diff --git a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.spec.ts b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.spec.ts
index 720e165b7609d2..2866d3702eeea3 100644
--- a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.spec.ts
+++ b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.spec.ts
@@ -1,6 +1,7 @@
 import {
   detectAIContentType,
   parseXmlTagSegments,
+  preprocessInlineXmlTags,
   tryParsePythonDict,
 } from 'sentry/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection';
 
@@ -127,6 +128,41 @@ describe('tryParsePythonDict', () => {
   });
 });
 
+describe('preprocessInlineXmlTags', () => {
+  it('replaces inline XML tags with italic markdown', () => {
+    const text = 'Before <thinking>inner thought</thinking> After';
+    expect(preprocessInlineXmlTags(text)).toBe('Before *thinking: inner thought* After');
+  });
+
+  it('leaves block-level tags at start of text untouched', () => {
+    const text = '<plan>step 1</plan> then more';
+    expect(preprocessInlineXmlTags(text)).toBe('<plan>step 1</plan> then more');
+  });
+
+  it('leaves block-level tags after newline untouched', () => {
+    const text = 'Some text\n<thinking>deep thought</thinking>';
+    expect(preprocessInlineXmlTags(text)).toBe(text);
+  });
+
+  it('handles mixed inline and block tags', () => {
+    const text =
+      'See the <code>details</code> here.\n<thinking>\ndeep thought\n</thinking>';
+    expect(preprocessInlineXmlTags(text)).toBe(
+      'See the *code: details* here.\n<thinking>\ndeep thought\n</thinking>'
+    );
+  });
+
+  it('trims whitespace from inline tag content', () => {
+    const text = 'before <tag>  spaced  </tag> after';
+    expect(preprocessInlineXmlTags(text)).toBe('before *tag: spaced* after');
+  });
+
+  it('strips nested XML tags from inline tag content', () => {
+    const text = 'Text <outer>before <inner>nested</inner> after</outer> more';
+    expect(preprocessInlineXmlTags(text)).toBe('Text *outer: before nested after* more');
+  });
+});
+
 describe('parseXmlTagSegments', () => {
   it('splits text with XML tags into segments', () => {
     const text = 'Before <thinking>inner thought</thinking> After';
@@ -157,9 +193,9 @@ describe('parseXmlTagSegments', () => {
   });
 
   it('returns single text segment when no XML tags', () => {
-    const text = 'just plain text';
-    const segments = parseXmlTagSegments(text);
-    expect(segments).toEqual([{type: 'text', content: 'just plain text'}]);
+    expect(parseXmlTagSegments('just plain text')).toEqual([
+      {type: 'text', content: 'just plain text'},
+    ]);
   });
 
   it('handles empty string', () => {
@@ -168,7 +204,20 @@ describe('parseXmlTagSegments', () => {
 
   it('handles tags with hyphens in names', () => {
     const text = '<my-tag>content</my-tag>';
-    const segments = parseXmlTagSegments(text);
-    expect(segments).toEqual([{type: 'xml-tag', tagName: 'my-tag', content: 'content'}]);
+    expect(parseXmlTagSegments(text)).toEqual([
+      {type: 'xml-tag', tagName: 'my-tag', content: 'content'},
+    ]);
+  });
+
+  it('extracts outer tag with nested tags preserved in content', () => {
+    const text =
+      '<bug_report>\n<location>file.ts</location>\n<description>a bug</description>\n</bug_report>';
+    expect(parseXmlTagSegments(text)).toEqual([
+      {
+        type: 'xml-tag',
+        tagName: 'bug_report',
+        content: '\n<location>file.ts</location>\n<description>a bug</description>\n',
+      },
+    ]);
   });
 });
diff --git a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.ts b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.ts
index 9e1161b576672d..bb39264a545e9f 100644
--- a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.ts
+++ b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection.ts
@@ -18,10 +18,7 @@ interface AIContentDetectionResult {
   wasFixed?: boolean;
 }
 
-/**
- * Best-effort conversion of a Python dict literal to a JSON-parseable string.
- * Returns the parsed object on success, or null if conversion fails.
- */
+/** Best-effort conversion of a Python dict literal to a JSON-parseable string. */
 export function tryParsePythonDict(text: string): Record<PropertyKey, unknown> | null {
   const trimmed = text.trim();
   if (!trimmed.startsWith('{') && !trimmed.startsWith('[')) {
@@ -50,10 +47,7 @@ export function tryParsePythonDict(text: string): Record<PropertyKey, unknown> |
   }
 }
 
-/**
- * Splits text into segments of plain text and XML-like tag blocks.
- * Matches `<tagname>...</tagname>` patterns (including multiline content).
- */
+/** Splits text into segments of plain text and XML-like tag blocks. */
 export function parseXmlTagSegments(text: string): ContentSegment[] {
   const segments: ContentSegment[] = [];
   const xmlTagRegex = /<([a-zA-Z][\w-]*)>([\s\S]*?)<\/\1>/g;
@@ -78,6 +72,19 @@ export function parseXmlTagSegments(text: string): ContentSegment[] {
   return segments;
 }
 
+/** Replaces inline XML tags with italic markdown, leaves block-level tags untouched. */
+export function preprocessInlineXmlTags(text: string): string {
+  const xmlTagRegex = /<([a-zA-Z][\w-]*)>([\s\S]*?)<\/\1>/g;
+  return text.replace(xmlTagRegex, (match, tagName, content, offset) => {
+    const isBlock = offset === 0 || /\n\s*$/.test(text.slice(0, offset));
+    if (isBlock) {
+      return match;
+    }
+    const stripped = content.replace(/<\/?[a-zA-Z][\w-]*>/g, '').trim();
+    return `*${tagName}: ${stripped}*`;
+  });
+}
+
 const XML_TAG_REGEX = /<([a-zA-Z][\w-]*)>[\s\S]*?<\/\1>/;
 
 const MARKDOWN_INDICATORS = [
@@ -91,16 +98,7 @@ const MARKDOWN_INDICATORS = [
   /^```/m, // code fences
 ];
 
-/**
- * Detects the content type of an AI response string.
- * Short-circuits on first match in priority order:
- * 1. Valid JSON (object/array)
- * 2. Python dict
- * 3. Partial/truncated JSON (fixable)
- * 4. Markdown with XML tags
- * 5. Markdown
- * 6. Plain text
- */
+/** Detects AI content type: JSON, Python dict, markdown-with-xml, markdown, or plain text. */
 export function detectAIContentType(text: string): AIContentDetectionResult {
   const trimmed = text.trim();
   if (!trimmed) {
diff --git a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.spec.tsx b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.spec.tsx
index 561f2389c8037d..82118cef34efb8 100644
--- a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.spec.tsx
+++ b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.spec.tsx
@@ -28,16 +28,32 @@ describe('AIContentRenderer', () => {
     expect(screen.getByText('name')).toBeInTheDocument();
   });
 
-  it('renders XML tags with styled wrappers', () => {
+  it('renders inline XML tags as italic text within the flow', () => {
     render(<AIContentRenderer text="Before <thinking>inner thought</thinking> After" />);
+    expect(screen.getByText(/thinking: inner thought/)).toBeInTheDocument();
+  });
+
+  it('renders block XML tags with styled wrappers', () => {
+    render(<AIContentRenderer text={'Text\n<thinking>inner thought</thinking>'} />);
     expect(screen.getByText('thinking')).toBeInTheDocument();
   });
 
-  it('renders XML tags with styled wrappers when inline', () => {
+  it('renders inline XML tags as italic text when inline', () => {
     render(
       <AIContentRenderer text="Before <thinking>inner thought</thinking> After" inline />
     );
-    expect(screen.getByText('thinking')).toBeInTheDocument();
+    expect(screen.getByText(/thinking: inner thought/)).toBeInTheDocument();
+  });
+
+  it('renders nested XML tags recursively', () => {
+    const text =
+      '<bug_report>\n<location>file.ts</location>\n<description>a bug</description>\n</bug_report>';
+    render(<AIContentRenderer text={text} />);
+    expect(screen.getByText('bug_report')).toBeInTheDocument();
+    expect(screen.getByText('location')).toBeInTheDocument();
+    expect(screen.getByText('description')).toBeInTheDocument();
+    expect(screen.getByText('file.ts')).toBeInTheDocument();
+    expect(screen.getByText('a bug')).toBeInTheDocument();
   });
 
   it('wraps plain text in MultilineText by default', () => {
diff --git a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.tsx b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.tsx
index 3b5e501a382ed3..b591d91e92760f 100644
--- a/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.tsx
+++ b/static/app/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentRenderer.tsx
@@ -9,12 +9,12 @@ import {MarkedText} from 'sentry/utils/marked/markedText';
 import {
   detectAIContentType,
   parseXmlTagSegments,
+  preprocessInlineXmlTags,
 } from 'sentry/views/performance/newTraceDetails/traceDrawer/details/span/eapSections/aiContentDetection';
 import {TraceDrawerComponents} from 'sentry/views/performance/newTraceDetails/traceDrawer/details/styles';
 
 interface AIContentRendererProps {
   text: string;
-  /** When true, renders content directly without a wrapper or raw/pretty toggle. */
   inline?: boolean;
   maxJsonDepth?: number;
 }
@@ -27,22 +27,23 @@ function XmlTagBlock({tagName, content}: {content: string; tagName: string}) {
       direction="column"
       padding="0 0 0 md"
       margin="sm 0"
-      style={{borderLeft: `3px solid ${theme.tokens.border.accent.moderate}`}}
+      style={{borderLeft: `2px solid ${theme.tokens.border.primary}`}}
     >
       <Container margin="0 0 xs 0">
         <Text size="xs" variant="muted">
           {tagName}
         </Text>
       </Container>
-      <Text italic>
-        <MarkedText as={TraceDrawerComponents.MarkdownContainer} text={content} />
-      </Text>
+      <MarkdownWithXmlRenderer text={content} />
     </Flex>
   );
 }
 
 function MarkdownWithXmlRenderer({text}: {text: string}) {
-  const segments = useMemo(() => parseXmlTagSegments(text), [text]);
+  const segments = useMemo(
+    () => parseXmlTagSegments(preprocessInlineXmlTags(text)),
+    [text]
+  );
 
   return (
     <Fragment>
@@ -61,10 +62,7 @@ function MarkdownWithXmlRenderer({text}: {text: string}) {
   );
 }
 
-/**
- * Unified AI content renderer that auto-detects content type and renders appropriately.
- * Handles JSON, Python dicts, partial JSON, markdown with XML tags, markdown, and plain text.
- */
+/** Auto-detects AI content type and renders appropriately. */
 export function AIContentRenderer({
   text,
   inline = false,