Data & Vendors.

This module walks through the providers we think an AI native firm needs. Cost can and should influence what you pick — it is entirely possible to work within a budget, but you should know what you are trading off. The most powerful datasets are frequently gated behind $20,000+ annual walls, and your operating capital decides whether that math works.

The good news: there are a lot of cheap, fast, powerful wins available now that did not exist two years ago. The goal here is to lay out the universe so you can see the landscape clearly and make easier decisions — not to tell you to buy everything!

Call recordings are some of the most valuable data points a firm generates. They are unstructured firm knowledge that, once transcribed, becomes searchable, summarizable, and feedable into every downstream AI workflow. Recording calls is therefore extremely important.

Some people will be reluctant on privacy grounds. In practice most founders are completely fine with it, and you do not have to record every single call — be selective, especially early. Start with the calls that matter most and expand from there. There will be a lot more to say about how to use this data once you have it.

Out of the box, we recommend Granola. It is the cleanest standalone notetaker for venture — it joins meetings silently, produces high quality transcripts, and makes the recordings easy to search and export. No configuration theater, no CRM glue required to get value. It can also put a privacy sensitive team at more ease because it does not record audio or video and does not even show up as a bot in meetings.

That said, if your CRM ships with call recording built in, use that instead. Attio Call Intelligence and Affinity Meetings both join calls natively and sync transcripts straight onto the relevant record, which removes a whole layer of integration work and keeps your conversation history where your relationship history lives.

You need data on the companies you are evaluating without having to ask for their data room or deck. There are really three options worth considering: PitchBook, Harmonic, and Specter. Pick one, it probably doesn't make sense to stack.

• PitchBook — best for hard data, especially financials. Not always right, and people love to dunk on it, but they have made real improvements and it is the most reliable source for the structured facts that actually matter in diligence. The API is especially useful for building automations and the list creating capabilities offer a good way to build out a universe. This is our recommendation.
• Harmonic — best for AI powered startup identification and sourcing. Strong native sync into Attio, which matters more than people realize when your CRM is the system of record. Their native AI has gotten incredibly advanced and they seem to be the darling / future of the space. A number of the largest and most storied VCs have large contracts with Harmonic.
• Specter — a credible third option in the same category, worth evaluating depending on your thesis, but unclear why a firm would choose this offering over Harmonic.
• Evertrace — best for early stage firms looking for alpha related to building relationships with stealth mode founders. Can directly enrich your CRM, connects natively to Attio. An add on option but not a replacement for one of the first 3 providers.

The next tier is your AlphaSense, Cap IQ, and sector specific research tools. Whether they are worth it depends entirely on how you do diligence and what you invest in.

• Capital IQ — makes sense if you run heavy diligence with a lot of comparable-company analysis. Strong on public comps, transaction data, and the kind of structured financial work later-stage firms live in.
• AlphaSense — extraordinary if you consume a lot of equity research and expert transcripts. Also very expensive. Justifiable for firms whose process genuinely depends on that corpus; overkill for most early-stage shops.
• Obviant & GovTribe — if you focus on defense or govtech, these are genuinely AI enabled tools that make sense of the complicated DoD world. Specialized but invaluable in thesis.
• Wokelo — an agentic research platform that spits out first-draft company, sector, and target reports in minutes. A real time-saver for teams that need to cover a lot of ground quickly without staffing a full analyst on every look.
• Hebbia — enterprise-grade multi-agent research over unlimited document sets with full citations. Built for the kind of document-heavy diligence — data rooms, CIMs, expert transcripts — where the answer has to be defensible.
• Rogo — an AI workspace tuned for finance workflows: comps, model updates, transcript Q&A, and IC-style memos against your own and licensed data. Increasingly common in growth and later-stage shops.
• Auquan — agentic platform that runs end-to-end research and monitoring workflows, from screening through IC memos and ESG / portfolio reporting. Worth a look if you want to automate the full diligence assembly line.
• Termina — on-demand quantitative diligence scans (growth quality, unit economics, PMF, benchmarks) delivered in about a day. A useful outside check on targets, especially when you want investor-grade numbers without spinning up an internal workstream.
• V7 Go — agentic document workflows for diligence: pull structured financials out of CIMs, normalize data-room uploads, and generate IC-ready outputs. Strongest where the bottleneck is unstructured documents at scale.
• Fintool & Quartr — public-markets pair. Fintool is an AI equity research agent across SEC filings and transcripts; Quartr is the cleanest structured access to earnings calls, slides, and filings across 13,000+ companies. Relevant if your thesis touches public comps or crossover work.
• Sacra & PrivCo — private-company research where PitchBook and Crunchbase thin out. Sacra publishes original reports and revenue models on growth-stage and pre-IPO names; PrivCo covers bootstrapped, non-VC-backed U.S. middle-market companies with financial estimates and M&A history.
• Elicit & Consensus — AI search over peer-reviewed literature with citations. The right tool when a deep-tech, life-sciences, or climate thesis turns on what the science actually says rather than what a deck claims.

Happenstance is the cleanest way we have found to query the partnership's collective network with AI. Point it at the team's LinkedIn and Outlook connections and ask it real questions in plain English — who do we know at this company, who could intro us here, who has worked on this problem. The data never leaves Happenstance, which makes the privacy story easy.

Affinity is meaningfully better than Attio at network mapping and best-path intros — visualizing who knows whom, surfacing the warmest route to a target, and keeping relationship intelligence alive across the partnership. Harmonic also offers really amazing network-intelligence capabilities here. For the network-driven VC firm this layer is critical, and it compounds: the bigger the firm, the bigger the universe of connections, and the more likely someone always knows someone.

Enrichment is how you turn a raw list of names or companies into something actionable. The right stack depends on geography, volume, and whether you need emails, phones, firmographics, technographics, or live web signals. Most firms end up with a primary workhorse and one or two point solutions for edge cases.

• Clay — the workhorse for enrichment. LinkedIn profiles, emails, company metadata, intent signals, the long tail of provider lookups you would otherwise stitch together yourself. It is the one tool we reach for whenever a workflow needs “take this list and make it useful.” Pairs naturally with the CRM and with anything you are sourcing or running outbound through.
• Apify — a scraper marketplace with thousands of useful tools others have built that you can easily drop into your workflows for orchestration. Think scraping a LinkedIn profile, extracting a transcript from a YouTube video, or pulling in search results and relevant articles about a company. A powerful complement whenever you need to pull live web data into a process.
• Apollo.io — a vertically integrated GTM platform built on a massive proprietary B2B contact and company database. Self-serve friendly, strong on email sequencing, and a practical all-in-one for teams that want enrichment and outbound execution in the same place.
• Cognism — compliance-first contact enrichment with a strong EMEA footprint and phone-verified mobile numbers. Worth considering if GDPR posture and international coverage matter for your sourcing motion.
• FullEnrich — a waterfall enrichment specialist that cascades 20+ data providers in sequence to maximize verified email and mobile hit rates. Especially useful for niche verticals where any single database has spotty coverage.
• Explorium — a B2B data infrastructure layer aggregating 50+ sources into a single API and MCP server. Good for firms building AI-native enrichment pipelines that need structured company, contact, and signal data at scale.
• BuiltWith — a technographic profiler that reveals the tech stack of any website and generates filterable lead lists by technology adoption. Handy for market mapping and competitive intelligence when a sector turns on specific infrastructure choices.

Feedly is the best and easiest tool we have seen for monitoring news sources, companies in your pipeline, and anything else you need to keep tabs on. You can build AI-powered feeds that surface what matters without the noise, and it is interoperable enough to drop into most workflows. It is also cheap — a rare combination of genuinely useful and genuinely affordable.

For how to put it to work, see the Build a Pipeline News Feed module in the Sourcing playbook, and the Pipeline News Feed automation that pipes Feedly into Notion.

Documentero is the best method by far for putting together templated documents automatically. It is especially useful when you have AI-generated content flowing through a workflow and need to drop it into a polished Word document before sharing it with a team member. No manual formatting, no copy-paste friction — just clean, repeatable document generation. By far the easiest connector within automation platforms like Zapier.

That said, Word documents are increasingly exportable directly from AI platforms, so if the tool you are working in already outputs formatted documents you may not need an extra layer. Working in Claude, for example, you can export to Word easily, and both the Claude and ChatGPT Word integrations are super valuable — every firm should connect them.