Client Thread — AI-Maintained Knowledge Files for Every Client Relationship
A business wanted everyone on their team to have a complete, current picture of every client relationship — without anyone maintaining it manually. We built an AI system that monitors inboxes across the business and keeps a living knowledge file per client, updated automatically.
The brief
Client conversations were happening across multiple inboxes, across multiple people. Knowledge stayed locked in individual inboxes and faded from memory. The brief: monitor the inboxes, know which threads relate to which client, extract what matters, and keep a single always-current file per client that anyone can read in 60 seconds before a call or meeting.
How it works
- Multi-inbox monitoring — all relevant inboxes watched continuously across the business
- Client identification — each thread matched to the right client by domain, name, and context
- Intelligent extraction — AI identifies decisions, concerns, commitments, and status changes
- Living knowledge files — each client has a structured current.md, updated automatically
- Context retention — important history kept even as new information arrives
- AI question answering grounded in real relationship history
What changed
Anyone can read a client's current file before a call and be fully up to speed. A conversation in one person's inbox is visible to the whole team. When the team asks the AI about a client, it answers from actual relationship history.
Client kept anonymous at their request.