DEMO

Building Blocks Q&A Using Headless IBM Bob

Ask any question about IBM Building Blocks and get instant, cited answers — powered entirely by Headless Bob's autonomous doc-crawling and streaming AI capabilities.

Building Blocks used:

ai.engineering

Looking for Building Block skills?

Skills provide best practices for IBM technologies

Download skills →

Products Used

IBM Bob

1 Review use case

IBM teams exploring the Building Blocks portfolio face a documentation discovery problem: 67+ pages spread across agents, AI Control Plane, data, and automation pillars with no conversational entry point. Sellers, architects, and developers spend time hunting for the right page instead of understanding how the blocks fit together.

The Building Blocks Q&A Bot solves this by putting Headless Bob in charge of everything beyond the UI shell. Bob autonomously crawls all 67 documentation pages, scores and ranks them by relevance to each incoming question, builds a grounded prompt with full context, and streams a markdown-formatted answer back to the user with source citations. Multi-turn conversation is preserved across follow-up questions within the same Bob thread, and the knowledge base can be refreshed on demand by clicking a single button — triggering another full Bob agentic crawl.

The demo runs on IBM Cloud OpenShift with two containerised services: the Headless Bob server (handling AI reasoning and thread management) and a Next.js frontend (handling routing, KB scoring, and SSE streaming to the browser). It showcases Bob's AI Agents and AI Control Plane building blocks in action — autonomous tool use, stateful multi-turn conversation, and real-time streaming — making it ideal for seller demos, architect walkthroughs, and internal enablement sessions.

Start Here

Watch this short overview

Preview the application

Explore key screens from the demo

2 Try the interactive application

Experience the interactive demo and explore its features. A demo script is provided to help when presenting the application to partners.

Try the demo

Have questions, want a walkthrough or found a bug?

Contact the demo owner for assistance.

Submit a request →

3 Explore how it was built

The demo consists of two containerised services deployed on IBM Cloud OpenShift. The Headless Bob server (port 8000) acts as the AI runtime — it manages conversation threads, executes agentic tool calls to crawl documentation pages, and streams token-by-token responses via Server-Sent Events. The Next.js frontend (port 3000) serves the chat UI and hosts four API routes: /api/chat scores the knowledge base by keyword relevance, selects the top 5 pages, injects them as context into a grounded prompt, and sends it to Bob; /api/stream proxies Bob's SSE stream directly to the browser; /api/threads manages thread listing and deletion; and /api/kb-status triggers Bob's agentic crawl to rebuild the knowledge base on demand. A shared CephFS PersistentVolumeClaim mounts knowledge-base.json into both pods simultaneously.

Architecture

4 Download and modify with IBM Bob

Download the source code and specifications for this application to set up and run it locally. Once downloaded, open the code in IBM Bob to modify and adapt it to your own use case.

Source Code

Building Blocks Q&A Bot — Demo Script

Overview

Demonstrates how Headless Bob autonomously crawls IBM Building Blocks documentation, builds a knowledge base, and answers natural-language questions with cited, streamed responses — all with the Next.js app acting as a thin UI shell.

Step 1 — Orient the Audience

Open the app URL — point out the dark green sidebar and the four suggestion cards, one per Building Blocks pillar: Agents, AI Control Plane, Data, and Automation.
Draw attention to the KB Ready badge in the top-right corner — explain that Headless Bob has already autonomously crawled all 67 Building Blocks documentation pages and the knowledge base is live and indexed.

Step 2 — Ask a Question from a Suggestion Card

Click one of the suggestion cards — for example: “What building blocks are available for building AI agents?”
Watch the answer stream word-by-word in real time directly from Headless Bob via SSE.
Point out the source citation chips that appear below the response — each chip links directly to the specific documentation page Bob used to ground its answer.

Step 3 — Follow-Up Question (Multi-Turn Conversation)

Type a follow-up question in the input box — for example: “How do these agents connect to watsonx.ai?”
Show that the response continues in the same thread (same entry visible in the left sidebar) — demonstrating Bob’s stateful multi-turn conversation capability.
Key talking point: Bob is maintaining full context across turns without any custom session management code in the app.

Step 4 — Revisit a Previous Thread

Click a previous thread in the left sidebar.
Show that the full conversation history — all prior questions and answers — loads back instantly.

Step 5 — Demonstrate Agentic KB Refresh

Click the ↻ refresh icon at the bottom of the sidebar.
Narrate: “Bob is now autonomously crawling all 67 pages of the Building Blocks documentation site and rebuilding the knowledge base live — this is Bob doing the work, not a script.”
Point out the last crawl timestamp in the bottom-left corner — this updates once the crawl completes (~5–10 minutes).

Step 6 — Key Talking Point

Emphasise to the audience:

“The Next.js app is a thin UI shell. Bob handles all the intelligence — crawling, reasoning, context injection, streaming, and thread management. This is exactly what Headless Bob enables: autonomous agentic capability with no custom AI orchestration code.”