot}: {.arg {p_names}}."
+      )
     }
 
     if (type == "parameters") {
@@ -122,7 +120,7 @@ autoplot.tune_results <-
 
 #' @export
 autoplot.resample_results <- function(object, ...) {
-  rlang::abort("There is no `autoplot()` implementation for `resample_results`.")
+  cli::cli_abort("There is no {.fn autoplot} implementation for {.cls resample_results}.")
 }
 
 # ------------------------------------------------------------------------------
@@ -322,7 +320,8 @@ plot_param_vs_iter <- function(x, call = rlang::caller_env()) {
   param_cols <- get_param_columns(x)
   pset <- get_param_object(x)
   if (is.null(pset)) {
-    rlang::abort("`autoplot()` requires objects made with tune version 0.1.0 or later.")
+    cli::cli_abort("{.fn autoplot} requires objects made with {.pkg tune}
+                    version 0.1.0 or later.")
   }
 
   # ----------------------------------------------------------------------------
@@ -378,7 +377,8 @@ plot_marginals <- function(x, metric = NULL, eval_time = NULL, call = rlang::cal
   param_cols <- get_param_columns(x)
   pset <- get_param_object(x)
   if (is.null(pset)) {
-    rlang::abort("`autoplot()` requires objects made with tune version 0.1.0 or later.")
+    cli::cli_abort("{.fn autoplot} requires objects made with {.pkg tune}
+                   version 0.1.0 or later.")
   }
 
   # ----------------------------------------------------------------------------
@@ -408,10 +408,10 @@ plot_marginals <- function(x, metric = NULL, eval_time = NULL, call = rlang::cal
     num_param_cols <- param_cols[is_num]
     chr_param_cols <- param_cols[!is_num]
     if (length(chr_param_cols) > 1) {
-      rlang::abort("Currently cannot autoplot grids with 2+ non-numeric parameters.")
+      cli::cli_abort("Currently cannot autoplot grids with 2+ non-numeric parameters.")
     }
     if (length(num_param_cols) == 0) {
-      rlang::abort("Currently cannot autoplot grids with only non-numeric parameters.")
+      cli::cli_abort("Currently cannot autoplot grids with only non-numeric parameters.")
     }
     num_val <- num_val[param_cols %in% chr_param_cols]
     names(num_val) <- chr_param_cols
@@ -521,7 +521,8 @@ plot_regular_grid <- function(x,
   param_cols <- get_param_columns(x)
   pset <- get_param_object(x)
   if (is.null(pset)) {
-    rlang::abort("`autoplot()` requires objects made with tune version 0.1.0 or later.")
+    cli::cli_abort("The {.fn autoplot} function requires objects made with
+                    {.pkg tune} version 0.1.0 or later.")
   }
 
   grd <- dat %>% dplyr::select(all_of(param_cols))
diff --git a/R/pull.R b/R/pull.R
index 63c5e6e6..89c5da2a 100644
--- a/R/pull.R
+++ b/R/pull.R
@@ -107,11 +107,10 @@ reduce_all_outcome_names <- function(resamples) {
   }
 
   if (n_unique > 1L) {
-    rlang::warn(paste0(
-      "More than one set of outcomes were used when tuning. ",
-      "This should never happen. ",
-      "Review how the outcome is specified in your model."
-    ))
+    cli::cli_warn(
+      "More than one set of outcomes were used when tuning. This should never
+       happen. Please review how the outcome is specified in your model."
+    )
   }
 
   outcome_names <- all_outcome_names[[1L]]
diff --git a/R/resample.R b/R/resample.R
index 237ce4d8..44e5b4ef 100644
--- a/R/resample.R
+++ b/R/resample.R
@@ -60,11 +60,10 @@ fit_resamples <- function(object, ...) {
 
 #' @export
 fit_resamples.default <- function(object, ...) {
-  msg <- paste0(
-    "The first argument to [fit_resamples()] should be either ",
-    "a model or workflow."
+  cli::cli_abort(
+    "The first argument to {.fn fit_resamples} should be either a model or workflow,
+    not {.obj_type_friendly {object}}."
   )
-  rlang::abort(msg)
 }
 
 #' @export
@@ -77,10 +76,7 @@ fit_resamples.model_spec <- function(object,
                                      eval_time = NULL,
                                      control = control_resamples()) {
   if (rlang::is_missing(preprocessor) || !is_preprocessor(preprocessor)) {
-    rlang::abort(paste(
-      "To tune a model spec, you must preprocess",
-      "with a formula or recipe"
-    ))
+    cli::cli_abort(tune_pp_msg)
   }
 
   control <- parsnip::condense_control(control, control_resamples())
diff --git a/R/select_best.R b/R/select_best.R
index b6e9bacd..ef69df29 100644
--- a/R/select_best.R
+++ b/R/select_best.R
@@ -32,7 +32,7 @@
 #' units). See details below.
 #' @param eval_time A single numeric time point where dynamic event time
 #' metrics should be chosen (e.g., the time-dependent ROC curve, etc). The
-#' values should be consistent with the values used to create `x`. The `NULL` 
+#' values should be consistent with the values used to create `x`. The `NULL`
 #' default will automatically use the first evaluation time used by `x`.
 #' @param call The call to be shown in errors and warnings.
 #' @return A tibble with columns for the parameters. [show_best()] also
@@ -72,16 +72,16 @@ show_best <- function(x, ...) {
 #' @export
 #' @rdname show_best
 show_best.default <- function(x, ...) {
-  rlang::abort("No `show_best()` exists for this type of object.")
+  cli::cli_abort("No {.fn show_best} exists for this type of object.")
 }
 
 #' @export
 #' @rdname show_best
-show_best.tune_results <- function(x, 
+show_best.tune_results <- function(x,
                                    ...,
-                                   metric = NULL, 
+                                   metric = NULL,
                                    eval_time = NULL,
-                                   n = 5, 
+                                   n = 5,
                                    call = rlang::current_env()) {
   rlang::check_dots_empty()
 
@@ -114,7 +114,7 @@ select_best <- function(x, ...) {
 #' @export
 #' @rdname show_best
 select_best.default <- function(x, ...) {
-  rlang::abort("No `select_best()` exists for this type of object.")
+  cli::cli_abort("No {.fn select_best} exists for this type of object.")
 }
 
 #' @export
@@ -141,7 +141,7 @@ select_by_pct_loss <- function(x, ...) {
 #' @export
 #' @rdname show_best
 select_by_pct_loss.default <- function(x, ...) {
-  rlang::abort("No `select_by_pct_loss()` exists for this type of object.")
+  cli::cli_abort("No {.fn select_by_pct_loss} exists for this type of object.")
 }
 
 #' @export
@@ -207,7 +207,7 @@ select_by_one_std_err <- function(x, ...) {
 #' @export
 #' @rdname show_best
 select_by_one_std_err.default <- function(x, ...) {
-  rlang::abort("No `select_by_one_std_err()` exists for this type of object.")
+  cli::cli_abort("No {.fn select_by_one_std_err} exists for this type of object.")
 }
 
 #' @export
diff --git a/R/tune_bayes.R b/R/tune_bayes.R
index fd2b1e02..9603b448 100644
--- a/R/tune_bayes.R
+++ b/R/tune_bayes.R
@@ -182,11 +182,10 @@ tune_bayes <- function(object, ...) {
 
 #' @export
 tune_bayes.default <- function(object, ...) {
-  msg <- paste0(
-    "The first argument to [tune_bayes()] should be either ",
-    "a model or workflow."
+  cli::cli_abort(
+    "The first argument to {.fn tune_bayes} should be either a model or workflow,
+    not {.obj_type_friendly {object}}."
   )
-  rlang::abort(msg)
 }
 
 #' @export
@@ -203,10 +202,7 @@ tune_bayes.model_spec <- function(object,
                                   initial = 5,
                                   control = control_bayes()) {
   if (rlang::is_missing(preprocessor) || !is_preprocessor(preprocessor)) {
-    rlang::abort(paste(
-      "To tune a model spec, you must preprocess",
-      "with a formula or recipe"
-    ))
+    cli::cli_abort(tune_pp_msg)
   }
 
   # set `seed` so that calling `control_bayes()` doesn't alter RNG state (#721)
@@ -840,7 +836,7 @@ check_time <- function(origin, limit) {
   }
   now_time <- proc.time()[3]
   if (now_time - origin >= limit * 60) {
-    rlang::abort(paste("The time limit of", limit, "minutes has been reached."))
+    cli::cli_abort("The time limit of {limit} minute{?s} has been reached.")
   }
   invisible(NULL)
 }
diff --git a/R/tune_grid.R b/R/tune_grid.R
index b3bbc250..fe67c6a6 100644
--- a/R/tune_grid.R
+++ b/R/tune_grid.R
@@ -244,11 +244,10 @@ tune_grid <- function(object, ...) {
 
 #' @export
 tune_grid.default <- function(object, ...) {
-  msg <- paste0(
-    "The first argument to [tune_grid()] should be either ",
-    "a model or workflow."
+  cli::cli_abort(
+    "The first argument to {.fn tune_grid} should be either a model or workflow,
+    not {.obj_type_friendly {object}}."
   )
-  rlang::abort(msg)
 }
 
 #' @export
@@ -257,10 +256,7 @@ tune_grid.model_spec <- function(object, preprocessor, resamples, ...,
                                  param_info = NULL, grid = 10, metrics = NULL,
                                  eval_time = NULL, control = control_grid()) {
   if (rlang::is_missing(preprocessor) || !is_preprocessor(preprocessor)) {
-    rlang::abort(paste(
-      "To tune a model spec, you must preprocess",
-      "with a formula or recipe"
-    ))
+    cli::cli_abort(tune_pp_msg)
   }
 
   control <- parsnip::condense_control(control, control_grid())
@@ -298,7 +294,7 @@ tune_grid.workflow <- function(object, resamples, ..., param_info = NULL,
   # Disallow `NULL` grids in `tune_grid()`, as this is the special signal
   # used when no tuning is required
   if (is.null(grid)) {
-    rlang::abort(grid_msg)
+    cli::cli_abort(grid_msg)
   }
 
   res <-
@@ -364,7 +360,10 @@ tune_grid_workflow <- function(workflow,
   )
 
   if (is_cataclysmic(resamples)) {
-    rlang::warn("All models failed. Run `show_notes(.Last.tune.result)` for more information.")
+    cli::cli_warn(
+      "All models failed. Run {.code show_notes(.Last.tune.result)} for more
+       information."
+    )
   }
 
   outcomes <- reduce_all_outcome_names(resamples)
diff --git a/R/tune_results.R b/R/tune_results.R
index da0bb114..caa74973 100644
--- a/R/tune_results.R
+++ b/R/tune_results.R
@@ -126,13 +126,15 @@ is_tune_results <- function(x) {
 
 peek_tune_results_outcomes <- function(x) {
   if (!is_tune_results(x)) {
-    rlang::abort("Internal error: `outcomes` can only be extracted from 'tune_results'.")
+    cli::cli_abort("Internal error: {.arg outcomes} can only be extracted from
+                   {.cls tune_results}.")
   }
 
   out <- attr(x, "outcomes", exact = TRUE)
 
   if (is.null(out)) {
-    rlang::abort("'tune_results' object doesn't have an 'outcomes' attribute.")
+    cli::cli_abort("The object of type {.cls tune_results} doesn't have an
+                    {.code outcomes} attribute.")
   }
 
   out
diff --git a/R/utils.R b/R/utils.R
index 28404c40..c94e70b2 100644
--- a/R/utils.R
+++ b/R/utils.R
@@ -8,9 +8,24 @@
 empty_ellipses <- function(...) {
   dots <- rlang::enquos(...)
   if (length(dots) > 0) {
-    msg <- "The `...` are not used in this function but one or more objects were passed: "
-    msg <- paste0(msg, paste0("'", names(dots), "'", collapse = ", "))
-    rlang::warn(msg)
+    nms <- names(dots)
+    no_name <- nms == ""
+    if (!any(no_name)) {
+      cli::cli_warn(
+        "The {.code ...} are not used in this function but {length(dots)}
+         object{?s} {?was/were} passed: {.val {names(dots)}}"
+      )
+    } else if (all(no_name)) {
+      cli::cli_warn(
+        "The {.code ...} are not used in this function but {length(dots)}
+         unnamed object{?s} {?was/were} passed."
+      )
+    } else {
+      cli::cli_warn(
+        "The {.code ...} are not used in this function but {length(dots)}
+         object{?s} {?was/were} passed."
+      )
+    }
   }
   invisible(NULL)
 }
diff --git a/tests/testthat/_snaps/acquisition.md b/tests/testthat/_snaps/acquisition.md
index e476ac2f..15616a6a 100644
--- a/tests/testthat/_snaps/acquisition.md
+++ b/tests/testthat/_snaps/acquisition.md
@@ -12,7 +12,7 @@
       conf_bound(function() 1)
     Condition
       Error in `conf_bound()`:
-      ! The `trade_off` function should have at least one argument.
+      ! The `trade_off()` function should have at least one argument.
 
 ---
 
@@ -36,7 +36,7 @@
       prob_improve(function() 1)
     Condition
       Error in `prob_improve()`:
-      ! The `trade_off` function should have at least one argument.
+      ! The `trade_off()` function should have at least one argument.
 
 ---
 
@@ -76,7 +76,7 @@
       exp_improve(function() 2)
     Condition
       Error in `exp_improve()`:
-      ! The `trade_off` function should have at least one argument.
+      ! The `trade_off()` function should have at least one argument.
 
 ---
 
diff --git a/tests/testthat/_snaps/augment.md b/tests/testthat/_snaps/augment.md
index 36bf4eb2..266bd2f4 100644
--- a/tests/testthat/_snaps/augment.md
+++ b/tests/testthat/_snaps/augment.md
@@ -22,7 +22,7 @@
       augment(fit_1, parameters = list(cost = 3))
     Condition
       Error in `augment()`:
-      ! 'parameters' should be a single row data frame
+      ! `parameters` should be a single row data frame.
 
 ---
 
@@ -30,7 +30,7 @@
       augment(fit_1, parameters = data.frame(cost = 3:4))
     Condition
       Error in `augment()`:
-      ! 'parameters' should be a single row data frame
+      ! `parameters` should be a single row data frame.
 
 ---
 
diff --git a/tests/testthat/_snaps/autoplot.md b/tests/testthat/_snaps/autoplot.md
index 76ce0bf4..4994fe66 100644
--- a/tests/testthat/_snaps/autoplot.md
+++ b/tests/testthat/_snaps/autoplot.md
@@ -4,7 +4,7 @@
       autoplot(knn_results, type = "performance")
     Condition
       Error in `autoplot()`:
-      ! `type = performance` is only used with iterative search results.
+      ! `type = "performance"` is only used with iterative search results.
 
 ---
 
@@ -12,7 +12,7 @@
       autoplot(knn_results, type = "parameters")
     Condition
       Error in `autoplot()`:
-      ! `type = parameters` is only used with iterative search results.
+      ! `type = "parameters"` is only used with iterative search results.
 
 # coord_obs_pred
 
diff --git a/tests/testthat/_snaps/bayes.md b/tests/testthat/_snaps/bayes.md
index c9fefcdc..700566fe 100644
--- a/tests/testthat/_snaps/bayes.md
+++ b/tests/testthat/_snaps/bayes.md
@@ -432,7 +432,7 @@
       param_info = extract_parameter_set_dials(rec_tune_1), iter = iter1, initial = iter2)
     Condition
       Error in `tune_bayes()`:
-      ! The first argument to [tune_bayes()] should be either a model or workflow.
+      ! The first argument to `tune_bayes()` should be either a model or workflow, not a <recipe> object.
 
 # argument order gives an error for formula
 
@@ -441,7 +441,7 @@
       param_info = extract_parameter_set_dials(svm_mod), initial = iter1, iter = iter2)
     Condition
       Error in `tune_bayes()`:
-      ! The first argument to [tune_bayes()] should be either a model or workflow.
+      ! The first argument to `tune_bayes()` should be either a model or workflow, not a <formula> object.
 
 # retain extra attributes and saved GP candidates
 
@@ -465,7 +465,7 @@
       tune:::check_bayes_initial_size(5, 3, FALSE)
     Message
       ! There are 5 tuning parameters and 3 grid points were requested.
-      * There are more tuning parameters than there are initial points. This is
+      * There are as many tuning parameters as there are initial points. This is
         likely to cause numerical issues in the first few search iterations.
 
 ---
@@ -474,7 +474,7 @@
       tune:::check_bayes_initial_size(5, 3, TRUE)
     Message
       ! There are 5 tuning parameters and 3 grid points were requested.
-      * There are more tuning parameters than there are initial points. This is
+      * There are as many tuning parameters as there are initial points. This is
         likely to cause numerical issues in the first few search iterations.
       * With racing, only completely resampled parameters are used.
 
@@ -492,28 +492,28 @@
     Code
       tune:::check_bayes_initial_size(5, 1, FALSE)
     Condition
-      Error:
+      Error in `tune:::check_bayes_initial_size()`:
       ! There are 5 tuning parameters and 1 grid point was requested.
-      * The GP model requires 2+ initial points. For best performance, supply more initial points than there are tuning parameters.
+      i The GP model requires 2+ initial points. For best performance, supply more initial points than there are tuning parameters.
 
 ---
 
     Code
       tune:::check_bayes_initial_size(5, 1, TRUE)
     Condition
-      Error:
+      Error in `tune:::check_bayes_initial_size()`:
       ! There are 5 tuning parameters and 1 grid point was requested.
-      * The GP model requires 2+ initial points. For best performance, supply more initial points than there are tuning parameters.
-      * With racing, only completely resampled parameters are used.
+      i The GP model requires 2+ initial points. For best performance, supply more initial points than there are tuning parameters.
+      i With racing, only completely resampled parameters are used.
 
 ---
 
     Code
       tune:::check_bayes_initial_size(1, 1, FALSE)
     Condition
-      Error:
+      Error in `tune:::check_bayes_initial_size()`:
       ! There is 1 tuning parameter and 1 grid point was requested.
-      * The GP model requires 2+ initial points. For best performance, supply more initial points than there are tuning parameters.
+      i The GP model requires 2+ initial points. For best performance, supply more initial points than there are tuning parameters.
 
 # missing performance values
 
diff --git a/tests/testthat/_snaps/case-weights.md b/tests/testthat/_snaps/case-weights.md
index 3fc5881e..616b4d68 100644
--- a/tests/testthat/_snaps/case-weights.md
+++ b/tests/testthat/_snaps/case-weights.md
@@ -4,9 +4,9 @@
       .use_case_weights_with_yardstick(1)
     Condition
       Error in `.use_case_weights_with_yardstick()`:
-      ! Unknown case weights object with class <numeric>. 
+      ! An object with <hardhat_case_weights> was expected, not a number.
       i Define a `.use_case_weights_with_yardstick()` method for this type to declare whether or not these case weights should be passed on to yardstick.
-      i See `?.use_case_weights_with_yardstick` for more information.
+      i See `?.use_case_weights_with_yardstick()` for more information.
 
 # `extract_case_weights()` errors if `col` doesn't exist
 
diff --git a/tests/testthat/_snaps/censored-reg.md b/tests/testthat/_snaps/censored-reg.md
index b6f546da..53a43be3 100644
--- a/tests/testthat/_snaps/censored-reg.md
+++ b/tests/testthat/_snaps/censored-reg.md
@@ -23,7 +23,8 @@
       Warning in `tune_grid()`:
       `eval_time` is only used for models with mode "censored regression".
       Warning:
-      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning. Did you want to [tune()] parameters?
+      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning.
+      Did you want to assign any parameters with a value of `tune()`?
     Output
       # Tuning results
       # 10-fold cross-validation using stratification 
diff --git a/tests/testthat/_snaps/checks.md b/tests/testthat/_snaps/checks.md
index d98ea891..2eea7eb3 100644
--- a/tests/testthat/_snaps/checks.md
+++ b/tests/testthat/_snaps/checks.md
@@ -58,7 +58,7 @@
       tune:::check_grid(grid, workflow)
     Condition
       Error in `tune:::check_grid()`:
-      ! The provided `grid` has the following parameter columns that have not been marked for tuning by `tune()`: 'other1', 'other2'.
+      ! The provided grid has 2 parameter columns (`other1` and `other2`) that have not been marked for tuning by `tune()`.
 
 # Missing required `grid` columns are caught
 
@@ -66,7 +66,7 @@
       tune:::check_grid(grid, workflow)
     Condition
       Error in `tune:::check_grid()`:
-      ! The provided `grid` is missing the following parameter columns that have been marked for tuning by `tune()`: 'deg_free'.
+      ! The provided grid is missing the following 1 parameter column that has been marked for tuning by `tune()`: "deg_free".
 
 # workflow objects
 
@@ -74,7 +74,7 @@
       tune:::check_workflow(x = wflow_2, check_dials = TRUE)
     Condition
       Error in `tune:::check_workflow()`:
-      ! The workflow has arguments whose ranges are not finalized: 'mtry'
+      ! The workflow has arguments whose ranges are not finalized: "mtry".
 
 ---
 
@@ -132,7 +132,7 @@
       tune:::check_metrics(yardstick::rmse, chi_wflow)
     Condition
       Error in `tune:::check_metrics()`:
-      ! The `metrics` argument should be the results of [yardstick::metric_set()].
+      ! The `metrics` argument should be the results of `yardstick::metric_set()`.
 
 # metrics must match the parsnip engine
 
@@ -302,7 +302,7 @@
         yardstick::rsq), ctrl = control_bayes())
     Condition
       Error in `tune:::check_initial()`:
-      ! `initial` should be a positive integer or the results of [tune_grid()]
+      ! `initial` should be a positive integer or the results of `tune_grid()`
 
 # check parameter finalization
 
@@ -310,21 +310,21 @@
       expect_no_error(p1 <- tune:::check_parameters(w1, data = mtcars, grid_names = character(
         0)))
     Message
-      i Creating pre-processing data to finalize unknown parameter: mtry
+      i Creating pre-processing data to finalize 1 unknown parameter: "mtry"
 
 ---
 
     Code
       expect_no_error(p2 <- tune:::check_parameters(w2, data = mtcars))
     Message
-      i Creating pre-processing data to finalize unknown parameter: mtry
+      i Creating pre-processing data to finalize 1 unknown parameter: "mtry"
 
 ---
 
     Code
       expect_no_error(p3_a <- tune:::check_parameters(w3, data = mtcars))
     Message
-      i Creating pre-processing data to finalize unknown parameter: mtry
+      i Creating pre-processing data to finalize 1 unknown parameter: "mtry"
 
 ---
 
@@ -332,5 +332,6 @@
       tune:::check_parameters(w4, data = mtcars)
     Condition
       Error in `tune:::check_parameters()`:
-      ! Some model parameters require finalization but there are recipe parameters that require tuning. Please use  `extract_parameter_set_dials()` to set parameter ranges  manually and supply the output to the `param_info` argument.
+      ! Some model parameters require finalization but there are recipe parameters that require tuning.
+      i Please use `extract_parameter_set_dials()` to set parameter ranges manually and supply the output to the `param_info` argument.
 
diff --git a/tests/testthat/_snaps/collect.md b/tests/testthat/_snaps/collect.md
index 2976215c..a8e051db 100644
--- a/tests/testthat/_snaps/collect.md
+++ b/tests/testthat/_snaps/collect.md
@@ -4,7 +4,7 @@
       collect_predictions(lm_splines %>% dplyr::select(-.predictions))
     Condition
       Error in `collect_predictions()`:
-      ! The `.predictions` column does not exist. Refit with the control argument `save_pred = TRUE` to save predictions.
+      ! The .predictions column does not exist. Please refit with the control argument `save_pred = TRUE` to save predictions.
 
 # `collect_predictions()` errors informatively applied to unsupported class
 
@@ -20,7 +20,7 @@
       collect_predictions(svm_tune, parameters = tibble(wrong = "value"))
     Condition
       Error in `filter_predictions()`:
-      ! `parameters` should only have columns: 'cost value'
+      ! `parameters` should only have columns: "cost value".
 
 # collecting notes - fit_resamples
 
diff --git a/tests/testthat/_snaps/compute_metrics.md b/tests/testthat/_snaps/compute_metrics.md
index de2bac52..3560f458 100644
--- a/tests/testthat/_snaps/compute_metrics.md
+++ b/tests/testthat/_snaps/compute_metrics.md
@@ -30,7 +30,7 @@
       compute_metrics("boop", metric_set(rsq))
     Condition
       Error in `compute_metrics()`:
-      ! No `compute_metrics()` method exists for this type of object.
+      ! No `compute_metrics()` method exists for a string.
 
 ---
 
diff --git a/tests/testthat/_snaps/conf-mat-resampled.md b/tests/testthat/_snaps/conf-mat-resampled.md
index 1ac63c0e..9b194e86 100644
--- a/tests/testthat/_snaps/conf-mat-resampled.md
+++ b/tests/testthat/_snaps/conf-mat-resampled.md
@@ -4,7 +4,7 @@
       conf_mat_resampled(svm_results)
     Condition
       Error in `conf_mat_resampled()`:
-      ! It looks like there are 5 tuning parameter combination(s) in the data. Please use the `parameters` argument to select one combination of parameters.
+      ! It looks like there are 5 tuning parameter combination in the data. Please use the `parameters` argument to select one combination of parameters.
 
 ---
 
@@ -28,7 +28,7 @@
       conf_mat_resampled(tibble::as_tibble(svm_results))
     Condition
       Error in `conf_mat_resampled()`:
-      ! The first argument needs to be an object with class 'tune_results'.
+      ! The first argument needs to be <tune_results> object, not a data frame.
 
 ---
 
diff --git a/tests/testthat/_snaps/engine-parameters.md b/tests/testthat/_snaps/engine-parameters.md
index 76194a9f..09f88212 100644
--- a/tests/testthat/_snaps/engine-parameters.md
+++ b/tests/testthat/_snaps/engine-parameters.md
@@ -4,7 +4,7 @@
       rf_tune <- rf_mod %>% tune_grid(mpg ~ ., resamples = rs, grid = 3)
     Condition
       Error in `check_param_objects()`:
-      ! The workflow has arguments to be tuned that are missing some parameter objects: 'corr.bias'
+      ! The workflow has arguments to be tuned that are missing some parameter objects: "corr.bias"
 
 ---
 
@@ -12,7 +12,7 @@
       p <- autoplot(rf_tune)
     Condition
       Error in `autoplot()`:
-      ! Some parameters do not have corresponding parameter objects and cannot be used with `autoplot()`: 'corr.bias'
+      ! Some parameters do not have corresponding parameter objects and cannot be used with `autoplot()`: `corr.bias`.
 
 ---
 
@@ -20,5 +20,5 @@
       rf_search <- rf_mod %>% tune_bayes(mpg ~ ., resamples = rs)
     Condition
       Error in `check_param_objects()`:
-      ! The workflow has arguments to be tuned that are missing some parameter objects: 'corr.bias'
+      ! The workflow has arguments to be tuned that are missing some parameter objects: "corr.bias"
 
diff --git a/tests/testthat/_snaps/filter-parameters.md b/tests/testthat/_snaps/filter-parameters.md
index 4001ac96..422b88ea 100644
--- a/tests/testthat/_snaps/filter-parameters.md
+++ b/tests/testthat/_snaps/filter-parameters.md
@@ -5,8 +5,7 @@
         `%^*#` = 1))
     Condition
       Error in `filter_parameters()`:
-      ! collect_metrics should have class 'tune_results'.
-      * svm_reg_results should have class 'tune_results'.
+      ! `collect_metrics(svm_reg_results)` should have class <tune_results>; a tibble was passed.
 
 ---
 
@@ -14,7 +13,7 @@
       filter_parameters(svm_reg_results, parameters = tibble::tibble(soup = 1))
     Condition
       Error in `filter_by_join()`:
-      ! There are no columns in 'parameters' that match with svm_reg_results
+      ! There are no columns in `parameters` that match with "svm_reg_results".
 
 ---
 
diff --git a/tests/testthat/_snaps/finalization.md b/tests/testthat/_snaps/finalization.md
index 52b78da8..38b240ac 100644
--- a/tests/testthat/_snaps/finalization.md
+++ b/tests/testthat/_snaps/finalization.md
@@ -4,5 +4,6 @@
       mod_1 %>% tune_grid(rec_1, resamples = rs, grid = 3)
     Condition
       Error in `check_parameters()`:
-      ! Some model parameters require finalization but there are recipe parameters that require tuning. Please use  `extract_parameter_set_dials()` to set parameter ranges  manually and supply the output to the `param_info` argument.
+      ! Some model parameters require finalization but there are recipe parameters that require tuning.
+      i Please use `extract_parameter_set_dials()` to set parameter ranges manually and supply the output to the `param_info` argument.
 
diff --git a/tests/testthat/_snaps/fit_best.md b/tests/testthat/_snaps/fit_best.md
index 3e8e951e..d9daaf20 100644
--- a/tests/testthat/_snaps/fit_best.md
+++ b/tests/testthat/_snaps/fit_best.md
@@ -42,11 +42,11 @@
 
 ---
 
-    There is no `fit_best()` method for an object with class `integer`.
+    No `fit_best()` exists for an integer.
 
 ---
 
-    There is no `fit_best()` method for an object with classes `tbl_df`, `tbl`, and `data.frame`.
+    No `fit_best()` exists for a tibble.
 
 ---
 
diff --git a/tests/testthat/_snaps/grid.md b/tests/testthat/_snaps/grid.md
index a975cfb4..7369ce99 100644
--- a/tests/testthat/_snaps/grid.md
+++ b/tests/testthat/_snaps/grid.md
@@ -43,7 +43,7 @@
         mtcars, v = 2))
     Condition
       Error in `tune_grid()`:
-      ! The first argument to [tune_grid()] should be either a model or workflow.
+      ! The first argument to `tune_grid()` should be either a model or workflow, not a <recipe> object.
 
 # argument order gives errors for formula
 
@@ -51,7 +51,7 @@
       tune_grid(mpg ~ ., helper_objects$lm_mod, rsample::vfold_cv(mtcars, v = 2))
     Condition
       Error in `tune_grid()`:
-      ! The first argument to [tune_grid()] should be either a model or workflow.
+      ! The first argument to `tune_grid()` should be either a model or workflow, not a <formula> object.
 
 # ellipses with tune_grid
 
@@ -59,7 +59,7 @@
       tune_grid(wflow, resamples = folds, grid = 3, something = "wrong")
     Condition
       Warning:
-      The `...` are not used in this function but one or more objects were passed: 'something'
+      The `...` are not used in this function but 1 object was passed: "something"
     Output
       # Tuning results
       # 10-fold cross-validation 
diff --git a/tests/testthat/_snaps/last-fit.md b/tests/testthat/_snaps/last-fit.md
index d49fc24d..6ea96647 100644
--- a/tests/testthat/_snaps/last-fit.md
+++ b/tests/testthat/_snaps/last-fit.md
@@ -21,7 +21,7 @@
       linear_reg() %>% set_engine("lm") %>% last_fit(f, split, something = "wrong")
     Condition
       Warning:
-      The `...` are not used in this function but one or more objects were passed: 'something'
+      The `...` are not used in this function but 1 object was passed: "something"
     Output
       # Resampling results
       # Manual resampling 
@@ -36,7 +36,7 @@
       last_fit(rec, lin_mod, split)
     Condition
       Error in `last_fit()`:
-      ! The first argument to [last_fit()] should be either a model or workflow.
+      ! The first argument to `last_fit()` should be either a model or workflow, not a <recipe> object.
 
 ---
 
@@ -44,20 +44,20 @@
       last_fit(f, lin_mod, split)
     Condition
       Error in `last_fit()`:
-      ! The first argument to [last_fit()] should be either a model or workflow.
+      ! The first argument to `last_fit()` should be either a model or workflow, not a <formula> object.
 
 # `last_fit()` when objects need tuning
 
-    2 arguments have been tagged for tuning in these components: model_spec and recipe. 
-    Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
+    2 arguments have been tagged for tuning in these components: model_spec and recipe.
+    i Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
 
 ---
 
-    1 argument has been tagged for tuning in this component: model_spec. 
-    Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
+    1 argument has been tagged for tuning in this component: model_spec.
+    i Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
 
 ---
 
-    1 argument has been tagged for tuning in this component: recipe. 
-    Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
+    1 argument has been tagged for tuning in this component: recipe.
+    i Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
 
diff --git a/tests/testthat/_snaps/misc.md b/tests/testthat/_snaps/misc.md
index 0ab3b02f..40edfd02 100644
--- a/tests/testthat/_snaps/misc.md
+++ b/tests/testthat/_snaps/misc.md
@@ -4,5 +4,5 @@
       tune:::empty_ellipses(a = 1)
     Condition
       Warning:
-      The `...` are not used in this function but one or more objects were passed: 'a'
+      The `...` are not used in this function but 1 object was passed: "a"
 
diff --git a/tests/testthat/_snaps/resample.md b/tests/testthat/_snaps/resample.md
index 9fc0e154..835082c9 100644
--- a/tests/testthat/_snaps/resample.md
+++ b/tests/testthat/_snaps/resample.md
@@ -74,7 +74,8 @@
       result <- tune_grid(lin_mod, mpg ~ ., folds)
     Condition
       Warning:
-      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning. Did you want to [tune()] parameters?
+      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning.
+      Did you want to assign any parameters with a value of `tune()`?
 
 # `tune_grid()` falls back to `fit_resamples()` - workflow variables
 
@@ -82,7 +83,8 @@
       result <- tune_grid(wf, folds)
     Condition
       Warning:
-      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning. Did you want to [tune()] parameters?
+      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning.
+      Did you want to assign any parameters with a value of `tune()`?
 
 # `tune_grid()` ignores `grid` if there are no tuning parameters
 
@@ -90,7 +92,8 @@
       result <- lin_mod %>% tune_grid(mpg ~ ., grid = data.frame(x = 1), folds)
     Condition
       Warning:
-      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning. Did you want to [tune()] parameters?
+      No tuning parameters have been detected, performance will be evaluated using the resamples with no tuning.
+      Did you want to assign any parameters with a value of `tune()`?
 
 # cannot autoplot `fit_resamples()` results
 
@@ -98,7 +101,7 @@
       autoplot(result)
     Condition
       Error in `autoplot()`:
-      ! There is no `autoplot()` implementation for `resample_results`.
+      ! There is no `autoplot()` implementation for <resample_results>.
 
 # ellipses with fit_resamples
 
@@ -106,7 +109,7 @@
       lin_mod %>% fit_resamples(mpg ~ ., folds, something = "wrong")
     Condition
       Warning:
-      The `...` are not used in this function but one or more objects were passed: 'something'
+      The `...` are not used in this function but 1 object was passed: "something"
     Output
       # Resampling results
       # 2-fold cross-validation 
@@ -122,7 +125,7 @@
       fit_resamples(rec, lin_mod, folds)
     Condition
       Error in `fit_resamples()`:
-      ! The first argument to [fit_resamples()] should be either a model or workflow.
+      ! The first argument to `fit_resamples()` should be either a model or workflow, not a <recipe> object.
 
 ---
 
@@ -130,7 +133,7 @@
       fit_resamples(mpg ~ ., lin_mod, folds)
     Condition
       Error in `fit_resamples()`:
-      ! The first argument to [fit_resamples()] should be either a model or workflow.
+      ! The first argument to `fit_resamples()` should be either a model or workflow, not a <formula> object.
 
 # retain extra attributes
 
@@ -152,16 +155,16 @@
 
 # `fit_resamples()` when objects need tuning
 
-    2 arguments have been tagged for tuning in these components: model_spec and recipe. 
-    Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
+    2 arguments have been tagged for tuning in these components: model_spec and recipe.
+    i Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
 
 ---
 
-    1 argument has been tagged for tuning in this component: model_spec. 
-    Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
+    1 argument has been tagged for tuning in this component: model_spec.
+    i Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
 
 ---
 
-    1 argument has been tagged for tuning in this component: recipe. 
-    Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
+    1 argument has been tagged for tuning in this component: recipe.
+    i Please use one of the tuning functions (e.g. `tune_grid()`) to optimize them.
 
diff --git a/tests/testthat/_snaps/utils.md b/tests/testthat/_snaps/utils.md
new file mode 100644
index 00000000..921251a3
--- /dev/null
+++ b/tests/testthat/_snaps/utils.md
@@ -0,0 +1,40 @@
+# check for no arguments pass to ...
+
+    Code
+      empty_ellipses(a = 5)
+    Condition
+      Warning:
+      The `...` are not used in this function but 1 object was passed: "a"
+
+---
+
+    Code
+      empty_ellipses(a = 5, b = 1)
+    Condition
+      Warning:
+      The `...` are not used in this function but 2 objects were passed: "a" and "b"
+
+---
+
+    Code
+      empty_ellipses(5)
+    Condition
+      Warning:
+      The `...` are not used in this function but 1 unnamed object was passed.
+
+---
+
+    Code
+      empty_ellipses(1, 5)
+    Condition
+      Warning:
+      The `...` are not used in this function but 2 unnamed objects were passed.
+
+---
+
+    Code
+      empty_ellipses(1, second = 2)
+    Condition
+      Warning:
+      The `...` are not used in this function but 2 objects were passed.
+
diff --git a/tests/testthat/test-utils.R b/tests/testthat/test-utils.R
new file mode 100644
index 00000000..17e00c74
--- /dev/null
+++ b/tests/testthat/test-utils.R
@@ -0,0 +1,18 @@
+test_that("check for no arguments pass to ...", {
+
+  expect_snapshot(
+    empty_ellipses(a = 5)
+  )
+  expect_snapshot(
+    empty_ellipses(a = 5, b = 1)
+  )
+  expect_snapshot(
+    empty_ellipses(5)
+  )
+  expect_snapshot(
+    empty_ellipses(1, 5)
+  )
+  expect_snapshot(
+    empty_ellipses(1, second = 2)
+  )
+})