How do you find emerging startups before they raise? A Crunchbase workflow, from a market you're watching to a committee-ready thesis.

By the time a category shows up on everyone's list, the good rounds are gone. The firms that win the early deals spot the niche first and reach the founder with a warm introduction before the round is competitive. A connected AI workflow gets you there in an afternoon. You describe the market you're interested in, and you get back the fastest-growing sub-categories, the specific startups inside them most likely to raise next, and, with your CRM connected, which of those companies fit your track record and who at your firm can already make the intro.

The workflow runs in the AI assistant your firm already uses, with two connectors switched on: Crunchbase for human-verified private-company and funding data, and Affinity for your own deal history and warm paths. Crunchbase tells you where the market is going while Affinity tells you where you have a right to win. Each step below is one plain-language prompt with the parameters in brackets, so you can copy it and change what's in the brackets to fit your mandate. The full sequence sits at the end, ready to run in order.

How do you spot a category before it's obvious?

Start above the company level. Ask which corners of a market are accelerating, and screen out the noise with real thresholds.

Using Crunchbase data, identify the emerging [B2B SaaS] categories with the fastest growth in [seed and Series A] funding over the last [18] months versus the prior [18] months. Only include categories with real scale: at least [50] companies, [10] recent early-stage deals, and [$50M] in recent funding. Rank them by funding growth. For each, give me a short market map: the trend driving it, a few notable startups, the most active investors, and why it still looks early.


The thresholds do the work. Without them the model returns categories that look explosive only because they started from zero; with them you get a ranked map of markets large enough to build a thesis on. Change the bracketed values to reset the sector, the stage, or the bar for scale.

How do you find the niche inside the category?

A category is too broad to act on. The money moves inside it, so follow it down a level.

Go deeper on the top category, [category name]. Break it into four to six sub-themes or product wedges. For each wedge, show seed and Series A funding, deal count, growth versus the prior period, notable companies, and the most active investors. Then tell me which wedges look most strategically important, and whether the growth looks durable or mostly hype.


Crunchbase resolves markets down to micro-industries, so the answer names the specific wedges where early dollars are concentrating rather than restating the macro trend you already knew.

Which startups are about to raise?

Now drop to the company level and ask the question that sources a deal.

Within [wedge], identify three or four fast-growing startups that are likely preparing for their next funding round. For each, explain the signals that point to a raise and why the company is strategically interesting.


Predicting who is about to raise is only useful if you can trust the signal behind it, which is the difference between verified data and an answer scraped off the open web.

"If you're a private capital firm betting your entire future on these investments, you really need to be able to trust that data." — Nick Hill, Crunchbase

How do you pressure-test a company before you reach out?

Before a name goes to your team, build the case and the counter-case in one session.

Write an investment memo for [company]: business model, why now, leadership team, full funding history and lead investors, market position and competitors, and growth signals. Then give me the bear case: the legitimate reasons a disciplined investor might pass. End with the top five diligence questions I should be asking before a first meeting.


One more prompt turns it into something you can hand the committee.

Turn that memo and bear case into a one-page investment thesis in plain terms: a short overview, the pros, and the cons, laid out as a committee-ready one-pager.


You walk in with the memo, the objections, and the questions already drawn up.

Where does your own track record come in?

Everything above works without your CRM. Connecting Affinity is what makes the list yours.

"You feed your last 12 months of wins back into Crunchbase, find more companies just like them, and now your own track record becomes the search query." — Kyle Turner, Head of Solutions Engineering, Affinity

Using my last 12 months of closed-won deals in Affinity as a reference set, find companies in [category or wedge] that resemble them, and rank the matches by how closely they fit the profile of deals we have actually won. For each, show whether anyone at my firm already has a relationship or a warm path in.


Now the search points at companies you have a right to win: the ones that match the deals you have closed, and the ones you can already reach. This is the step that needs your own data, and it is where the search stops being generic.

What you end up with

A path from "which markets are heating up?" to a committee-ready thesis on a specific startup that fits your track record, in an afternoon rather than the weeks a manual version takes. Crunchbase supplies the verified market data; Affinity supplies the deal history and the warm path; and because the two connect directly, your own wins shape what the search returns. Every deal you close makes the next one easier to find.

The full workflow, copy and run

Here is the whole sequence in order. Copy each step into the AI assistant your firm uses, with Crunchbase connected throughout and Affinity connected for the last step. Change what's in the brackets to fit your mandate.

1. Spot the category

Using Crunchbase data, identify the emerging [B2B SaaS] categories with the fastest growth in [seed and Series A] funding over the last [18] months versus the prior [18] months. Only include categories with real scale: at least [50] companies, [10] recent early-stage deals, and [$50M] in recent funding. Rank them by funding growth. For each, give me a short market map: the trend driving it, a few notable startups, the most active investors, and why it still looks early.


2. Find the niche inside it

Go deeper on the top category, [category name]. Break it into four to six sub-themes or product wedges. For each wedge, show seed and Series A funding, deal count, growth versus the prior period, notable companies, and the most active investors. Then tell me which wedges look most strategically important, and whether the growth looks durable or mostly hype.


3. Find who's about to raise

Within [wedge], identify three or four fast-growing startups that are likely preparing for their next funding round. For each, explain the signals that point to a raise and why the company is strategically interesting.


4. Build the memo and the bear case

Write an investment memo for [company]: business model, why now, leadership team, full funding history and lead investors, market position and competitors, and growth signals. Then give me the bear case: the legitimate reasons a disciplined investor might pass. End with the top five diligence questions I should be asking before a first meeting.


5. Make the committee one-pager

Turn that memo and bear case into a one-page investment thesis in plain terms: a short overview, the pros, and the cons, laid out as a committee-ready one-pager.


6. Match it to your own track record (Affinity connected)

Using my last 12 months of closed-won deals in Affinity as a reference set, find companies in [category or wedge] that resemble them, and rank the matches by how closely they fit the profile of deals we have actually won. For each, show whether anyone at my firm already has a relationship or a warm path in.


The last step only works with your own deals connected

The final prompt needs your firm's own history in the loop, and that is the Affinity MCP: your relationship and deal data connected to the AI assistant you already use, so the search knows which companies you have a right to win and who can make the introduction. To get started with Affinity MCP, follow our step-by-step guide here.

More AI workflows across the deal lifecycle

This is one stage of a four-part series on connecting AI to where your firm's work already lives.

• Start here: Putting AI to work across the private capital deal lifecycle—the full thesis and all four workflows
• Sourcing (Grata): How do you find off-market companies to acquire?
• Diligence (AlphaSense): How do you use AI for due diligence?
• Monitoring (Lumonic): How do you monitor a private credit portfolio?

FAQ

How do you use AI to find emerging startups before they raise?

Ask an AI assistant connected to a verified data source like Crunchbase to rank the fastest-growing sub-categories of a market, then to name the specific startups inside them showing signs of preparing a round. Connecting your CRM lets the same workflow prioritize the startups that match your past wins and where you already have a relationship.

Can you trust AI's answer on which companies are about to raise?

Only if you can see the source. A general model will produce an answer with no traceable basis; a workflow grounded in human-verified funding data returns predictions you can tie back to real signals, which matters when the decision is an investment.

What is a micro-industry?

A narrow sub-segment inside a broader market category, granular enough to show where early-stage dollars are concentrating. Working at the micro-industry level surfaces specific wedges and companies instead of a macro trend everyone can already see.

How does a CRM improve AI deal sourcing?

It grounds market data in your firm's own history. With a CRM like Affinity connected, the workflow can weight the search toward categories where you've won before and companies you already have a warm path to, so the output reflects where your firm has a right to win, weighting your own track record over raw market growth.

What is the best AI workflow for deal research?

A strong AI deal-research workflow connects the model to a verified data source so it can rank the fastest-growing sub-categories of a market, name the specific companies likely to raise next, and assemble a memo with sources you can check. Connecting your CRM adds your own track record, so the research weights toward the companies and categories where your firm has already won.