Automation and AI

AI has made most investment workflows cheap to automate. The hard part has moved to the organization: who owns the scoring model, who sees which deals, and which parts of an investment committee memo still need a human to write them. Andre Retterath (Data Driven VC) shares what he's learned from the funds furthest along, what works, and what fails. We'll cover why a scoring model launched before the data is clean can lose the deal team's trust, and why knowing who at the firm knows a founder, CEO or banker depends on reliable email and calendar data. We'll also get into permission-aware MCP setups and why proprietary data, judgment and access are still where the edge comes from, whether you invest in startups, buyouts, credit or real assets.

Attendees will leave with:

• Permissions and control: role-based access in a central MCP, so sensitive deals and data stay with the people who should see them, plus review gates for skills your team builds
• Memos people actually read: how one fund found 80%+ of its IC papers repeated decks, data rooms and websites, and rebuilt its format around the deal team's judgment
• Sourcing and scoring: a two-layer scoring design (hard filters that reflect your mandate, plus a learned layer), measuring coverage by whether you saw the deals that mattered, and structured reviews of the ones you missed
• Relationship data as the foundation for reaching management teams and founders through the people your firm already knows