How Antler Built a Firm-Wide Meeting-Intelligence Layer with MeetGeek

Antler runs cohorts in 15 countries and 25+ cities at once. Rather than roll out a note-taker, they made MeetGeek the standardized layer that captures every meeting across the firm, then pulled that data into their own centralized AI through MeetGeek’s privacy-oriented API.

‍

At a glance

• Company: Antler, a global early-stage VC firm and startup generator
• Scale: 15 countries, 25+ cities
• Integrations: MeetGeek native integrations (Slack, Notion, HubSpot, CRM), 10,000+ apps via Zapier, and the API
• Use case: Firm-wide meeting capture feeding an in-house investment AI, via MeetGeek’s API

‍

Meeting intelligence is the practice of capturing and structuring every conversation, then using it to drive work, giving automated workflows the context they need and triggering the actions that follow. It isn’t only about capture; it’s about what the meeting sets in motion.

‍

At Antler, that layer is MeetGeek. More than a note-taking app, it’s the meeting-data infrastructure beneath the firm’s entire operation, including the centralized AI they built in-house.

‍

Every meeting is captured, transcribed, and normalized into one structured format: founder interviews, partner discussions, portfolio reviews, internal syncs. From there it flows into the tools partners already work in and, through MeetGeek’s privacy-oriented API, into that AI.

‍

That difference between a note-taker and a data layer is the story. For an early-stage investor operating across continents at once, standardized, privacy-controlled meeting data isn’t a convenience. It’s the substrate an investment AI runs on, and the shared record every partner works from and builds on.

‍

Who is Antler?

Antler is a global early-stage venture capital firm and startup generator. Unlike traditional VC firms that back companies that already exist, Antler invests in individual founders before there’s a team, a product, or even a finalized idea, running structured cohort-based residency programs in cities worldwide.

‍

That model is deliberately intensive. Each cohort runs on its own calendar, with its own founders and its own partners running interviews. Antler’s edge is identifying exceptional founders before the evidence is obvious. That judgment depends entirely on who can access what was said across the organization, and on whether that record is structured enough to reason over.

What breaks down at Antler’s scale?

At Antler’s scale, cohorts run in parallel across every time zone. Every meeting produces something that has to travel: a conviction signal on a founder, a portfolio update, a decision made in a partner sync. All of it has to reach the colleagues who need it, and a system that can act on it. If those conversations live as loose notes scattered across inboxes and docs, no human and no AI can see the full picture.

‍

Fixing that takes more than asking partners to share notes. It requires every meeting to arrive already structured, in one queryable format, automatically, the moment it ends, with no one waiting on a write-up.

‍

How does Antler capture every meeting?

Antler captures every meeting automatically through MeetGeek, regardless of how or where it happens. It isn’t a founder-interview tool bolted on; every conversation across the organization enters the same structured record.

‍

Operating across dozens of markets and cultures, Antler’s partners meet in very different ways, and what’s appropriate to record, and how, shifts with the setting, the expectations in the room, and the person’s role.

‍

MeetGeek gives them several apps to match: the Meeting Agent bot on Zoom, Google Meet, and Teams; the mobile app for in-person conversations; and the desktop app and Chrome extension for calls where they'd rather not have a bot.

‍

How does that data reach the tools teams already use?

At Antler, structured meeting data flows into those tools automatically. No one moves it by hand. The routing runs on MeetGeek’s native integrations, including Slack, Notion, HubSpot, and leading CRMs, and on 10,000+ apps via Zapier, Make, and n8n. Anything more bespoke runs through the API.

‍

Delivery is only half of it. Because MeetGeek treats a finished meeting as an event, the record doesn’t just land somewhere to be read later. It can trigger the next step on its own: posting a summary to the right channel, updating a record, or kicking off a multi-step workflow.

‍

The meeting becomes the thing that sets work in motion, not a document someone has to remember to act on.

‍

How does Antler’s own AI run on MeetGeek?

Antler’s build starts with MeetGeek’s API, which exposes transcripts, summaries, highlights, and performance KPIs through REST and webhooks.

‍

The API also offers EU and US regional endpoints and privacy controls, so Antler decides exactly what data flows where before any of it reaches an AI. For an investor handling confidential founder, partner, and portfolio conversations, that control matters as much as the data itself.

‍

On top of MeetGeek’s API, Antler feeds the normalized data into a centralized AI system they built in-house. The result isn’t individual transcripts on request. It’s a single body of knowledge the full partnership can reason across, spanning every conversation Antler has captured, in any cohort, from any city.

A partner preparing for a second meeting can pull everything Antler knows about a founder across every prior interaction, and an investment committee can look for patterns across dozens of conversations from a single cohort.

‍

Because MeetGeek already solved capture, transcription, and normalization, Antler could focus their build on the AI layer that is actually their edge.

‍

Separately, partners use MeetGeek’s MCP integration with Claude to build their own personal automations on top of the same meeting-data layer.

‍

One layer ends up working at two levels at once: it makes each partner more effective on their own meetings, and it gives the firm as a whole intelligence across all of them.

How did MeetGeek become Antler’s firm-wide standard?

MeetGeek became standard at Antler through a phased, long-term partnership, not a one-time purchase. That work covered feature gaps, configured org-wide settings, and standardized how meetings are captured and aggregated. Today the same practice runs across a globally distributed team, so information from any meeting or cohort is available to whoever needs it, or whatever system needs it, regardless of time zone or city.

‍

What does this mean for other distributed teams?

Antler’s setup is specific, but the principle isn’t. Any organization running meetings across time zones hits the same failure mode: insight trapped with whoever attended.

‍

The fix is the same, too: capture that’s automatic and uniform, a record that lands in the tools people already use, and data structured cleanly enough for a system to reason over.

‍

Get those right, and the payoff is that no signal gets lost between locations, and decisions get made with full context regardless of who produced the conversation.

‍

For a distributed team, that’s an operational edge, not a convenience.

Key takeaways

• Meeting intelligence means every meeting, from founder interviews to partner discussions, portfolio reviews, and internal syncs, is captured, structured, and made available to the full partnership automatically, not left with whoever was in the room.
• Antler operates across 15 countries and 25+ cities, with meetings happening continuously and independently across every time zone.
• MeetGeek captures every meeting regardless of format, including bot-free via the desktop app and Chrome extension, and every recording produces the same output: transcript, summary, highlights, action items, and metadata.
• MeetGeek routes that record into the tools teams already use, including Slack, Notion, HubSpot, CRMs, and 10,000+ apps via Zapier, and a finished meeting can trigger the next step automatically rather than waiting to be read.
• Using MeetGeek’s public API, Antler feeds that data into a centralized AI they built themselves, giving the full partnership one body of knowledge to reason across, from any conversation, cohort, or city. Partners also use the MCP integration with Claude to build their own tools on the same layer.
• Adoption came through a phased, long-term partnership: covering feature gaps, configuring org-wide settings, and standardizing practice across a globally distributed team.
• The result is full-context decisions at global scale, with MeetGeek as Antler’s meeting-data infrastructure, not just a note-taker.

‍

If your team is losing context across meetings and time zones, the problem usually isn’t that people take bad notes. It’s that the conversation never becomes data anyone else can use. 

‍

MeetGeek turns every meeting into a shared, structured record, routes it into the tools your team already works in, and lets it trigger the work that follows. Try MeetGeek for free and make every conversation part of the system your team runs on, instead of a note someone has to remember to write.