Prompts for Affinity MCP (Claude, ChatGPT & other tools)

​

Prep & research

• “Brief me on my meeting with [Firm] tomorrow — last three interactions, any open notes, and their relationship strength with our team.”
• “Summarize our full relationship history with [Company] — interactions, notes, and key moments over the past three years.” (value scales with how long you’ve used Affinity)
• “What were the key discussion points and action items from my last call with [Company]?”
• “Who on our team has the strongest connections at [Company]? Rank them.”
• “Pull the transcript from our meeting with [Company] last week and give me the highlights.”
• “Here’s the attendee list for [conference] — cross-reference against Affinity and rank by relationship strength and pipeline status.” (for lists over ~100 companies, ask it to work in batches)

​

Search & find

• “Find family offices in [City] on our [List] we haven’t contacted in 90 days.” (include the list name — accuracy drops without it)
• “Search [List] for companies with a status of ‘Active’, sorted by date added.”
• “Find companies working on [thesis] in our network.”
• “Give me a full list of every [stage] company we’re tracking, with their [field].” · “How many companies are in our CRM?”
• “Who is our warmest path to [Person]? Show it as a graph.”
• “Show me how the Stage field on [Company]‘s pipeline entry has changed over time.” (field-value history — not available in AI Chat)
• “Search our notes for any mention of ‘[topic]’ across our deals.” · “Search our files for [Company]‘s latest pitch deck.” (needs Deal Assist)

​

CRM updates

• Move a stage / update fields — “Move [Company] to ‘Due Diligence’ and log a note that we received their materials today.” · “On [Company]‘s entry in [List], set Stage to ‘Diligence’ and Deal Lead to [Name].”
• Log a note — “Log a note on [Company]: discussed Series B timeline, lead investor hesitant on valuation, follow up in two weeks.” · “Attach a note to my meeting with [Company] recapping the action items.” (a note on the meeting — this can trigger Suggested Field Updates)
• Create records — “Create an Affinity company for [Company] (domain [x]), add [Person] as a contact, and create an opportunity for them on our [Pipeline] list.”
• Pitch deck / document → record — “Parse this pitch deck and create a new Affinity record with the relevant company fields populated.” (review the auto-filled fields before saving)
• Update across a list — “For every company we met at [conference] this week, set Follow-up Status to ‘Reviewed’.” (the model works entry-by-entry; for very large lists, do it in batches)
• Define structure — “Add a ‘Lead Source’ text field on companies.” · “Add a ‘Not Relevant’ option to the Status field on our [List].”
• Log an interaction — “Log the call I had with [Person] this morning about pricing.”
• Set reminders — “Set a reminder to request year-end financials from [Company] on January 1.” · “Recurring reminder to check in with [Person], assigned to me.”

​

Reports & multi-step

• Pipeline reports on demand — “Give me a pipeline summary — deals by stage, total value at risk, and which opportunities have gone cold this month.” (great for ad-hoc; not a replacement for a persistent dashboard yet)
• Outreach sequencing — “Pull my [segment] contacts with their relationship strength and last-contact date, and draft a personalized follow-up for each.” (drafts only — you review and send)
• Write scores back — “Here are this week’s [model] scores by company — write each into the [field] on the matching Affinity record.”
• Transcript → fields — “Extract the funding details from this meeting transcript and update the relevant fields on our pipeline entry.”
• Duplicate cleanup — “Surface likely duplicate company records so we can merge them.”
• Combine with your other tools — “Cross-reference [Company] against our notes in Notion and Affinity’s relationship data, and tell me who to loop in.” (connect that tool’s own MCP too)

​

The five MCP skills

• Meeting prep (affinity-meeting-prep) — “Use Affinity to build my brief for today’s external meetings.”
• Warm intro (affinity-warm-intro) — “Find intro paths to [target] and draft the outreach.”
• Market map (affinity-market-map) — “Build a market map for [thesis], score by fit, and shortlist the top companies.”
• Data migration (affinity-data-migration) — “Migrate this CSV from [prior system] into Affinity.”
• Event setup (affinity-event-setup) — “Set up tracking for [event].”

​

Good to know

• Plan: MCP is on Scale, Advanced, and Enterprise. Everything runs with your own permissions — MCP can only see and change what you can.
• Feature-gated tools: note/file search need Deal Assist; meeting tools need unified events onboarding.
• Known edges: very large lists (~100+) can hit your AI tool’s context limit — batch them. File write-back (pushing files into Affinity) and reading email bodies aren’t supported — pair with that tool’s own MCP. Location/geographic filtering isn’t native yet.