73 Strings

The removal test separates this vendor from its extraction peers. Take the models out of the extraction module and documents have to be keyed by hand, but take them out of the valuation module and a working product remains: discounted cash flow modelling, market comparables, benchmarking, scenario periods, version control and an audit trail are finance and software engineering rather than model work, and that engine was built by valuation practitioners.

The company's own chief executive frames it the same way, describing the data substrate as the thing that makes the models trustworthy rather than the other way around. Models compress analyst effort on top of a product that would still function without them, which is the established B pattern in this index.

Human judgement is designed in rather than bolted on, which is appropriate given the output is an opinion of value. The company states that automated outputs including data sources, peer comparisons and the valuation arithmetic are presented in standardised formats for human verification, that analysts can override assumptions such as discount rates and growth projections in support of the reasonableness testing international valuation standards require, and that a full audit trail and version control sit underneath. What is missing for a higher grade is any account of who is permitted to override, whether overrides are flagged as such downstream, and whether any sampling or independent check runs across them.

Traceability is well covered and validation is not. The company publishes an audit trail, version control, standardised presentation of automated outputs and full data lineage, all of which let a reviewer follow how a number was produced. None of it establishes whether the number is right. No extraction accuracy rate appears, no error rate by document type, no drift monitoring and no model documentation.

The absence that stands out most in this particular domain is back testing: private markets valuations are eventually settled by realised exits, which makes marks uniquely checkable against outcomes, and no published material compares the platform's valuations to what assets actually sold for.

Two clients are named publicly with their own scale attached, a French listed investment company at roughly thirty five billion euros and a Belgian holding company at over nine billion, and the company states its client base collectively manages close to ten trillion dollars across private equity, growth equity, venture capital and private credit.

A case study with one of the largest alternative managers in the world exists and a global head of financial affairs at a thirty one billion dollar private equity fund is quoted on outcomes, though that firm is not named. What holds this below an A is the shape of the numbers: the speed claim is a round multiple with no client attached, and the scale figures describe the buyers rather than the result. Prestige backing does not lift this axis.

The company markets a connected client community in which clients benefit from shared insight across some of the world's leading alternative asset managers, and sells access to market comparables, performance metrics and investor benchmarks inside the same platform.

Those two facts raise the obvious question of whether benchmark and comparable data is derived from client portfolios, because the material a general partner uploads is the unpublished financial performance of companies it competes to buy and sell. Nothing published states where client data stops and platform data begins, whether contribution is optional, or whether a manager can see comparables drawn from a rival's book. This is the second vendor in this lane with an unresolved network effect claim.

The data class is portfolio company financials, fund level performance and investor reporting rather than consumer financial information, so the statutory privacy exposure that shapes this axis elsewhere is largely absent and the grade reflects what is published rather than a penalty for the category.

Searched the platform, security and company pages for a privacy statement covering handling of client and portfolio company information, retention periods, deletion on exit or processing locations, and located none. For a platform holding the unpublished financials of thousands of private companies that never agreed to anything with this vendor, that silence is worth naming.

One attestation is referenced and it takes finding. A product page answering questions about audit readiness states that the extraction audit trails are covered by a service organisation control type two certification, which is a named attestation rather than a vague assurance and lifts this above the vendors in this lane that publish nothing at all. It stays at C for three reasons.

The reference sits inside a product answer rather than on any security or compliance page, no trust centre, report request route or scope statement accompanies it, and no second standard appears beside it. For a company whose clients manage close to ten trillion dollars and whose platform holds the unpublished financials of thousands of private companies, a single passing mention is a thin public security position.

No licence is held or required and the company states its regulatory position more clearly than most in this index, positioning the product against international valuation standards, fund reporting obligations and audit scrutiny, and describing outputs as audit defensible.

It goes further and claims to be the only solution in its category carrying no independence or service conflicts, which is a substantive regulatory positioning statement in a field where valuation advisers have long been criticised for selling advice and valuations to the same client. That claim sits awkwardly beside the company's own valuation specialist team delivering expert work alongside the software, and nothing published reconciles the two. Independence of the valuation function remains the manager's duty, not the tool's.

A protected class frame does not fit a valuation platform and forcing one would miss the exposure entirely. The bias that matters here is directional. The party operating the model is the party whose management and performance fees are computed from the resulting net asset value, and the platform openly supports overriding the discount rate and growth assumptions that drive that number. Any systematic drift toward optimistic marks is a transfer from investors to managers.

Nothing published states whether overrides are aggregated, compared against realised exit outcomes, or disclosed to limited partners and auditors. A second and quieter version of the same problem is that comparables drawn from a client pool weighted to large managers will fit small and non United States assets worst.

The published promise is high and the published commitment behind it is absent. Outputs are described as audit ready, defensible and results you can trust, and they land in fund financial statements, capital account statements and fee calculations, so an error travels straight into audited numbers and investor reporting.

Nothing states a service level, a warranty, a correction obligation or a remedy, and nothing states the opposite either, meaning no disclaimer was located clarifying that the valuation conclusion remains the manager's own. The audit trail makes an error traceable after the fact, which is useful and is not recourse.

Large language models are acknowledged in general terms as the foundation for agentic capabilities on the roadmap, and that is where the disclosure stops. No provider is named, no distinction is drawn between models built in house and models called from an external service, and nothing states whether client documents or portfolio company financials pass to a third party during processing.

The contrast inside this lane is direct: a close competitor answers the same question explicitly and by name, which shows the disclosure is possible rather than commercially impossible. Until it is made, an institution cannot tell which counterparties its confidential portfolio data actually reaches.

The integration story is internal rather than external. Extraction, monitoring and valuation run as three modules over one private capital data model, which is a real architectural claim and the reason time series stay consistent from one reporting cycle to the next, and the platform ingests financial and operational inputs from multiple sources and exports into presentation formats used for investment committee and investor reporting.

What is not published is a named connection to anything a client already runs. No fund accounting system, administrator, general partner portal, market data provider or accounting platform is identified as an integration partner, which leaves a buyer unable to judge how the platform reaches the systems holding the rest of its record.

Delivery is cloud hosted software and the published material says little beyond that. Searched for hosting regions, residency commitments, single tenant options and any account of where processing occurs, and located none. That gap carries weight for this particular client base, which includes European managers subject to their own data location expectations and Middle Eastern sovereign investors who commonly impose them by contract, and the company reorganised in 2026 specifically to serve those regions from a single global model.

Every route through the published material ends at a demo booking. No rates, no tier ladder, no unit of charge and no indication of whether pricing follows seats, funds, portfolio companies or valuations produced, which matters because those units scale very differently for the same firm.

The company also pairs software with a team of valuation specialists, and nothing distinguishes what is licensed from what is delivered as expert work, so a buyer cannot see where the software price ends and the service price begins.

Coverage runs both sides of the private capital relationship and across the asset owner tier above it. The published client base spans pension funds, sovereign wealth funds, limited partners and general partners, including multi strategy alternative managers, across North America, Europe and the Middle East, with client assets stated at close to ten trillion dollars.

Strategy coverage is enumerated rather than implied, taking in private equity, private credit, venture capital and infrastructure, and the company reorganised in 2026 around a single global operating model to serve those segments consistently across regions. Breadth inside private markets is genuine, and the honest boundary is that it does not extend to banks, insurers or retail institutions.