fix: spotter viz story — clean prompt-only format, include full model URL

- Remove ### Step N: headers and Expected result/What to look for lines
from both story prompts — Spotter Viz is an AI agent, it only needs
the raw prompts
- Pass full model URL (ts_url/#/data/tables/model_guid) so Spotter can
identify the data source without asking for confirmation

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

chat_interface.py

@@ -2271,7 +2271,8 @@ To change settings, use:
 - **End with action**: Show how insights lead to decisions""")
     
     def _generate_ai_spotter_story(self, company_name: str, use_case: str,
-                                    model_name: str = None, liveboard_name: str = None) -> str:
+                                    model_name: str = None, model_url: str = None,
+                                    liveboard_name: str = None) -> str:
         """Pure AI-generated Spotter Viz story — no matrix reference.
         AI decides what a compelling liveboard story looks like for this company + use case.
         """
@@ -2285,6 +2286,8 @@ To change settings, use:
         company_context = f"Company: {company_name}\nUse Case: {use_case}"
         if model_name:
             company_context += f"\nData Source/Model: {model_name}"
+        if model_url:
+            company_context += f"\nModel URL: {model_url}"
         if liveboard_name:
             company_context += f"\nLiveboard Name: {liveboard_name}"
         if hasattr(self, 'demo_builder') and self.demo_builder:
@@ -2313,7 +2316,8 @@ To change settings, use:
             return f"*(Generation failed: {e})*"
 
     def _generate_matrix_spotter_story(self, company_name: str, use_case: str,
-                                        model_name: str = None, liveboard_name: str = None) -> str:
+                                        model_name: str = None, model_url: str = None,
+                                        liveboard_name: str = None) -> str:
         """Matrix/ThoughtSpot-recommended Spotter Viz story.
         Builds from the vertical×function matrix (KPIs, liveboard_questions, story controls, persona).
         AI writes it — adds narrative — but every step comes from the matrix.
@@ -2334,6 +2338,8 @@ To change settings, use:
         use_case_name = uc_cfg.get("use_case_name", use_case)
 
         matrix_context = f"Company: {company_name}\nUse Case: {use_case_name}\nData Source/Model: {data_source}"
+        if model_url:
+            matrix_context += f"\nModel URL: {model_url}"
         if liveboard_name:
             matrix_context += f"\nLiveboard Name: {liveboard_name}"
         if persona:
@@ -3932,10 +3938,18 @@ Ask these questions to showcase ThoughtSpot's AI capabilities:
                 
                 # Generate Spotter Viz Stories (both matrix and AI versions)
                 try:
+                    _model_name = results.get('model', None)
+                    _model_guid = results.get('model_guid', None)
+                    _ts_url = (self.settings.get('thoughtspot_url') or '').rstrip('/')
+                    if _model_guid and _ts_url:
+                        _model_url = f"{_ts_url}/#/data/tables/{_model_guid}"
+                    else:
+                        _model_url = None
                     _story_args = dict(
                         company_name=company_name,
                         use_case=use_case,
-                        model_name=results.get('model', None),
+                        model_name=_model_name,
+                        model_url=_model_url,
                         liveboard_name=results.get('liveboard', None),
                     )
                     self.spotter_story_matrix = self._generate_matrix_spotter_story(**_story_args)

prompts.py

@@ -571,58 +571,58 @@ Create a bullet outline demo script with:
 - Spotter questions to ask live
 - Closing value proposition""",
 
-    "spotter_viz_story_matrix": """You are creating a ThoughtSpot-recommended Spotter Viz story.
+    "spotter_viz_story_matrix": """You are writing a Spotter Viz story — a sequence of prompts a user pastes one at a time into Spotter Viz, ThoughtSpot's AI liveboard builder.
 
-Spotter Viz is an AI agent in ThoughtSpot that builds Liveboards through natural language prompts.
+IMPORTANT: Spotter Viz is an AI agent. Output ONLY the prompts themselves — no step headers, no labels, no "what to look for" commentary. The prompts are the output.
 
-You have been given the ThoughtSpot-recommended KPIs, visualizations, and liveboard questions for this exact vertical and function. Your job is to turn those into a natural, conversational demo script — a sequence of prompts a user would type into Spotter Viz to build the liveboard step by step.
+You have been given the ThoughtSpot-recommended KPIs, visualizations, and liveboard questions for this exact vertical and function. Every KPI and visualization listed must appear as a prompt.
 
 {context}
 
 ---
 
-Write the story as a numbered sequence of prompts. Every KPI and visualization listed above must appear as a step. Format each step as:
+Output a numbered list of prompts. Each item is ONLY the prompt text — nothing else.
 
-### Step N: [Brief label]
-> "[The exact prompt to type into Spotter Viz]"
-
-**What to look for:** [1 sentence on what the viz reveals — use the insight from the matrix if provided]
+Example format:
+1. "Create a new liveboard for [Company] using the data model [model name] at [model URL]. This is for [persona] to monitor [goal]."
+2. "Add KPI tiles with monthly sparklines for [metric1], [metric2], and [metric3]. Show the latest value prominently with a 12-month trend."
+3. "Add a bar chart showing [metric] by [dimension]."
 
 Rules:
-- Step 1: set context — company name, data source, overall goal
-- KPI steps: include time dimension and granularity (e.g. "revenue for the last 12 months as a monthly KPI sparkline")
-- Chart steps: keep the metric name and dimension from the matrix (e.g. "revenue by region as a bar chart")
-- Add 1-2 narrative sentences between steps to tell a story — what are we looking for, what does it mean?
-- Final step: the "aha moment" — name the key insight or outlier from the story controls
-- Use the company name and real metric names from the matrix above
-- Do not invent KPIs or visualizations that are not in the matrix""",
+- Prompt 1: company name, full model URL, persona, and business goal
+- KPI prompts: name every metric, specify monthly sparklines and 12-month trend
+- Chart prompts: name the metric, chart type, and dimension to slice by
+- Final prompt: the "aha moment" — highlight the key outlier or insight from the story controls
+- Use real metric names from the matrix above
+- Do not invent KPIs or visualizations that are not in the matrix
+- Do not add any text outside the numbered prompts""",
 
-    "spotter_viz_story": """You are creating a Spotter Viz story for ThoughtSpot's AI-powered liveboard builder.
+    "spotter_viz_story": """You are writing a Spotter Viz story — a sequence of prompts a user pastes one at a time into Spotter Viz, ThoughtSpot's AI liveboard builder.
 
-Spotter Viz is an AI agent in ThoughtSpot that creates, structures, and styles Liveboards through natural language prompts. Users type conversational requests and the agent builds the dashboard step by step, allowing iterative refinement.
+IMPORTANT: Spotter Viz is an AI agent. Output ONLY the prompts themselves — no step headers, no labels, no "expected result" commentary. The prompts are the output.
 
-Your job is to write a sequence of natural language prompts that a user would type into Spotter Viz to build a liveboard for this demo scenario. The prompts should be conversational, specific, and progressively build the liveboard from scratch.
+Your job is to write 6-10 natural language prompts that build a compelling liveboard for this demo scenario, from scratch.
 
 {context}
 
 ---
 
-Write a Spotter Viz story as a numbered sequence of prompts. Format each step as:
-
-### Step N: [Brief label]
-> "[The exact prompt to type into Spotter Viz]"
+Output a numbered list of prompts. Each item is ONLY the prompt text — nothing else.
 
-**Expected result:** [1 sentence describing what Spotter Viz should create]
+Example format:
+1. "Create a new liveboard for [Company] using the data model [model name] at [model URL]. This is for [persona] to monitor [goal]."
+2. "Add KPI tiles with monthly sparklines for [metric1], [metric2], and [metric3]. Show the latest value prominently with a 12-month trend."
+3. "Add a bar chart showing [metric] by [dimension]."
 
 Rules:
-- Step 1 should set context: company name, data source, and overall goal
-- Steps 2-3 should add KPIs with sparklines (mention time granularity)
-- Steps 4-6 should add key visualizations (charts, breakdowns, comparisons)
-- Steps 7-8 should refine: add styling, rename the liveboard, organize into groups/tabs
-- Final step should call out the "aha moment" — the key data story or outlier to highlight
-- Use the company name and real column/metric names from the context above
-- Keep prompts natural — write how a business user would actually talk
-- Include 6-10 steps total""",
+- Prompt 1: company name, full model URL, persona, and business goal
+- Prompts 2-3: KPIs with monthly sparklines and 12-month trend
+- Prompts 4-6: key charts — name the metric, chart type, and dimension
+- Prompts 7-8: refinements — styling, liveboard name, organize into groups
+- Final prompt: the "aha moment" — the key insight or outlier to highlight
+- Use real column/metric names from the context above
+- Keep prompts conversational — how a business user would actually talk
+- Do not add any text outside the numbered prompts""",
 }
 
 DEFAULT_TEMPLATE = """You are helping create a {vertical} {function} demo.