LiveStack Finance Edition Runbook (#27)

* build-plan update

* big-star-retail-ai-agent container

Hosts compose files to spin up locally on any environment.

* update big-star-retial

* Created .env.example

* livelabs-livestack-big-star-retail-screenshots-update

* some code changes in the stack.

some code changes in the stack.

* Changes to Compose yml and stack.

* main.py changes

In main.py:324, the / route was using the old template API:

  return templates.TemplateResponse("index.html", {"request": request})

  With the FastAPI/Starlette version in this container, TemplateResponse now expects:

  TemplateResponse(request, name, context=...)

  So the old call was effectively passing the context dict where the template name should go, which caused:

  TypeError: unhashable type: 'dict'

  I changed it to:

  return templates.TemplateResponse(request, "index.html")

  That’s what removed the 500 Internal Server Error on GET /. Nothing in the database, ORDS, or Ollama stack was changed.

* Seer Sporting Goods LiveStack WMS ID: 12050

* resized images

resized images

* html patch

* update ll-demo-agent LiveStack markdown

* Adding Windows Powershell Commands

* Minor change

* social-commerce-the-tim-app-livestack-and-guide

* Oracle Augmented Card Authorisation App

* Updated image sizes

* Updated naming

* oracle-augmented-bank updates to match new screenshots

* resized images

* Updated Seer Equities story

* port reference updates in documents.

* Finance Edition LiveStack Runbook

* Update image sizes

Author: mkowalik-git

finance-edition/guide/conclusion/conclusion.md

@@ -0,0 +1,57 @@
+# Conclusion and Business Outcomes
+
+## Introduction
+
+This closing lab consolidates the Seer Equity Bank Finance LiveStack story. The demo starts with a governed finance data foundation and ends with AI-assisted operators, auditors, analysts, and service teams working from the same Oracle-backed source of truth.
+
+Estimated Time: 10 minutes
+
+![Agent Console outcome view](images/agent-console-outcome.png)
+
+### Objectives
+
+In this lab, you will:
+- Review the final operator narrative across the scenes.
+- Connect visible application workflows to Oracle AI Database 26ai capabilities.
+- Summarize the business outcome for a stakeholder discussion.
+
+## Task 1: Review the end-to-end story
+
+1. Return to **Risk & Operations Dashboard** and note the operational summary cards.
+2. Open **Regulatory & Market Signals** and **Financial Crime Network** to review how signal and graph evidence explain the risk picture.
+3. Open **Predictive Risk & Revenue Analytics** and **Agent Console** to review how the app turns governed data into recommendations and logged actions.
+
+Expected result:
+- You can describe how the demo moves from trusted finance data to risk insight, investigation, service response, predictive analytics, and agent-assisted decisions.
+
+## Task 2: Review the Oracle evidence
+
+1. In any scene, expand or review the **Oracle Internals** panel.
+2. Identify the Oracle capability badges and SQL or PL/SQL evidence for the current workflow.
+3. Compare scenes that use different engines: JSON Duality, Vector Search, Property Graph, Spatial, VPD, OML, Select AI-style natural-language SQL, and agent audit tables.
+
+Expected result:
+- The same application shell shows multiple finance workflows, but Oracle AI Database 26ai remains the shared data and execution layer.
+- The Oracle Internals panel gives the presenter concrete evidence instead of relying on generic architecture claims.
+
+## Task 3: Capture the stakeholder narrative
+
+1. Summarize the demo in one sentence: Seer Equity Bank serves finance data, risk intelligence, fraud relationships, service coverage, analytics, and AI workflows from one governed Oracle data foundation.
+2. Identify which scene matters most for your audience:
+   - Executives: dashboard outcomes and predictive risk.
+   - Risk and fraud teams: regulatory signals and crime network.
+   - Operations teams: client service coverage and case routing.
+   - Data leaders: data foundation, dataset manager, and Oracle Internals.
+3. Use **Use Your Own Data** as the next-step call to action when a customer wants to map the story to their own finance data.
+
+Expected result:
+- You have a concise value narrative and a recommended next step for a follow-up customer workshop.
+
+## Task 4: Why this matters?
+
+Financial institutions often split dashboards, search, fraud graphs, geospatial service logic, ML, and AI agents across separate systems. This LiveStack demonstrates the business alternative: keep governed data close to Oracle, expose it through application scenes, and let every operator workflow reuse the same trusted foundation.
+
+## Credits & Build Notes
+- **Author** - LiveLabs Team
+- **Last Updated By/Date** - LiveLabs Team, 2026-05-11
+- **Build Notes** - The conclusion references the current app navigation and the final Agent Console scene.

finance-edition/guide/conclusion/images/agent-console-outcome.png

@@ -0,0 +1,147 @@
+# Download the LiveStack
+
+## Introduction
+
+This lab shows how to run the Seer Equity Bank Finance LiveStack from the portable package. The archive in this solution is `finance-livestack.zip`, and it expands directly into the working directory. Do not expect a nested `stack/` folder: `compose.yml`, `Containerfile`, `frontend/`, `backend/`, `db/`, `scripts/`, and `verification/` live at the extracted bundle root.
+
+Estimated Time: 30 minutes
+
+### Objectives
+
+In this lab, you will:
+- Prepare a clean working directory for the portable package.
+- Extract `finance-livestack.zip` into that directory.
+- Start Oracle Database, ORDS, Ollama, and the finance application with Podman Compose.
+- Validate the health endpoint and open the application.
+- Stop the stack cleanly when the demo is complete.
+
+## Task 1: Prepare the working directory
+
+1. Create a clean folder outside of `Downloads`.
+
+    ```bash
+    <copy>
+    mkdir -p ~/finance-livestack
+    <copy>
+    ```
+
+2. Move into the folder.
+
+    ```bash
+    <copy>
+    cd ~/finance-livestack
+    <copy>
+    ```
+
+3. Move the downloaded archive into the folder.
+
+    ```bash
+    <copy>
+    mv ~/Downloads/finance-livestack.zip .
+    <copy>
+    ```
+
+Expected result:
+- The file `finance-livestack.zip` is in a clean directory that will become the bundle root.
+
+## Task 2: Extract the package
+
+1. Extract the archive.
+
+    ```bash
+    <copy>
+    unzip finance-livestack.zip
+    <copy>
+    ```
+
+2. Confirm that the archive expanded directly into the current directory.
+
+    ```bash
+    <copy>
+    ls compose.yml Containerfile frontend backend db scripts verification
+    <copy>
+    ```
+
+Expected result:
+- You see `compose.yml` in the current directory.
+- You do not need to `cd stack` or `cd finance-livestack` after extraction.
+
+## Task 3: Start the LiveStack
+
+1. Start all services.
+
+    ```bash
+    <copy>
+    podman compose up -d --build
+    <copy>
+    ```
+
+2. Watch the service state.
+
+    ```bash
+    <copy>
+    podman compose ps
+    <copy>
+    ```
+
+3. If this is the first run on the machine, allow time for the Oracle Database image, ORDS image, Ollama image, ONNX embedding model, and `llama3.2` model to download and warm up.
+
+Expected result:
+- The `db`, `ords`, `ollama`, and `app` services move toward a healthy state.
+- The database bootstrap creates the `LIVESTACK` schema and loads the demo data automatically.
+- The app service listens on host port `8505`.
+
+## Task 4: Validate health and open the application
+
+1. Check the application health endpoint.
+
+    ```bash
+    <copy>
+    curl http://localhost:8505/api/health
+    <copy>
+    ```
+
+2. Open the LiveStack UI in a browser.
+
+    ```bash
+    <copy>
+    open http://localhost:8505
+    <copy>
+    ```
+
+3. If you are not on macOS, open `http://localhost:8505` manually in your browser.
+
+Expected result:
+- The health check returns a healthy JSON response after Oracle is ready.
+- The browser opens the Seer Equity Bank LiveStack with the control tower, left navigation, dataset action, and Oracle Internals panel.
+
+## Task 5: Stop the stack when finished
+
+1. Stop and remove running containers while preserving volumes.
+
+    ```bash
+    <copy>
+    podman compose down
+    <copy>
+    ```
+
+2. Use this command only when you intentionally want to delete the local Oracle and Ollama volumes for a complete reset.
+
+    ```bash
+    <copy>
+    podman compose down -v
+    <copy>
+    ```
+
+Expected result:
+- `podman compose down` stops the local LiveStack cleanly.
+- Demo data remains available on the next startup unless you use the explicit volume-removal command.
+
+## Task 6: Why this matters?
+
+The package is designed to be portable: one archive, one compose file, and one browser URL. Keeping the extraction flow flat and the startup command simple reduces setup drift, while the health endpoint confirms that the Oracle-backed app is ready before the demo starts.
+
+## Credits & Build Notes
+- **Author** - LiveLabs Team
+- **Last Updated By/Date** - LiveLabs Team, 2026-05-11
+- **Build Notes** - The local run path is derived from the current `finance-livestack.zip` archive and root-level `compose.yml`.

finance-edition/guide/download-livestack/download-livestack.md

@@ -0,0 +1,57 @@
+# Seer Equity Bank Finance LiveStack
+
+## Introduction
+
+This workshop is the demo runbook for the Seer Equity Bank Finance LiveStack. The application shows how a financial services team can serve trusted data, risk signals, fraud investigations, service coverage, transaction views, machine learning outcomes, natural-language questions, and agent workflows from one Oracle AI Database 26ai foundation.
+
+Estimated Demo Time: 90 minutes
+
+![Seer Equity Bank LiveStack control tower](images/control-tower.png)
+
+### Objectives
+
+In this workshop, you will:
+- Run the Seer Equity Bank LiveStack scene by scene.
+- Use the left navigation to move through finance data, risk, fraud, spatial, transaction, analytics, natural-language, and agent workflows.
+- Inspect the Oracle Internals panel in each scene to connect the visible workflow to Oracle AI Database 26ai capabilities.
+- Use the dataset manager to validate, replace, or restore the demo dataset.
+- Run the portable LiveStack locally from `finance-livestack.zip` with Podman Compose.
+
+### Prerequisites
+
+This workshop assumes you have:
+- A browser with access to the running LiveStack UI.
+- Podman and Podman Compose when running the downloaded LiveStack locally.
+- Network access for first-run container pulls and model downloads.
+- Basic familiarity with financial services operations, fraud review, compliance signals, and service coverage workflows.
+
+## Workshop Flow
+
+- Download and run the portable finance LiveStack.
+- Scene 1: Seer Equity Bank Control Tower.
+- Scene 2: Seer 26ai Data Foundation.
+- Scene 3: Risk and Operations Dashboard.
+- Scene 4: Regulatory and Market Signals.
+- Scene 5: Financial Crime Network.
+- Scene 6: Client Service Coverage.
+- Scene 7: Client Transactions and Cases.
+- Scene 8: Predictive Risk and Revenue Analytics.
+- Scene 9: Ask Seer Equity Data.
+- Scene 10: Agent Console.
+- Scene 11: Use Your Own Data.
+- Conclusion and business outcomes.
+
+## Learn More
+
+- Oracle AI Database 26ai converged data platform.
+- Oracle JSON Relational Duality Views.
+- Oracle AI Vector Search.
+- Oracle Machine Learning and `DBMS_DATA_MINING`.
+- Oracle Spatial and Graph.
+- Oracle Virtual Private Database and row-level security.
+- Oracle REST Data Services.
+
+## Credits & Build Notes
+- **Author** - LiveLabs Team
+- **Last Updated By/Date** - LiveLabs Team, 2026-05-11
+- **Build Notes** - Screenshots were captured from the running React application shell at `http://127.0.0.1:5173`. Full live metrics populate after the Podman stack brings Oracle Database, ORDS, Ollama, and the app service online.

finance-edition/guide/introduction/images/control-tower.png

@@ -0,0 +1,56 @@
+# Scene 1 Seer Equity Bank Control Tower
+
+## Introduction
+
+The control tower is the opening scene for the finance LiveStack. It orients the presenter around the application story: Seer Equity Bank serves finance data, risk workflows, API-backed applications, and AI experiences from one Oracle AI Database 26ai foundation.
+
+Estimated Time: 8 minutes
+
+![Seer Equity Bank control tower](images/control-tower.png)
+
+### Objectives
+
+In this lab, you will:
+- Open the Seer Equity Bank LiveStack.
+- Review the left navigation and the primary story paths.
+- Use the quick actions to move into the dashboard, data foundation, or natural-language experience.
+
+## Task 1: Open the control tower
+
+1. Open the running application at `http://localhost:8505`.
+2. Confirm the left navigation starts with **Seer Equity Bank Control Tower**.
+3. Review the workflow stage track: **Ingest**, **Transform**, **Serve Data**, **Serve AI**, and **Govern**.
+
+Expected result:
+- The application opens to the control tower scene.
+- The presenter can introduce the demo as a finance data and AI workflow, not as a collection of disconnected dashboards.
+
+## Task 2: Review the main story paths
+
+1. In the welcome panel, review the capability cards for converged queries, natural-language SQL, data applications, graph and spatial analytics, vector search, and REST access.
+2. Click **Start with Dashboard**.
+3. Return to the control tower by clicking **Seer Equity Bank Control Tower** in the left navigation.
+4. Click **Review Schema & Data**, then return again.
+5. Click **Ask Your Data**, then return again.
+
+Expected result:
+- Each quick action moves to a real application scene.
+- The left navigation remains the control point for the whole demo story.
+
+## Task 3: Identify the persistent operator controls
+
+1. Locate the top action labeled **Use Your Own Data**.
+2. Locate the left sidebar **Customer name** control.
+3. Locate the active dataset label in the sidebar footer.
+
+Expected result:
+- The user sees that the LiveStack can be narrated as the default Seer Equity Bank demo or adapted to a customer name and customer dataset.
+- Dataset status stays visible throughout the app.
+
+## Task 4: Why this matters?
+
+This first scene gives the audience a mental model for the rest of the workshop. Every later scene should be positioned as another operator view on the same governed Oracle-backed finance data foundation.
+
+## Credits & Build Notes
+- **Author** - LiveLabs Team
+- **Last Updated By/Date** - LiveLabs Team, 2026-05-11

finance-edition/guide/introduction/introduction.md

@@ -0,0 +1,55 @@
+# Scene 10 Agent Console
+
+## Introduction
+
+The Agent Console demonstrates finance agent workflows. Users ask questions that route to market and compliance, client service routing, or transaction revenue agents. Ollama handles reasoning while Oracle provides data access, SQL and PL/SQL execution, and durable action logging.
+
+Estimated Time: 12 minutes
+
+![Agent Console](images/agent-console.png)
+
+### Objectives
+
+In this lab, you will:
+- Open the Agent Console.
+- Choose a runtime profile.
+- Ask an agent question.
+- Review recent agent actions and Oracle evidence.
+
+## Task 1: Open the Agent Console
+
+1. Click **Agent Console** in the left navigation.
+2. Review the runtime profile selector.
+3. Review the **Chat with AI Agents** examples and the **Recent Agent Actions** section.
+
+Expected result:
+- The scene opens as an agent workbench, not a static chat demo.
+- The user sees that agent decisions are expected to become auditable records.
+
+## Task 2: Ask an agent question
+
+1. Click an example **Ask** tile, or type a finance operations question in the input.
+2. Click **Send**.
+3. Review the response and the agent team that handled the request.
+4. Review **Recent Agent Actions** after a successful action.
+
+Expected result:
+- With the full stack healthy, the app routes the question to an agent team, executes Oracle-backed tools, and records an action.
+- Recent actions show status, confidence, entity context, and payload summary.
+
+## Task 3: Inspect Oracle Internals
+
+1. Review the **Oracle Internals** panel.
+2. Point out the agent teams, Oracle SQL and PL/SQL tools, `agent_actions`, `event_stream`, vector retrieval, and in-database ML references.
+3. Explain that Ollama reasons, while Oracle remains the data and execution layer.
+
+Expected result:
+- The audience can see how agent behavior becomes traceable through Oracle-backed data and audit surfaces.
+
+## Task 4: Why this matters?
+
+Enterprise AI agents need more than a chat interface. This scene shows how finance agents can stay grounded in governed Oracle data, execute controlled SQL or PL/SQL tools, and leave an audit trail for operators and reviewers.
+
+## Credits & Build Notes
+- **Author** - LiveLabs Team
+- **Last Updated By/Date** - LiveLabs Team, 2026-05-11