Partner playbook
Sales team memory, in an afternoon
Every account, contact, call and decision in one engine, so each salesperson and their AI tools start from the full picture. About three hours with the client.
The steps
- Create the engine and write its purpose (15 minutes). One engine for the client’s sales team. In a few sentences, say what it is for: accounts, contacts, calls, decisions and next steps, and what does not belong in it. AI tools read the purpose to decide what to save there.
- Read the client’s website (10 minutes). Give the engine the client’s website when you create it. It suggests notes, values, voice and colours; keep what is right. Nothing is saved until you keep it.
- Bring in what already exists (30 minutes). Drop the team’s account plans, price lists and past proposals into Files, as files, whole folders or a .zip. Text and PDFs become searchable, and a folder name becomes a tag.
- Connect each salesperson’s AI tool (45 minutes). Each person connects the tool they already use: ChatGPT, Claude, Claude Code or Cursor, at Can read and save. Each person’s own tool counts as one of the engine’s agents, so check the plan has room.
- Show them what to save (30 minutes). Walk through one real account: the customer, the people at it, the last meeting, the decisions made and why, and the open tasks with an owner and a due date. The AI tool saves each one as its own memory.
- Invite the team (15 minutes). Invite every salesperson and their manager. Users are free on every plan. Give managers admin, the team member, and anyone who only reads viewer.
- Hand over the habits (45 minutes). Show Ask for questions about an account, Summaries for what happened this week, and the Map for how people, customers and decisions link. At the start of each session, every AI tool gets a brief of what changed.
- Decide who owns it. Keep the engine on your account and bill the client your way, or hand it over: the client accepts it and takes over paying, and authorship stays. Everything can be exported as Markdown and JSON at any time.
Where the team uses a CRM or a call recorder, their own AI tool reaches it through its own connectors, and Context Engine is the memory it reads from and saves to.