Gain

Models do a great deal here, powering search across the graph, market mapping that identifies patterns and likely buyers, watchlists that surface opportunities without being asked, agentic workflows turning data into action, and a prediction of whether a company is receptive to acquisition. The company also describes proprietary language model capabilities.

Against that, the hybrid is stated openly rather than hidden: hundreds of researchers verify profiles by hand, intelligence comes partly from its own origination team and senior operator network, and the platform is marketed as analyst verified. Strip the models and a large human research operation with a proprietary graph on millions of private companies remains, and that is what the business was for its first several years.

Human verification is a property of the data rather than a step around it, with hundreds of researchers focused on the depth of individual profiles and the output described as analyst verified investment grade intelligence, which means a person has checked what the machine assembled before a dealmaker sees it. For a research product that is the control that matters, because the failure mode is a confident fabrication reaching an investment committee.

Newer capabilities push the other way, with agentic workflows turning data into action and watchlists surfacing opportunities automatically, and nothing describes whether agent generated output passes through the same verification layer as the underlying profiles.

The company's central claim is also its principal control: intelligence is human verified, with hundreds of researchers checking profile depth and the output described as investment grade and analyst verified rather than machine generated.

In a private markets research product, where financial data is not filed publicly and errors propagate into valuations, a staffed verification layer addresses the failure mode directly and costs real money to maintain, which makes it a commitment rather than a claim. Supporting evidence includes published analysis of more than ten thousand deals produced with a strategy consultancy.

What is missing is measurement: no accuracy rate, coverage completeness figure or validation result appears, and nothing states how the acquisition receptivity prediction is evaluated against actual outcomes.

One of the strongest customer lists in this index, and it is named rather than described. Private equity users include four of the largest global buyout firms, advisory users include three prominent independent investment banks, and the consulting and accounting side covers all three leading strategy firms and all four of the largest professional services groups.

Penetration is stated in coverage terms rather than counts: trusted across more than a trillion dollars of private capital, used by every one of the top strategy and accounting firms, and by between 70 and 80 percent of the world's twenty largest M&A advisory houses. Headcount passed 130 across six offices on three continents, and recognition includes a data provider of the year award from a private equity trade publication and placement in national and European growth rankings.

This is the sharpest instance in the index of the problem first recorded at ToltIQ, and the vendor has at least named a product against it. Four of the largest global buyout firms sit on one platform alongside three independent investment banks that advise on the same transactions, and each is invited to layer its own notes, custom fields and relationship signals onto shared profiles, so competing bidders and their advisers are contributing proprietary deal intelligence into a common system.

Enterprise grade data governance is offered as one of four named product vectors, which is more than most vendors provide and is the right place to address it. What is absent is substance: no statement of what that governance actually guarantees, whether a firm's contributed intelligence is invisible to rivals, or whether search and watchlist behaviour is isolated between customers pursuing the same target.

No data protection agreement, retention schedule or subprocessor list was located. The primary payload is company and market information rather than consumer records, which lowers the stakes, but two things complicate it: profiles of private companies necessarily carry information about the individuals who own and run them, and the platform ingests each customer's own notes, custom fields and relationship management signals, which is that firm's proprietary deal intelligence rather than public data. European headquarters and offices place the business under a strict regime by default, and nothing published describes how either category is held.

Enterprise grade data governance and security is named as a product vector and no framework, certification, attestation or trust centre was located behind it. That gap is conspicuous given the customer base, since global buyout firms and investment banks impose demanding vendor security requirements and all of them have evidently been satisfied privately, and a platform holding competing firms' proprietary deal intelligence on shared infrastructure is precisely where a published control set would carry most weight.

No supervisor, statute or instrument is named. One capability makes the omission worth noting: a model that predicts whether a company is open to acquisition, sold to the parties who would acquire it, sits close to the territory that market abuse and inside information rules govern, and while inference from public signals is lawful, a platform assembling acquisition receptivity across a trillion dollars of capital operates in a space where the boundary matters. Nothing published addresses it, and no data licensing or professional standard is identified either.

No consumer decision applies and two adapted exposures are worth recording. The first follows from the acquisition receptivity model: a company scored as open to a sale is approached, one scored as closed is not, and the company itself never learns it has been assessed or on what basis, so a model's read of public signals shapes which businesses receive investment interest at all. The second is coverage.

The graph is stated to include every company above ten employees, but private company disclosure varies enormously by jurisdiction, so depth will be far greater in Western European and North American markets than elsewhere, and a business in a thinly documented market is less findable and therefore less likely to be funded. No coverage breakdown or model validation across markets was located.

No guarantee, indemnity or falsifiable accuracy commitment was located, which is notable for a product describing its output as investment grade and used in decisions worth hundreds of millions. Human verification gives a customer a basis for confidence and no stated remedy where a verified profile proves wrong.

The company being profiled has nothing at all: it is researched, scored for acquisition receptivity and presented to potential buyers without knowledge, and no correction route exists for a business misrepresented in a graph that thirteen named institutions are reading.

Sources are described by type rather than named, covering the company's own origination team, senior operator experts, the private equity and M&A ecosystem, and public and local research gathered by its analysts, which tells a buyer the shape of the sourcing without identifying any provider. Language model capability is described as proprietary.

What is not disclosed is any external data supplier behind a graph covering millions of companies, half a million deals and eleven thousand advisers, and no subprocessor list or hosting arrangement was located.

Five distinct consumption modes are published and they include the two dealmakers actually use. Two way synchronisation with relationship management systems keeps research and execution aligned in both directions rather than pushing data one way, a spreadsheet add in meets analysts where the modelling happens, bulk data exports let a firm build its own stack on the underlying graph, interface feeds support automated consumption, and structured data can be piped directly into a customer's own language model workflows.

Integration is treated as one of four named product vectors rather than as a feature list, and the stated ambition is to replace more than twenty point solutions with one system, which only works if the connections are real.

No hosting provider, region selection, residency commitment or private deployment option was located. Offices span the Netherlands, the United Kingdom, Germany, Poland, the United States and India, and customers include European institutions whose own arrangements would normally require a stated position, particularly given that the platform holds each firm's proprietary deal notes alongside public market data.

No pricing, packaging or basis of charge is published. The product is structured across four distinct vectors covering data, integration, agents and governance, which almost certainly price separately, and bulk data and interface feeds are offered alongside seat based platform access, so the commercial shape is visibly complex and entirely undescribed. Nothing indicates whether charge scales with users, data volume, coverage or modules taken.

Buyers span private equity investors, M&A advisers, corporate development teams, strategy consultancies and professional services firms, with offices across Europe, North America and India supporting a genuinely global footprint. Functional coverage runs the full deal lifecycle from sourcing and market mapping through research and diligence to execution and portfolio tracking, which is why the company describes itself as replacing twenty or more point solutions.

Two limits hold this at B: the financial institution portion is private capital and investment banking specifically rather than the wider institutional market, and a substantial share of the named customer base consists of consultancies and accounting firms, which are advisers rather than institutions.