Beyond the Pitch Deck: How Top Seed Investors Are Using AI to Find the Next Unicorn
You know the drill. Every week, a dozen inbound decks land in your inbox, each claiming to be “the AI for X” or “the Y of Z.” Sourcing is noisy, pattern recognition is fatigued, and your pipeline feels like a firehose of buzzwords.
But here’s what I learned from talking to the partners who made this year’s Seed 100 and Seed 40 lists: they’re not just investing in AI—they’re using AI to do their jobs better, faster, and with sharper conviction.
These aren’t experiments with ChatGPT for fun. These are production-grade workflows that turn field notes into insights, board decks into searchable knowledge bases, and cold emails into warm introductions.
Let’s break down exactly how three top-tier VCs are weaving AI into their daily operating rhythm, and what you can steal for your own team.
Salil Deshpande: The AI Agent That Runs the Back Office
General Partner, Uncorrelated Ventures
Salil didn’t just dabble with AI. He trained an agent to run his entire back office.
Think about that for a second. How many hours do you lose every week to scheduling, email triage, and portfolio admin? Salil’s agent handles all of it—end-to-end scheduling, email campaigns, and even LinkedIn invitation management. It auto-accepts portfolio founders and relevant VCs while flagging others for his manual review.
But the real killer app is what Salil calls a “portfolio doctor.”
Here’s how it works:
- Anyone on his team or an investor can email a subject line like: “What’s the ARR of Cast.ai?”
- The agent digs through board decks and investor updates stored in Salil’s Dropbox.
- It synthesizes an answer with specific citations and replies directly to the email thread.
No more digging through folders. No more “I’ll get back to you next week.” The output is instant, cited, and actionable.
The GTM lesson: Your own team can build a similar “portfolio doctor” for customer health. Imagine an agent that surfaces churn risks from support tickets, ingested call transcripts, and product usage data—and emails the CS team a weekly report with direct citations. That’s not sci-fi. That’s Salil’s Monday morning.
Ann Miura-Ko: Turning Field Notes Into a Pattern Recognition Engine
Partner, Floodgate
Ann took a deliberately analog approach first. She spent months visiting a dozen AI-native companies in person—from 4-person startups to Ramp—documenting how they actually operate. Not their pitch decks. Their operations.
Then she used AI to ingest all those field notes, cross-reference them across companies, and surface patterns she’d never see from any single visit.
This is pattern recognition at scale. No single founder meeting gives you the full picture. But when you feed a dozen operations playbooks into an AI model, you start seeing convergence: how AI-native teams structure their engineering sprints, how they pitch to enterprise buyers, how they think about pricing.
Ann calls this the “empirical backbone of my analysis of the AI-pilled startup.” And those patterns are already reshaping how she evaluates investments and founders.
The GTM lesson: Your sales team probably has dozens of win-loss debriefs, discovery call notes, and churn analyses sitting in spreadsheets or CRMs. Stop treating them as individual data points. Ingest them all into an AI tool and ask: What patterns emerge from our top 20 wins vs. our top 20 losses? You’ll get back a pattern you can act on—not a gut feeling.
Anne Dwane: Agentic Workflows That Keep Networks Fresh
Cofounder and General Partner, Village Global
Anne put it succinctly: “Networks don’t stay fresh on their own.”
She’s using agentic workflows to pinpoint the “first-call” angels in today’s most promising founder ecosystems. Think of it as dynamic network mapping. Instead of relying on static spreadsheets or stale LinkedIn relationships, Anne’s AI continuously monitors where top angels are investing, which founders they back, and how those networks evolve.
The result? Her team always knows who to call first when a hot new startup emerges.
The GTM lesson: Your sales team can apply the same logic to account mapping. Instead of manually tracking who’s connected to whom in your target accounts, use an agentic workflow that monitors job changes, mutual connections, and recent funding events. When a champion moves to a new company, your AI should flag it and suggest a re-engagement sequence. Networks that stay fresh win.
What This Means for Your Revenue Team
Here’s the throughline across all three investors: they’re not using AI to replace judgment. They’re using AI to augment pattern recognition, compress time, and keep knowledge accessible.
Here are three actions you can take this week:
1. Build Your Own “Second Brain” for Customer Data
Take a page from Salil. Set up an agent that ingests your support tickets, call transcripts, and product usage data. Train it to answer questions like: “What’s the NPS of our top 10 accounts?” or “Which features are most correlated with churn?” Make it email you a weekly synthesis with citations.
2. Run a Cross-Deal Pattern Analysis
Do what Ann did, but with your own data. Pull your last 20 win-loss records, discovery call notes, and competitive battle cards. Ingest them into an AI tool and ask: “What do our top wins have in common that our losses don’t?” The answer will surprise you—and it’ll be backed by data, not anecdotes.
3. Automate Your Network Monitoring
Follow Anne’s lead. Set up a workflow that monitors your CRM, LinkedIn, and funding news for changes in target accounts. When a champion moves, an investor syndicate closes a new round, or a competitor makes a key hire, get an alert. Then act fast.
The Bottom Line
The best investors aren’t just looking for AI startups. They’re using AI to become better investors. And the same logic applies to revenue teams: the ones who win in 2025 won’t be the ones who use AI to generate more emails. They’ll be the ones who use AI to see patterns that humans miss, compress months of analysis into hours, and keep their networks alive and actionable.
So go ahead. Build your own portfolio doctor. Ingest your field notes. Let your agent map the network.
Because the startups that back you—and the customers that buy from you—deserve a partner who’s already operating at machine speed.
Want more GTM playbooks like this? Subscribe to B2B Pulse for weekly deep dives into how the best revenue teams are winning with data, AI, and relentless execution.