Skip to content

AI & Intelligence

Intelligence built on your actual records.

AI in Trelyn is a layer over the system, not the system itself. It reads the deals, meetings, projects and conversations your team already keeps, only what the person asking may see, and turns them into briefings, suggestions and drafts. People decide what changes.

AI features: Growth and Pro.

How AI works in Trelyn

Useful because it knows the context. Safe because it stays in its lane.

  • From your records

    Briefings and answers cite the activities, meetings and documents they come from.

  • Only what you may see

    Every answer is drawn from records the person asking has access to.

  • People apply changes

    Suggestions change nothing until someone applies or confirms them.

  • Apart from the numbers

    Reports and automations are deterministic; AI does not compute or trigger them.

Where it helps

On the record, where the decision is made.

01Deal Intelligence

Know which deals need you, and why.

A score with momentum, buying readiness and qualification, the risks, objections and stakeholder gaps behind it, and a recommended next action with a “why this?”. It is a suggestion: no field is changed and no task is created unless you do it.
  • Score, momentum and buying readiness
  • Risks, objections and stakeholder gaps

02Meeting Intelligence

Prepared before the call, followed up after it.

Before a meeting, a briefing per person: what to know, questions to ask, objectives, what to avoid. After it, the transcript is turned into key outcomes, commitments and proposed CRM updates, applied only when you select them, and a follow-up email draft in the tone you pick.
  • Pre-meeting briefing on meetings and bookings
  • CRM updates: apply selected, or reject
  • Transcripts from Fathom

03Summaries and briefings

Where a client stands, in a paragraph.

Summaries on companies, contacts, deals, engagements, projects and deliverables, and account briefings that collect opportunities, risks and key contacts. Relationship Intelligence maps the stakeholders at a company or on a deal: champions, relationship strength and who to involve next.
  • Summaries on six record types
  • Stakeholder maps for companies and deals

And around the work

Smaller things that save time every day.

  • Opportunities

    Expansion and new-deal opportunities surfaced on your accounts, create a deal or dismiss with a reason.

  • Writing help

    A careful copy edit you review before using it.

  • The assistant

    Ask about anything you can see; answers come with their sources.

  • Search by meaning

    Find notes and documents by what they say, not only their words.

MCP for Claude and ChatGPT

Your CRM, in the AI client you already use.

Connect Claude or ChatGPT to Trelyn over MCP and work with your companies, deals, delivery and files from there, with your own permissions, not a shared admin key.

  • A personal token per connection, shown once and stored as a hash
  • Read only, or read and write
  • Your current role decides what the connection can reach
  • Revoke at any time; write actions are logged
  • MCP access: every plan

Questions

About AI in Trelyn.

Does the AI change records by itself?
No. Intelligence panels suggest; they do not edit the CRM. Proposed updates from a meeting change nothing until someone applies them. The assistant writes a note or your own task straight away when you ask, and asks you to confirm anything else before it happens.
What can the AI see?
Only what the person asking may see. Retrieval and every answer go through the same permissions as the rest of the workspace, so a team member never gets an answer drawn from a record they cannot open.
Can we use Trelyn without AI?
Yes. Trelyn is a complete system without it: the CRM, delivery, portal, automations and reports do not depend on AI. AI features appear where your plan includes them and your workspace has them switched on.
Does the AI send emails or contact clients?
No. A follow-up email is drafted for you to copy or open in your own email app; you send it.
How does MCP access work?
Each person creates their own connection with a token that is shown once and stored only as a hash. A connection is read only or read and write, works with that person's current permissions, can be revoked at any time, and its write actions are logged.