How to build an investor list with AI

Most founders build their investor list by opening Crunchbase, filtering by "fintech seed," and emailing 200 people. Then they wonder why they get a 3% response rate.

The process is backwards. Databases are for filling gaps. Your network is where you start. The founders who close fast know this. They spend the first two weeks mining warm paths, not cold prospects, and they only reach for a database once they've exhausted who they already know.

AI agents change how fast you can do this. What used to take two weeks of manual research can now happen in a day. Here's the full process, including which tools and skills to use, so you can build a list that actually converts.

We built a free fundraising toolkit with skills and agents for every stage of this. That's what I'm referencing throughout. If you want context on how these work, this post covers the full picture.

Why your investor list is probably wrong right now

The typical founder approach: Google "top seed VCs," find a blog post with 50 names, add them all to a spreadsheet, start emailing. It's not that this never works. It's that it's leaving 80% of your warm paths on the table.

The founders I watched raise the fastest at LaunchHouse, the accelerator I ran, always started with the same question: who do I already know that is an investor, or knows investors? That's the question that unlocks everything. And it's not something you can answer with a database. You have to look at your actual network.

The second mistake: targeting investors who are a bad fit. A VC who focuses on Series B enterprise software is not going to back your pre-seed consumer app. Every cold email you send them is noise that burns your reputation in that network. Narrow targeting beats broad targeting every single time.

The 6-phase process

The build-investor-list skill breaks this into six phases. Here's what each one does and why it's in this order.

Phase 1: Pull your contacts

Before you touch any database, export everything you have. Gmail, LinkedIn, calendar, Twitter. The goal is a complete picture of your network. Most people are surprised how many investors they already know (or are one hop away from).

If you're using Micro, this step is automatic. Your contacts are already pulled in, enriched, and relationship-scored from your actual communication history. You skip straight to analysis.

Phase 2: Enrich your contacts

A name without context isn't useful. You need titles, current companies, and whether someone is actually an investor. For this you can use Clay, Apollo, or Proxycurl. Or just run the enrichment through the lead-research-assistant skill, which chains multiple data sources automatically.

The goal here is turning a raw contact list into structured data you can actually reason about.

Phase 3: Classify and score relationships

This is where most people's process falls apart. They have a list of 1,000 contacts but no way to identify which 50 are investors, which 200 are founders who can make warm intros, and which 30 are connectors who know everyone.

Run the vc-investor-analyst agent on your enriched contact list and let it classify and score. It looks at someone's title, their portfolio, their investment history, and produces a fit score against your specific company profile. You tell it your stage, sector, and raise amount. It tells you which of your contacts are worth prioritizing and why.

Relationship strength matters here too. Knowing someone from college who went on to become a partner at Sequoia is different from having met them at a conference once. The system scores recency, frequency, and interaction history so you can see actual relationship strength, not just a name in your phone.

Phase 4: Build your tiered target list

Now you build the list. Three tiers:

• Tier 1: Direct warm connections to investors who are a strong fit. These get your best outreach, the most personalized emails, and the most follow-up.
• Tier 2: One-hop warm intros. You know someone who knows them. These get targeted ask emails for intros.
• Tier 3: Database fills. Investors who fit your thesis but where you have no warm path. Cold outreach only after you've worked T1 and T2.

How many investors should be on your list? The right number depends on your stage. Pre-seed: 60-100, with more angels at smaller check sizes. Seed: 60-100. Series A: 30-60, more targeted. Aim for 60% or more warm intro coverage. If you're below 50%, you need to spend more time on network mining before starting outreach.

The worst thing you can do is start with 500 names and fire off generic emails. It signals inexperience and it works at maybe 1/10th the rate of a tight, warm list.

Phase 5: Map your warm intro paths

For every T1 and T2 investor, you need to know who will make the intro. Not just "I think Sarah might know them." You need a specific person, a specific ask, and a specific timeline.

The build-investor-list skill generates a warm intro map: for each target investor, it shows you the mutual connections in your network and suggests the best path. In Micro, this is built into the contact view. You can see at a glance who on your team is connected to any investor, and how strong that connection is.

Warm intros convert at 20-30% versus 1-3% for cold. That spread is not subtle. The intro path is the fundraising process.

Phase 6: Output the list

The skill generates a structured CSV with firm name, partner name, tier, intro path, relationship strength score, and current status. From there you run outreach. The investor-outreach skill takes the list and writes personalized first emails for each investor based on their actual portfolio and stated thesis.

Where to find investors for your startup

Once you've mined your network, here are the best sources for filling the remaining gaps with database targets:

Source	Best for	Cost
Signal (NFX)	AI-matched investor recommendations based on your company profile	Free
OpenVC	20K+ verified investors across stage and sector	Free
AngelMatch	110K+ angels and VCs with verified emails	Freemium
Crunchbase	Largest public funding database, investment history search	Freemium
Ramp VC Database	2,200+ active investors with contact info, curated	Free
LaunchGravity	Who's actively investing right now, founder intelligence	Paid

My honest ranking: Signal and OpenVC first, because they're free and match you algorithmically. Crunchbase for research on specific firms once you have names. AngelMatch if you want email addresses at scale. LaunchGravity if you want real-time signal on who's actively writing checks.

Do not pay for PitchBook unless you're running a Series A or later and need fund performance data. The free tier alternatives cover 95% of what you need at seed.

For a broader take on how AI is changing fundraising, see how founders are using AI to raise venture capital.

How to research individual investors before you reach out

Once you have names, you need to research each one before sending anything. The vc-investor-analyst agent does this for you at scale: give it your list, it produces a fit score and a personalized outreach hook for each investor based on their portfolio, public statements, and thesis.

The research you actually need on each investor:

1. Have they funded companies in your sector in the last 18 months? (Investment recency matters.)
2. Do they have a portfolio conflict? (Competitor companies in their portfolio means a pass, usually.)
3. Do they lead, or do they follow? (You want at least one lead investor on your list.)
4. What's the right check size for this fund? (A $5M fund is not writing a $2M seed check.)
5. Is there a warm path, and who in your network makes the most credible intro?

Investing 20-40 hours in research before outreach sounds like a lot. It's not. It's the difference between a 25% response rate and a 3% response rate. The math on that time investment is obvious.

What changes when AI does the work

This whole process used to take two or three weeks. Export your contacts manually, research each investor by hand, build the spreadsheet, figure out intro paths from memory. Most founders skip half the steps because it's too slow, and they pay for it with bad response rates.

AI agents compress it. The build-investor-list skill and the vc-investor-analyst agent handle most of Phases 2-5 automatically. Enrichment, classification, scoring, research, intro mapping. What was two weeks is now a few hours of reviewing output and approving actions.

The skill runs inside Claude Code or Cursor. Copy the SKILL.md file into your .claude/skills/ directory. Done. The fundraising guide has setup instructions.

In Micro, Phases 1-3 are already handled. Contacts pulled from your email, LinkedIn, and calendar. Enriched automatically. Relationship strength scored from real interaction history. You skip straight to building the tiered list.

The thing most people miss

Fundraising is a relationship problem, not a research problem.

Every founder thinks they need better data. More names, better emails, a bigger list. Turns out what they need is warmer paths and tighter targeting. The founders who close the fastest don't have the most investors on their list. They have the most warm paths per investor on their list.

AI gets you to those paths faster. But the relationship work is still yours to do. The tool just means you're not spending two weeks in a spreadsheet before you start.

Start with the agentic fundraising toolkit. The build-investor-list skill is where to begin.