Flanks

Models do the work the company itself identifies as the hard part: extracting holdings from non standardised documents, reports and statements where private equity, real estate and art positions exist only in unstructured form, alongside newer analytical tools for portfolio analysis and advisory workflows.

The company is explicit that liquid account data sitting in different portals could be handled with basic scraping or a simple feed, and that unstructured alternatives are the real bottleneck. Against that, the asset is the connection network itself, more than 600 links to institutions across 33 countries built and maintained under a regulatory authorisation, which is engineering and licensing work rather than modelling and is what a bank actually buys. This is the Sikoia position.

Validation is a named function rather than an implied one, with the platform stated to manage aggregation, validation and enrichment while ensuring full control and data integrity, and reconciliation across sources means figures are checked against each other rather than accepted. That is the right control for a data layer, because the failure mode is a wrong number propagating silently into advice.

Output is structured data and reporting for an adviser to act on rather than a recommendation or an instruction, so nothing is decided on a client's behalf. What is absent is any description of what happens when validation fails, whether an adviser is told a position could not be reconciled, or how uncertainty in an extracted alternative asset value is surfaced.

Validation and reconciliation are built into the product rather than promised around it, with data standardised, cross checked and enriched before it reaches an adviser, and data integrity named as an explicit objective. For a data infrastructure business that is the correct control, because reconciliation against the source is the only way to catch an extraction error before it becomes a reported valuation.

Scale gives some assurance too, since half a million portfolios monthly across 600 connections means format breakages surface quickly. What is missing is measurement: no extraction accuracy, reconciliation break rate or coverage completeness figure is published, and for alternative assets read from documents that is the number an institution would most want.

Scale is stated with unusual precision: more than 600 connections to financial institutions, over 500,000 investment portfolios processed monthly across 33 countries, 40,000 users and more than 39 billion euros in tracked assets.

One named customer is decisive, since one of Europe's largest banks selected the platform to supply the automated connectivity, daily collection and standardisation behind its new multi custodian aggregation service for private banking clients, and that data feeds the bank's wealth management platform running on the world's dominant portfolio risk technology. Total funding stands at around 23.55 million euros, with a Series B in February 2025 led by a specialist fintech investor alongside a major growth fund. Headcount is between 51 and 100.

No data boundary statement was located. More than 500,000 portfolios pass through the platform monthly on behalf of banks, family offices and advisers that compete for the same clients, and extraction models improve with exposure to more custodian formats and more alternative asset documentation, so what one institution's data teaches the system is directly valuable to the next. Nothing states whether extraction learning is shared, whether a client can decline to contribute, or how holdings data is separated between institutions serving overlapping households.

The privacy position rests on a licence rather than an assurance, which is the strongest form available. Authorisation as an account information service provider means data access operates under the statutory consent framework, so the client authorises each connection and can withdraw it, and the subject is a party to the arrangement rather than an object of it. Supervision by a national central bank under European Central Bank oversight adds a standing examiner.

Held at B because no retention schedule, subprocessor list or deletion commitment was located, and the platform assembles the most complete picture of a wealthy household's holdings that exists anywhere, across banks that individually see only their own part.

No attestation, certification, trust centre or enumerated framework was located. Regulatory authorisation brings supervisory requirements around systems and controls that an examiner can inspect, which is a genuine floor and not the same as a published assurance artifact. One of Europe's largest banks has completed vendor assessment on this platform and connected it to its private banking client data, so the review has been passed at a demanding standard privately.

The vendor's own authorisation is published with its supervisor named, which is the Sikoia position and the strongest form this axis recognises: the company operates as an account information service provider regulated by the Spanish central bank under European Central Bank oversight, so its data access rights derive from a licence a reader can verify rather than from a contract.

Four European instruments are additionally named as the market context, covering the current and forthcoming payment services directives, the open finance framework and the anticipated revision of the investment services regime, with the company's own strategy officer explaining that the latter will likely require advisers to account for a client's pre existing wealth before giving advice, which is the regulatory case for the product stated precisely.

No consumer credit decision applies and two adapted points are worth recording. The inclusion argument is real and structural rather than aspirational: advisers have historically worked without knowing a client's global asset allocation, and advice given on partial information is worse advice, so consolidating holdings improves the quality of what a client receives, with the company framing its purpose as extending wealth services to a broader base of affluent investors.

Against it, extraction quality will not be uniform. Alternative assets arrive as non standardised documents, coverage across 600 institutions and 33 countries will be deeper in some markets than others, and a household whose holdings sit in less well covered institutions or harder to read formats receives a less complete picture without being told it is less complete.

No commercial guarantee or indemnity was located, and the recourse that exists is statutory and therefore harder to withdraw. Authorisation as an account information service provider brings consent and withdrawal rights over the client's data and places the company under a supervisor with examination powers, so a client retains control over the connections feeding the picture built about them.

Reconciliation means an erroneous position can in principle be traced to the source that produced it. What is missing is anything describing correction where an extracted alternative asset valuation proves wrong, or notification of an adviser whose reporting rested on it.

The data chain is disclosed by category and quantified, covering more than 600 connections spanning banks, brokers, custodians, pension providers and alternative asset platforms across 33 countries, accessed through multiple secure methods, so a buyer understands the breadth and the mechanism even without individual names. The downstream consumer is named, which is unusual and useful.

What is not disclosed is the model layer, with no provider identified for the document extraction that handles alternative assets or for the newer analytical tools, and no subprocessor list or hosting arrangement located.

The integration surface is the product and it is evidenced at both ends. Upstream, more than 600 connections reach banks, brokers, custodians, pension providers and alternative asset platforms across 33 countries, established through multiple secure methods rather than a single technique, which is what allows coverage where open banking rails do not reach.

Downstream, the most significant disclosure in this profile: aggregated data feeds a major European bank's wealth management platform, which itself runs on the dominant institutional portfolio and risk technology, so this company's output is consumed inside the system most large asset managers use. Being named as a supplier into that stack is a stronger integration statement than any connector list.

No hosting provider, region selection, residency commitment or private deployment option was located. Operating under a European authorisation with central bank supervision implies processing within that regime, and it is inferred rather than stated, and a platform holding consolidated household wealth data for private banks across 33 countries would ordinarily be expected to publish where that data rests.

No pricing, packaging or basis of charge was located. The platform is explicitly modular, spanning aggregation, reporting, portals and analytics, which implies component pricing, and connection breadth across 600 institutions suggests cost may scale with coverage. Nothing indicates whether charge falls per portfolio, per connection, per user or as an enterprise licence.

Six buyer types are served through the same infrastructure, spanning private banks, retail banking groups, family offices, asset managers, independent advisers and wealth managers, across 33 countries with more than 600 institutional connections.

Asset coverage is the second dimension and the harder one, reaching beyond liquid holdings at banks and brokers into pensions and alternative assets including private equity, real estate and art, with both financial and non financial assets consolidated. Function spans aggregation, reconciliation, enrichment, reporting, simulation, client portals and analytics, so an institution can take the data layer alone or the whole platform.