diff --git a/automation/workflow.py b/automation/workflow.py
index 5a695cf..b186a3a 100644
--- a/automation/workflow.py
+++ b/automation/workflow.py
@@ -33,12 +33,25 @@ DISCUSSION_SUFFIXES = (
 SUMMARY_SUFFIX = ".sum.md"
 MENTION_PATTERN = re.compile(r"@(\w+|all)")
 
+# Patterns to extract structured markers (support both plain and markdown bold)
+# Matches: DECISION: text, **DECISION**: text, decision: text
+DECISION_PATTERN = re.compile(r'(?:\*\*)?DECISION(?:\*\*)?\s*:\s*(.+?)(?=\s*(?:\*\*QUESTION|\*\*ACTION|\*\*SUGGESTION|VOTE:)|$)', re.IGNORECASE | re.DOTALL)
+QUESTION_PATTERN = re.compile(r'(?:\*\*)?(?:QUESTION|Q)(?:\*\*)?\s*:\s*(.+?)(?=\s*(?:\*\*DECISION|\*\*ACTION|\*\*SUGGESTION|VOTE:)|$)', re.IGNORECASE | re.DOTALL)
+ACTION_PATTERN = re.compile(r'(?:\*\*)?(?:ACTION|TODO)(?:\*\*)?\s*:\s*(.+?)(?=\s*(?:\*\*DECISION|\*\*QUESTION|\*\*SUGGESTION|VOTE:)|$)', re.IGNORECASE | re.DOTALL)
+
 
 def extract_structured_basic(text: str) -> dict[str, list]:
     """
-    Derive structured discussion signals using lightweight pattern matching.
+    Derive structured discussion signals using regex pattern matching.
 
-    Recognises explicit markers (Q:, TODO:, DONE:, DECISION:) and @mentions.
+    Recognises markers anywhere in text (not just at line start):
+    - DECISION: / **DECISION**: - Architectural/technical decisions
+    - QUESTION: / **QUESTION**: / Q: - Open questions needing answers
+    - ACTION: / **ACTION**: / TODO: - Action items with optional @assignee
+    - @mentions - References to participants
+
+    Also supports legacy line-start markers: ASSIGNED:, DONE:
+    Questions ending with '?' are auto-detected.
     """
     questions: list[dict[str, str]] = []
     action_items: list[dict[str, str]] = []
@@ -66,7 +79,6 @@ def extract_structured_basic(text: str) -> dict[str, list]:
         analysis = remainder.strip() if participant else stripped
         if not analysis:
             continue
-        lowered = analysis.lower()
         participant_name = participant or "unknown"
 
         if timeline_data is None:
@@ -75,30 +87,47 @@ def extract_structured_basic(text: str) -> dict[str, list]:
                 "summary": _truncate_summary(analysis),
             }
 
-        # Questions
-        if lowered.startswith("q:") or lowered.startswith("question:"):
-            _, _, body = analysis.partition(":")
-            question_text = body.strip()
-            if question_text:
-                questions.append(
-                    {"participant": participant_name, "question": question_text, "status": "OPEN"}
+        # Extract decisions using regex (finds markers anywhere in line)
+        for match in DECISION_PATTERN.finditer(analysis):
+            decision_text = match.group(1).strip()
+            if decision_text:
+                decisions.append(
+                    {
+                        "participant": participant_name,
+                        "decision": decision_text,
+                        "rationale": "",
+                        "supporters": [],
+                    }
                 )
-        elif analysis.endswith("?"):
-            question_text = analysis.rstrip("?").strip()
+
+        # Extract questions using regex (finds markers anywhere in line)
+        for match in QUESTION_PATTERN.finditer(analysis):
+            question_text = match.group(1).strip()
             if question_text:
                 questions.append(
                     {"participant": participant_name, "question": question_text, "status": "OPEN"}
                 )
 
-        # Action items
-        if lowered.startswith(("todo:", "action:")):
-            _, _, body = analysis.partition(":")
-            action_text = body.strip()
+        # Also catch questions that end with '?' and don't have explicit marker
+        if '?' in analysis and not QUESTION_PATTERN.search(analysis):
+            # Simple heuristic: if line ends with ?, treat as question
+            if analysis.rstrip().endswith('?'):
+                question_text = analysis.rstrip('?').strip()
+                # Avoid duplicate if already extracted
+                if question_text and not any(q['question'] == question_text for q in questions):
+                    questions.append(
+                        {"participant": participant_name, "question": question_text, "status": "OPEN"}
+                    )
+
+        # Extract action items using regex (finds markers anywhere in line)
+        for match in ACTION_PATTERN.finditer(analysis):
+            action_text = match.group(1).strip()
             if action_text:
+                # Extract assignee from @mention in the line
                 assignee = None
-                match = MENTION_PATTERN.search(line)
-                if match:
-                    assignee = match.group(1)
+                mention_match = MENTION_PATTERN.search(action_text)
+                if mention_match:
+                    assignee = mention_match.group(1)
                 action_items.append(
                     {
                         "participant": participant_name,
@@ -107,15 +136,17 @@ def extract_structured_basic(text: str) -> dict[str, list]:
                         "assignee": assignee,
                     }
                 )
-        elif lowered.startswith("assigned:"):
-            _, _, body = analysis.partition(":")
-            action_text = body.strip()
+
+        # Legacy support for plain text markers at line start
+        lowered = analysis.lower()
+        if lowered.startswith("assigned:"):
+            _, _, action_text = analysis.partition(":")
+            action_text = action_text.strip()
             if action_text:
-                # Extract assignee from @mention in the line
-                assignee = participant_name  # Default to participant claiming it
-                match = MENTION_PATTERN.search(line)
-                if match:
-                    assignee = match.group(1)
+                assignee = participant_name
+                mention_match = MENTION_PATTERN.search(line)
+                if mention_match:
+                    assignee = mention_match.group(1)
                 action_items.append(
                     {
                         "participant": participant_name,
@@ -125,8 +156,8 @@ def extract_structured_basic(text: str) -> dict[str, list]:
                     }
                 )
         elif lowered.startswith("done:"):
-            _, _, body = analysis.partition(":")
-            action_text = body.strip()
+            _, _, action_text = analysis.partition(":")
+            action_text = action_text.strip()
             if action_text:
                 action_items.append(
                     {
@@ -137,20 +168,6 @@ def extract_structured_basic(text: str) -> dict[str, list]:
                     }
                 )
 
-        # Decisions
-        if lowered.startswith("decision:"):
-            _, _, body = analysis.partition(":")
-            decision_text = body.strip()
-            if decision_text:
-                decisions.append(
-                    {
-                        "participant": participant_name,
-                        "decision": decision_text,
-                        "rationale": "",
-                        "supporters": [],
-                    }
-                )
-
         # Mentions
         for match in MENTION_PATTERN.finditer(line):
             mentions.append(