What AI tools are most useful for investor deal sourcing in India?

Claude (Anthropic) for startup research, competitive landscape analysis, and pitch deck gap analysis. Antigravity for workflow automation — enriching new leads and managing follow-up sequences. The combination saves 8-10 hours per week of research and admin overhead.

How does AI improve startup evaluation for investors?

AI handles the cognitive scaffolding — competitive analysis, pitch deck review, follow-up drafting — freeing up mental bandwidth for higher-value work: attending events, having genuine founder conversations, and making judgment calls that require real context.

Are there AI tools that don't work well for investors?

AI meeting transcription tools (manual note-taking produces better real-time synthesis) and fully AI-generated outreach to founders (lower conversion than personally written messages — founders can tell the difference between genuine curiosity and a generated message).

The AI Tools I Actually Use for Deal Sourcing (And What They Replaced)

There’s a lot of content about AI tools for investors. Most of it is written by people who have tried a product for two weeks and written a listicle. This is not that.

These are the tools I use every single day in my work as an ecosystem builder at Delhi Angels — sourcing startups, evaluating founders, building pipelines. I’ll tell you what I use, what I replaced, how long it took to actually work, and where it still falls short.

The Problem AI Was Solving for Me

Before I list the tools, it’s worth understanding the underlying workflow problem. Deal sourcing at the early stage in India is fundamentally a research and communication problem. On any given week, I’m doing some combination of:

Researching 5–10 new startups that came through referrals or event interactions
Preparing context on founders before calls
Summarising pitch decks for review meetings
Following up with founders across different stages of the pipeline
Tracking what I’ve already seen to avoid duplication

All of this is cognitively intensive but largely repeatable. That’s exactly where AI earns its keep.

Claude for Research and Communication

Claude (Anthropic) is the tool I use most heavily and the one that’s changed my workflow the most.

For startup research: Given a founder’s LinkedIn, a website, and a one-liner, I can ask Claude to identify the competitive landscape, surface comparable companies globally, and flag potential concerns — in about 3 minutes. What used to take 20–30 minutes of manual searching now takes a fraction of that. The output isn’t perfect, but it gives me a strong starting framework to interrogate and verify.

For pitch deck analysis: I’ll feed the text content of a pitch deck and ask Claude to identify what’s missing against standard early-stage evaluation criteria — no clear retention numbers, no mention of competitive differentiation, CAC not mentioned but customer acquisition strategy described. It produces a structured gap analysis I can bring into the call as a question list.

For follow-up emails: I hate writing formulaic follow-up emails. I give Claude the context of the conversation — what the founder said, what my concerns were, what the next step is — and ask it to draft a concise follow-up that sounds like me. I edit it, but the base is there in 30 seconds.

The key thing I’ve learned about using Claude effectively: specificity in the prompt produces specificity in the output. “Analyse this startup” produces generic output. “Identify the three biggest risks in this business model given that their primary customer is an Indian SMB in tier-2 cities” produces something actually useful.

Antigravity for Workflow Automation

Antigravity is where the repetitive, structured parts of my workflow live.

The biggest workflow it handles: when a new startup comes in via referral or form submission, it automatically pulls the company URL, LinkedIn profiles, and any attached deck, runs a basic profile enrichment, and creates a structured entry in my tracking system. I used to do this manually for every lead. It was 15–20 minutes per startup. Now it’s essentially zero.

I’ve also used it to set up automated follow-up sequences — if a founder hasn’t responded to an initial note in 7 days, a reminder goes out. Not a generic reminder; a contextual one that references the specific conversation. The setup took time, but it runs itself now.

What Antigravity doesn’t do well: anything that requires judgment. It’s excellent at structured, rule-based processes. The moment a situation requires nuance — a founder who’s pivoted mid-conversation, a deck that doesn’t fit standard categories — it needs a human in the loop. I’ve learned not to over-automate the parts of the workflow that actually require thinking.

The Stack That Didn’t Work

In the interest of being honest: I’ve tried tools that didn’t stick.

AI meeting transcription tools — I tried two of these. In theory, having an automatic transcript of every founder call sounded useful. In practice, I found that the act of taking notes manually forces me to synthesise in real time, which produces better output than a raw transcript I have to re-read later. I stopped using them after a month.

AI-generated outreach to founders — I tested using AI to draft cold outreach to founders I wanted to connect with. The messages were fine. But they didn’t convert as well as messages I wrote myself. There’s something about the specificity of genuine human curiosity that AI can approximate but not replicate, and founders can feel the difference. I use AI to draft and edit, not to replace, outreach.

What This Has Actually Freed Up

The honest accounting: the tools above have probably saved me 8–10 hours a week of research, admin, and communication overhead. That’s time I’ve redirected into the higher-value parts of the work — attending more events, having more genuine founder conversations, thinking more carefully about the deals in the pipeline.

The counterintuitive thing about AI tools in this context: their value isn’t just time saved. It’s cognitive load reduced. When I’m not spending mental energy on the research scaffolding, I show up to founder conversations more present and more prepared. That quality of attention is what actually moves deals forward.

If you’re building AI tools for investors or ecosystem builders in India, I’d be interested in the conversation. Reach out here.

Related reading: How I Evaluate Early-Stage Startups: A Framework from 100+ Reviews — the evaluation framework these tools are built to support. And: Inside Delhi NCR’s Startup Ecosystem — the ecosystem context where this deal sourcing work actually happens.