TCI Convergence Engine

A personalised convergence briefing engine for the World Economic Forum and Capgemini Tech Convergence Initiative. Executives answer six questions and receive a tailored 4-6 page briefing grounded in TCI research.

What it does

An executive selects their sector, role, time horizon, and strategic intent, then describes their specific situation. The engine runs a 4-agent pipeline (Interpreter, Mapper, Synthesiser, Verifier) against the TCI corpus (253 technology subcomponents, 23 combinations, 235 cross-cutting mappings, two years of report text) and generates a structured briefing with convergence maps, opportunities and risks, action items, and full source citations.

This is a research instrument, not a chatbot. One intake, one briefing.

Tech stack

• Framework: Next.js 16 (App Router)
• Styling: Tailwind CSS 4
• LLM: Anthropic Claude (claude-sonnet-4-20250514) via @anthropic-ai/sdk
• Language: TypeScript 5
• Runtime: Node.js 20+

No database, no vector store, no RAG pipeline. The corpus is small and structured enough for deterministic filtering and direct LLM synthesis.

Quick start

Prerequisites

• Node.js 20 or later
• An Anthropic API key

Install

cd app
npm install

Configure

cp .env.example .env.local

Edit .env.local and add your Anthropic API key:

ANTHROPIC_API_KEY=sk-ant-...

Run

npm run dev

Open http://localhost:3000 in your browser.

Project structure

app/
  src/
    app/
      page.tsx                 Intake form (6-step wizard)
      briefing/page.tsx        Briefing display and export
      api/briefing/route.ts    POST endpoint, chains 4 agents
    lib/
      agents.ts                Agent implementations (Interpreter, Mapper, Synthesiser, Verifier, Orchestrator)
      corpus.ts                Corpus loading, filtering, and caching
  docs/
    PRD.md                     Product requirements
    architecture.md            System architecture
    agents.md                  Agent definitions and schemas
    skills.md                  Skill definitions
    corpus.md                  Data dictionary
    editorial-rules.md         Briefing voice and style
    decisions.md               Architectural decision records
    roadmap.md                 Phase plan and progress

../corpus/                     TCI corpus data (JSON + report text)
../agents/                     Agent prompt definitions
../skills/                     Skill specifications

Documentation

See docs/ for detailed documentation:

• PRD -- product requirements and scope
• Architecture -- system design and data flow
• Agents -- agent pipeline and schemas
• Skills -- corpus query capabilities
• Corpus -- data dictionary and update guide
• Editorial Rules -- briefing voice and style
• Decisions -- architectural decision records
• Roadmap -- phase plan

Built by

World Economic Forum and Capgemini Tech Convergence Initiative (TCI), Year 3.

License

Internal use. Not for public distribution.