IThe 30-Second Verdict
Grata is one of the strongest crawl-based company-search tools on the market — and a poor fit for a pure venture-stage thesis. Both of those are true at the same time, which is exactly why the one-line takes never quite land.
I build sourcing and thesis-scoring systems for funds, so the question I care about isn't "is Grata good software" — it clearly is — but "does it fit the shape of what a VC is actually hunting?" Pull apart what Grata claims, what its buyers consistently report, and how its index is actually built, and the answer splits cleanly. Grata is genuinely excellent at surfacing established, under-the-radar private companies that structured databases miss. It gets noisier, and structurally mismatched, the moment the thesis is young, venture-stage startups.
The verdict, plainly: if you source lower-middle-market or PE-style deals, Grata earns its (steep, quote-only) price. If your thesis is seed-stage venture, you'll spend more time filtering than sourcing — and tools like Harmonic, Synaptic, or Specter fit the shape of what you're hunting far better. Grata was the one deep-dive we didn't isolate in our four-way breakdown of Grata, PitchBook, Harmonic, and Specter, so here it is on its own.
IIWhat Grata Actually Claims to Do
Grata's pitch is that its search understands what you mean instead of reacting to the keywords you type. The company calls its newest approach Agentic Search — software that, in Grata's words, "actively interprets intent, explores options, and refines results on the user's behalf," adjusting its own queries and testing assumptions as it goes rather than waiting for you to add another filter.
In plain language: with the older style of search, every question you asked required another question. You'd search, skim, tighten, search again. Grata's claim is that the agentic version does that tightening for you — ask for "founder-owned HR services companies" and it plans and runs the follow-on refinements itself instead of handing the next step back to you.
Grata backs this with its own numbers. In a benchmark comparing its newer Intelligence Search against its earlier agent approach on the same company-search test, the newer system scored 0.684 on a ranking-quality measure (NDCG@50 — essentially, how well the genuinely relevant companies cluster near the top of the list) versus 0.526 for the old one, about 30% better. It also answered in 5.6 seconds instead of 12.5, and Grata reports that identical search plans produced identical results across all nine groups and 36 runs in its consistency tests.
Take that for what it is: vendor-published, on a vendor-built benchmark. It tells you Grata's engineering is getting faster and more consistent. It doesn't tell you whether the tool matches your thesis — which is the only test that matters. So the rest of this piece is about fit, drawn from what independent buyers actually report and how the index is built underneath the pitch.
IIIWhere a Vendor Demo Stops Being Useful
Every sourcing tool looks brilliant in a demo, because the demo query is chosen to look brilliant. "Show me vertical SaaS companies in the Southeast" returns a crisp, impressive page. A real fund thesis isn't that. It has edge cases, carve-outs, a stage ceiling, two continents, and a definition of "climate" or "fintech" that's obvious to the partners and invisible to a crawler. The gap between the clean demo query and the messy real thesis is exactly where a tool's true fit lives — and it's the gap a sales deck is built to hide.
That's the lens to judge Grata through: not "does the demo look good" — it does — but "what happens when someone phrases a real thesis, with its exclusions and stage limits, and sees what comes back?" Enough independent buyers have run that experiment that the pattern is clear and consistent. Here's what they report.
IVWhere Grata Is Genuinely Sharp
Grata's edge is real, and worth being specific about. Grata runs natural-language, machine-learning search over a web crawl of tens of millions of private companies — it reads company websites, not a registry of who raised money. So it surfaces the companies that structured databases miss entirely: bootstrapped software vendors, regional service firms, family-owned manufacturers with real revenue and no venture backing.
For the parts of a thesis that are company-shaped rather than round-shaped — "a business doing X in Y region" — this is where Grata shines. It finds operators with a thin public footprint, no Crunchbase profile, and no press, pulled straight from how they describe themselves on their own sites. If your definition of a great lead is "a company nobody else has found yet," Grata's crawl is a legitimately strong way to find it. That strength is reflected in its review standing, which I'll get to under pricing — the people buying Grata are mostly buying it for this job, and mostly happy they did.
VWhere Buyers Hit Noise
The recurring complaint in buyer reviews is "result noise," and it's the other half of the honest picture. A verified G2 reviewer describes it exactly: "Sometimes the keyword-based search can surface companies that are not fully relevant to the intended criteria," which means results "occasionally require additional manual filtering and review to remove less relevant targets." Independent analysis puts it more bluntly: broad queries "pull in nearby businesses that you have to prune by hand." The tool, as that same review sums it up, "narrows the field. It does not hand you a clean list."
In practice, "nearby" means companies that share your thesis's vocabulary without sharing its substance — the right sector at the wrong stage, the right stage in the wrong sector, or an adjacent service business that merely sounds like a fit. None of it is random; it's all plausible, which is precisely what makes pruning it slow. And getting the semantic search to return what you actually meant takes practice — there's a real learning curve to phrasing a thesis in a way the model reads the way you intended. This isn't a tool you buy on Monday and trust on Tuesday.
aThe Stage Mismatch: Built for Lower-Middle-Market, Not Venture
Most of the noise buyers describe traces back to one structural fact: Grata was never built for venture. It was founded in 2016 to cover lower-middle-market companies with real revenue but almost no public data trail — exactly the firms a PE or corp-dev team hunts. The same roundup is direct about the mismatch: if your "deal thesis centers on venture-stage startups or large-cap public companies," you should "pick something else," because "Grata's lower-middle-market focus will miss most targets."
This matters more for a VC than it first looks. Grata's index is built around established, revenue-generating private companies — and most of the private market never touches venture money at all. A seed-stage venture thesis is chasing the opposite kind of company: young, often pre-revenue, defined by its round and its team more than by a decade of web presence. So the crawl is strongest on exactly the companies a venture thesis is least likely to be hunting. For PE and corp dev, that's a feature. For a seed fund, it's the root of the noise.
VIWhat Grata Actually Costs
Grata doesn't publish pricing — every number below is buyer-reported, and your quote will depend on seats, data access, and CRM integrations. Treat it as a range, not a rate card.
Those figures come from a roundup of buyer-reported pricing, which also notes Grata's review standing: 4.8 out of 5 on G2 across 79 reviews, and 4.3 out of 5 on Capterra across 9. People who buy it mostly like it — which fits, because the people buying it are mostly buying it for the job it's good at.
The CRM integrations are a real part of the value at that price. Grata connects natively with Salesforce, HubSpot, Affinity, Pipedrive, and DealCloud, across two tiers: a basic tier that pushes company and contact records and filters out companies already in your pipeline, and an advanced tier that syncs automatically and pulls up to 25 custom CRM fields back into Grata. If you're going to prune a lot by hand, having that pruning flow straight into your CRM is worth something.
One honest caveat on the data: contact data is reported to be thinner than the company data, coverage is strongest in North America and thins out internationally, and the opaque pricing makes comparison shopping slow. Budget for a second vendor if reaching people — not just finding companies — is part of the job.
VIIThe Acquisition Context: Grata Is Now Part of Datasite
If you're evaluating Grata today, you're not really evaluating the 2016 startup. Grata was founded in New York and had raised $35M across three rounds before Datasite acquired it in June 2025, with Datasite's controlling shareholder committing $500 million to expand its intelligence solutions. Co-founders Andrew Bocskocsky and Nevin Raj stayed on to run Grata as a business unit inside Datasite.
The data footprint has moved fast since. Datasite acquired deal-sourcing provider Sourcescrub in August 2025 and folded its data — drawn from more than 220,000 sources covering 16 million companies — into Grata. Then, in May 2026, Datasite acquired the Swedish private-market firm Valu8, pushing Grata's European coverage to 70 million companies, on top of its July 2025 expansion into the UK, France, Germany, and Australia.
Read the direction of travel. Datasite is a deal-execution business, and the roadmap is pointed at M&A and deal-workflow depth, plus broader international company coverage — not at tuning the index for venture-stage startups. If anything, every acquisition deepens the thing Grata is already good at and widens the gap with what a VC thesis needs.
VIIIThe Verdict: Who Should Actually Buy Grata
Grata is a buy — for the right desk. If you're in PE, corp dev, or a lower-middle-market-focused fund hunting established, under-the-radar private companies, the crawl is genuinely differentiated and the price is defensible. Grata's own pricing page leans on exactly these buyers, citing outcomes like a firm that "sources 15 deals in its first year" and another that "saves 3 hours per day" on sourcing and diligence — PE-style shops, doing PE-style sourcing.
For a pure venture-stage thesis, I'd pass. Not because Grata is bad, but because you'd be paying lower-middle-market money to fight a stage mismatch, then hand-pruning the difference every week. For that job, the venture-native tools are closer to the shape of what you're sourcing — see Specter's own honest verdict and what Synaptic actually does for venture-stage sourcing. Match the tool to the thesis, not to the demo.
IXFAQ
Can a sourcing tool's output be scored against our fund's specific thesis automatically?
Yes — this is a lot of what I build. Instead of pruning a noisy list by hand every week, we wire a thesis-scoring step into your existing dealflow: every candidate company, inbound deck, or sourcing export gets scored against your fund's actual criteria and routed by the result. I've written up the approach in scoring every inbound deck against your fund's own thesis. The point of a scoring layer is to turn "Grata narrows the field" into "here's the shortlist, ranked."
Do we have to replace our CRM to get better sourcing?
No. Black Matter VC is an AI studio and consultancy — I build and operate custom automation on top of the stack you already run, whether that's Affinity, Attio, HubSpot, or Salesforce. The sourcing and scoring layer sits on your CRM; it doesn't ask you to migrate it.
How does Black Matter actually work with a fund?
It's a build-and-operate retainer: $10k/month to design, ship, and run the AI systems your fund needs — dealflow capture, thesis scoring, portfolio monitoring, LP reporting — as a solo, operator-led studio rather than a software license you have to staff and maintain yourself. You keep the tools you like, including the sourcing vendors; I make them do more of the work.
— Michael Rouveure