Affinity vs Twenty

The tracked source is Affinity's blog index rather than a changelog, so most of what lands here is article chrome and case studies. Underneath that, two genuine product items are visible: Affinity Lists rebuilt for faster performance with advanced filtering, and an Affinity MCP Server in beta that connects deal data to AI tools including Claude, ChatGPT and Gemini. The surrounding material — Speedinvest tracking 500+ LP relationships through automatic capture, a year-in-review post, an executive appointment — is positioning around relationship intelligence rather than shipped change.

Affinity is pushing on two fronts that reinforce each other: making the CRM's own surfaces fast enough to live in, and making its relationship graph reachable from whatever AI tool a dealmaker already has open. The consistent claim across the entries is automatic capture — email and meeting data collected without manual entry — which is what makes both the network mapping and the agent access worth anything. The category framing is relationship intelligence, not CRM.

The MCP server is in beta and is the obvious next thing to move; watch for it to gain write access to pipeline records rather than read-only querying, since the existing material already describes updating a pipeline from an AI tool.