diff --git a/python-package/querychat/prompt/prompt.md b/python-package/querychat/prompt/prompt.md
index 62d1ea17f..fcb00a5ed 100644
--- a/python-package/querychat/prompt/prompt.md
+++ b/python-package/querychat/prompt/prompt.md
@@ -74,6 +74,16 @@ Example of question answering:
 
 If the user provides a vague help request, like "Help" or "Show me instructions", describe your own capabilities in a helpful way, including examples of questions they can ask. Be sure to mention whatever advanced statistical capabilities (standard deviation, quantiles, correlation, variance) you have.
 
+### Showing example questions
+
+If you find yourself offering example questions to the user as part of your response, wrap the text of each prompt in `<span class="suggestion">` tags. For example:
+
+```
+* <span class="suggestion">Suggestion 1.</span>
+* <span class="suggestion">Suggestion 2.</span>
+* <span class="suggestion">Suggestion 3.</span>
+```
+
 ## DuckDB SQL tips
 
 * `percentile_cont` and `percentile_disc` are "ordered set" aggregate functions. These functions are specified using the WITHIN GROUP (ORDER BY sort_expression) syntax, and they are converted to an equivalent aggregate function that takes the ordering expression as the first argument. For example, `percentile_cont(fraction) WITHIN GROUP (ORDER BY column [(ASC|DESC)])` is equivalent to `quantile_cont(column, fraction ORDER BY column [(ASC|DESC)])`.
diff --git a/python-package/querychat/querychat.py b/python-package/querychat/querychat.py
index 4e492fb19..58bb56071 100644
--- a/python-package/querychat/querychat.py
+++ b/python-package/querychat/querychat.py
@@ -136,7 +136,7 @@ def df_to_html(df: IntoFrame, maxrows: int = 5) -> str:
     df_short = nw.from_native(df).head(maxrows)
 
     # Generate HTML table
-    table_html = df_short.to_pandas().to_html(index=False)
+    table_html = df_short.to_pandas().to_html(index=False, classes="table table-striped")
 
     # Add note about truncated rows if needed
     if len(df_short) != len(ndf):
diff --git a/python-package/querychat/static/css/styles.css b/python-package/querychat/static/css/styles.css
index 7596748f4..033ce722c 100644
--- a/python-package/querychat/static/css/styles.css
+++ b/python-package/querychat/static/css/styles.css
@@ -7,7 +7,6 @@
   flex-direction: column;
   height: 100%;
   width: 100%;
-  background-color: var(--bslib-sidebar-main-bg);
 }
 
 .querychat-messages {
diff --git a/r-package/inst/prompt/prompt.md b/r-package/inst/prompt/prompt.md
index 154ce0cc8..9ed80f43e 100644
--- a/r-package/inst/prompt/prompt.md
+++ b/r-package/inst/prompt/prompt.md
@@ -80,6 +80,16 @@ Example of question answering:
 
 If the user provides a vague help request, like "Help" or "Show me instructions", describe your own capabilities in a helpful way, including examples of questions they can ask. Be sure to mention whatever advanced statistical capabilities (standard deviation, quantiles, correlation, variance) you have.
 
+### Showing example questions
+
+If you find yourself offering example questions to the user as part of your response, wrap the text of each prompt in `<span class="suggestion">` tags. For example:
+
+```
+* <span class="suggestion">Suggestion 1.</span>
+* <span class="suggestion">Suggestion 2.</span>
+* <span class="suggestion">Suggestion 3.</span>
+```
+
 ## DuckDB SQL tips
 
 * `percentile_cont` and `percentile_disc` are "ordered set" aggregate functions. These functions are specified using the WITHIN GROUP (ORDER BY sort_expression) syntax, and they are converted to an equivalent aggregate function that takes the ordering expression as the first argument. For example, `percentile_cont(fraction) WITHIN GROUP (ORDER BY column [(ASC|DESC)])` is equivalent to `quantile_cont(column, fraction ORDER BY column [(ASC|DESC)])`.