Build the Investable Universe
Define the outer edge of what could ever be in pipeline — companies, investors, themes.
Building the universe is, at the start, a CRM + data provider mission. Before you fire a single shot at an AI or data driven sourcing motion, you need a clean ocean to fish in. This module covers how you assemble that ocean for the first time.
Prerequisites from the systems blueprint
If you followed the systems blueprint, you already have your CRM stood up with team emails and calendars synced, and any data from prior systems ported across. If you migrated from Affinity to Attio (or vice versa), the port is genuinely just a few day exercise (or less). It's not as scary as it looks. Don't let migration anxiety delay the rest of this work.
Top down: start with a data provider
The first decision is which provider sits at the head of the funnel. Three we'd recommend evaluating for a sourcing database:
Deep historical coverage, strong on later stages and institutional rounds. The default for funds that need defensible deal data.
Built for early-stage sourcing — strong founder and team graph, fast filters, designed for outbound motions.
Web, hiring, and product signal layered on top of company data. Useful when you want momentum signals, not just raised rounds.
Deeper writeups live in the vendor profiles: PitchBook, Harmonic, and Specter.
Filter the universe
The provider is just raw material. The work is filtering it down into the universe your thesis actually addresses. Layer filters until what remains is plausibly investable:
Sizing the list
It's better to over-capture than under. Typical list sizes vary with stage:
Export into the CRM
Once the filtered list is right, export it into the CRM as the seed of your pipeline. The easiest path is almost always a CSV with two key columns:
- Company name — for humans.
- Website — the unique identifier the CRM dedupes on.
Other fields are nice to have, but website is the spine. If two records share a domain, they're the same company.
Bring in your existing companies
The startup database gives you a fresh pool of names, but it is only half the picture. You also have internal data — companies that have already synced into your CRM from email and calendar history. If you are not starting a firm from scratch, a lot of work has already been done with those companies. The goal is to get those existing records out of the dead-company catch-all and into the Deal Pipeline list alongside the new export.
You can start with sorting and filters. That might surface some obvious candidates, but it rarely catches everything.
The better approach is to use an AI attribute in a CRM like Attio that can read the full record detail and classify each company as a pipeline startup or not. Create a yes / no AI attribute, run it across every record, then filter on yes and add those companies to the list. That way your pipeline contains both the fresh external universe and the internal relationships you have already been building.
The Deal Pipeline list
In Attio or Affinity, create a dedicated Deal Pipeline list and load these records there. Do not live out of the default Companies or People objects — those are catch-alls and quickly become noise. The Deal Pipeline is one of the most important assets the firm owns; treat it as a first-class construct from day one.
Keep the universe fresh
The initial export is a snapshot. The universe drifts the moment you finish loading it — new companies incorporate, raise, and ship every day. A one-time CSV decays into a stale list within a quarter. The fix is to wire two recurring inflows into the Deal Pipeline so the CRM replenishes itself without manual re-exports.
Thesis-shaped deal feed
Stand up a saved search at your provider that mirrors the same sector / stage / geo / recency filters you used for the initial pull. Schedule it to fire on a cadence — daily for early stage, weekly for later — and pipe new hits straight into the Deal Pipeline list. Any company that newly matches your thesis lands in CRM without anyone touching a spreadsheet.
Active watchlist with alerts
Maintain a named watchlist in PitchBook, Harmonic, or whichever provider you use — the specific companies the partnership is actively tracking. Turn on alerts for funding events, headcount jumps, leadership changes, and product launches. When a watched company moves, the alert routes into CRM as a fresh signal on the existing record.
Together, the push feed grows the universe and the pull watchlist keeps the records inside it current. Both terminate in the same Deal Pipeline list, so the partnership has one place to look — and the CRM never drifts more than a day or two out of date.
High-volume deal newsletters (StrictlyVC, Term Sheet, Axios Pro Rata) are a third source of net-new companies — the catch is they're mostly noise relative to any specific thesis. The Deal Newsletter Filter prompt is the on-ramp: paste a thesis include / exclude list, hand it the email body, and it returns only the rounds that fit — ready to drop into the Deal Pipeline via the Ingest Newsletter Deals automation.
From here
With a Deal Pipeline list populated, deduped on website, and refreshed on a cadence, you have a living ocean. The next module is where it gets interesting: enriching the pipeline — turning thin records into decision-grade objects a partner can act on.