AI Context Layer Demo
The context layer demo lives here.
This is the demo companion to the CRM AI context layer guide. It shows how a workflow loads business rules before it researches a record, builds context, drafts an outreach email, and keeps a rep in the loop.
What the demo proves
The point isn't the fictional company or the outreach use case. The point is the operating pattern: keep the rules outside the prompt, load only the guidance the workflow needs, and make the generated output show what it used.
- Rules live in git as reusable context.
- n8n runs the workflow and calls the model.
- Postgres stands in for CRM data.
- The draft stays grounded and rep-reviewed.
Current state
The source repo is public. The hosted interactive version is the next step: database, seed data, n8n webhooks, and a safe fallback mode need to be wired behind this URL before I call it live.
Until then, this page is the canonical home for the demo. The guide links here first, and this page will become the interactive runner when the app is deployed.