What shipped in July: Ascend, and the infrastructure it runs alongside

Ascend got most of July's attention, and deservedly so. But three releases shipped alongside it, and together they change how much you can do with the relationship and deal data already sitting in Affinity. This is the recap: what launched, what it solves, and how the pieces connect.

The thread running through all of it is control. Whether you want an agent Affinity built, an integration your engineers are wiring up, or a data warehouse your team already runs, July's releases give you more ways to act on the same underlying data.

Key takeaways

• Affinity Ascend went live on July 22 with three agents (Meeting Prep, Warm Intros, Data Update) and four MCP skills callable from Claude or ChatGPT.
• List Search and Filter is now available in both MCP and the API, so agents and integrations can query lists, companies, and people directly instead of pulling everything down first.
• Databricks Data Share entered open beta: self-service, admin-enabled, covering 30+ tables and refreshing every two hours.
• New Lists heads to general availability in a phased, cohort-by-cohort rollout starting in August.

ICYMI: Affinity Ascend is live

If you missed it, here is the short version: Affinity Ascend went live on July 22, and it puts AI to work on the parts of the job that used to eat your morning.

Three agents are now live. Meeting Prep assembles a briefing on the people and companies you are about to meet, so you walk in already knowing the relationship history. Warm Intros traces the shortest path from your firm's network to a target, so you can ask the right colleague for the introduction instead of going in cold. Data Update reads your meeting notes and suggests the field changes they imply, so your CRM stays current without the manual cleanup.

Alongside the agents, Ascend introduced four MCP skills you can call from inside Claude or ChatGPT: Meeting Prep, Warm Intro, Data Migration, and Event Setup. Because they run where your team already works, the relationship context you need shows up in the same window you are already typing in, rather than behind another login.

"Our team already works inside Claude all day, so having our CRM context available immediately in the same interface using Affinity's MCP connector just feels natural. Instead of digging through tabs before a meeting, we can ask Affinity for relationship history and recent interactions directly in Claude." — Joff Redfern, GP, Menlo Ventures

Read the full Ascend announcement

List Search and Filter comes to MCP and the API

Agents are only as fast as the data they can reach, and until now reaching the right records meant grabbing far too many. On the MCP side, an agent had to pull a whole list into memory just to filter it afterward, which burned tokens and fell apart on large lists. On the API side, there was no way to filter lists, companies, or people across global, enrichment, or list-specific fields at all, so pipeline integrations, data-quality audits, and sync jobs were harder to build than they should have been.

Now both surfaces can search and filter lists, companies, and people directly. Filters chain together with AND/OR logic, support a full range of operators across text, date, and number fields, and return results in the sort order you specify. That means an agent can answer a question like "which companies on our dealflow list are in Watching status, owned by me, with no contact logged in 2026" in one targeted query. No more pulling the entire list into memory to sift through it afterward.

The practical payoff is a new class of agent. A data-quality agent can scan a list for empty required fields. A pipeline agent can pull a weekly status summary on a schedule. A productivity agent can surface the specific deals a person owns that match a set of criteria. On the API side, the same capability lets you resolve a person by their LinkedIn URL or trigger an integration only when narrow conditions are met, rather than syncing everything and filtering downstream.

Databricks Data Share, now in open beta

Getting Affinity data into a warehouse used to mean building and babysitting API scripts or middleware, working around schema quirks, rate limits, and performance overhead just to keep the warehouse current. Databricks Data Share removes that tax.

An admin turns it on from Settings > Integrations > Databricks. No engineering project, no pipeline to maintain. The share covers more than 30 tables, including notes, transcripts, interactions, relationship strength scores, and associations, and it refreshes every two hours so your warehouse stays close to live. It also surfaces records that have been deleted from Affinity, which gives your team a clean trail for audit and data-integrity work. With this update, your analysts can point their existing Databricks tooling at Affinity data directly, and spend their time on analysis instead of plumbing.

Databricks Data Share documentation

New Lists just got a whole lot better

New Lists is a ground-up rebuild, not a retrofit. That shows up in ways you notice immediately: faster performance, AND/OR filtering, and an editing model built around how deal teams actually work through a list day to day.

A lot has already shipped since New Lists first opened to firms. You can rename, delete, duplicate, and favorite lists. You can create and manage dropdown, ranked dropdown, and status fields, and reorder or recolor options right from the column header. Removing a record is a single click with one-click undo. A misstep costs you a second, not a confirmation dialog.

This release also adds CSV export and the ability to create entities directly within a list.

Ready to try it? Opt in now from your list settings. If you'd rather wait, your account will move over automatically starting in August. You'll get advance notice before anything changes, and Classic remains available throughout.

How July's releases fit together

Look at the four releases side by side and the shape is clear. Ascend is Affinity’s new agent platform. Databricks Data Share is a bridge to a warehouse your team already runs. List Search and Filter in the API is a foundation your team can build on. New Lists is the workspace where the day-to-day happens. These are four different points of access to one data layer underneath.

That is the point worth holding onto from July: you are not locked into one way of working with your Affinity relationship and deal data. You can let an agent act on it, pipe it into your own systems, or work it directly in a faster interface. The data is the constant; how you reach it is increasingly your call.

Enjoyed this? More from this series

What shipped in June: more Affinity data where your firm already works

FAQ

What is Affinity Ascend?

Ascend is Affinity's AI agent platform. It launched on July 22 with three agents (Meeting Prep, Warm Intros, and Data Update) and five MCP skills (Meeting Prep, Warm Intro, Market Map, Data Migration, and Event Setup) that you can call from Claude or ChatGPT.

What does List Search and Filter do in MCP and the API?

It lets agents and integrations search and filter lists, companies, and people directly, with AND/OR logic, operators across text, date, and number fields, and custom sorting, instead of pulling an entire list down and filtering it afterward.

What is Databricks Data Share?

It is a self-service integration, enabled by an admin from Settings > Integrations > Databricks, that shares more than 30 Affinity tables, including notes, transcripts, interactions, and relationship strength scores, into your Databricks warehouse. It refreshes every two hours and also surfaces deleted records for audit purposes. It is currently in open beta.

When does New Lists become generally available?

New Lists moves to GA in a phased rollout by cohort beginning in August. Each account is notified before its cohort switches, you can opt in early, and reverting to Classic remains an option throughout.

What is not available in New Lists at GA?

Bulk editing, bulk email, inline filtering, inline entity creation, and inline option creation are not part of general availability.

Want the full detail on the agent platform? Read the Ascend announcement