FNZ
FNZ is a global wealth management platform founded in 2003 in New Zealand by Adrian Durham and headquartered in London, providing end to end infrastructure spanning custody, trading, advice, reporting and compliance to financial institutions rather than to advisers directly. It reports more than 650 financial institution partners, over 26 million end investors and close to two trillion dollars of assets on platform, with roughly 6,000 staff across more than 30 countries and annual revenue above 1.4 billion dollars. Ownership is institutional, with CDPQ, Generation Investment Management, CPP Investments, Motive Partners and Temasek among its backers.
Founder Adrian Durham moved from group chief executive to a non executive role in 2024 and Blythe Masters was appointed group chief executive, with Roman Regelman as group president. FNZ Advisor AI, launched in August 2025, is a generative model product embedded directly in the platform that prepares meetings, drafts client follow ups and surfaces next best conversation prompts, described by the company as the first in a planned series.
It rests on a five year global partnership with Microsoft that places Azure AI Foundry at the centre of the platform and brings Azure OpenAI, Microsoft Fabric and Microsoft 365 Copilot into both client facing and middle and back office workflows. Named clients include Raymond James.
Capability Axes
Capability grades
15 of 15 axes rated · 5 graded A or B
A platform trading since 2003 whose first model product shipped in August 2025 and is described by the company itself as the first in a planned series. Strip it and everything remains: custody, trading, advice, reporting and compliance infrastructure carrying close to two trillion dollars for 650 institutions.
Included on the Clearwater and MyComplianceOffice precedent because the model output feeds a regulated advice interaction rather than sitting beside it, generating the preparation and the conversation prompts that shape what an adviser puts to a client. The same reasoning built InvestCloud's meeting product in this pocket.
No gate, threshold, confidence measure, escalation path or sampling audit is described, and no sequencing statement was located from the vendor beyond the product being positioned as assistive to advisers. An independent review refers to embedded governance safeguards as a differentiator, which is an evaluator's summary phrase with no described content behind it.
That places this at the Avaloq position rather than the Ruleguard one: the oversight claim is carried by a third party and asserted as an outcome, not described as a control. A system drafting client communications and prompting what an adviser says next is shaping a regulated interaction and warrants more.
No accuracy, precision or recall figure, benchmark or validation method was located, and none of the four properties this index accepts as evidence of model risk discipline is present. What makes this entry unusual is that an independent evaluator reached the same conclusion in public, identifying limited visibility into performance benchmarks and uncertainty about whether the product adapts across different adviser segments as the central open questions. When a paid analyst reviewing the launch cannot find the numbers either, the gap is a disclosure choice rather than an artefact of how this index searches.
Named institutional clients and an independent analyst review of the model product specifically, which is more scrutiny than most of this pocket attracts. Off an A because no quantified outcome attaches to any named client, and the analyst review makes the gap explicit rather than leaving it to inference: it identifies limited visibility into performance benchmarks for validating return on investment as a central question for enterprises evaluating the product. That is the first time in this sweep an independent evaluator has named the measurement gap as a buyer risk rather than the index inferring it.
The most consequential stewardship position located in this index, because the vendor advertises the thing every other vendor leaves unsaid. Its own launch material states that with more than 650 institution partners, over 26 million end investors and close to two trillion dollars on platform, it provides access to one of the largest wealth management data sets in the world, and that this is what enables the model product to support smarter and faster decision making.
Scale across competing institutions is presented as the product's advantage. No boundary statement accompanies it: nothing describes whether one institution's client data informs anything served to another, whether models are trained or tuned per client, or what an institution can require to be excluded. The question is not implied here, it is raised by the vendor and left open.
No privacy programme, retention position, data processing terms or subject rights framework was located. The scale makes the absence unusually consequential: this platform holds the holdings, transactions and personal circumstances of more than 26 million identified individuals, almost none of whom have any relationship with the vendor or would recognise its name, and a model product now reads across that record to generate communications about them.
No certification, attestation, report type, audit scope, penetration testing summary or trust centre was located. An administrator holding close to two trillion dollars of client assets and the records of 26 million investors will hold formal credentials and will produce them in a procurement process, and none of them are published where a buyer or an end investor can find them. Temenos and Avaloq, the two closest peers by scale in this pocket, both publish theirs.
Materially different from the pure software vendors in this pocket. The group delivers custody, administration and tax wrapper services alongside the platform, which are authorised activities in the markets it operates rather than technology supply, so it sits inside the regulatory perimeter its clients occupy rather than beside it.
Off an A because no specific authorisation was enumerated in the material located and, more importantly, no regulator has supervised, tested or admitted the model layer itself, which is the bar CleverChain set.
No framework, bias testing, fairness evaluation, model documentation or independent assessment of an artificial intelligence management system was located. The company publishes substantial thought leadership on artificial intelligence in wealth management, including a large multi market study conducted with a research firm, and calls for responsible deployment across the value chain.
Publishing research about how the industry should govern models is not disclosure of how this vendor governs its own, and the same pattern is recorded against ACA Group and Luthor elsewhere in this index.
Nothing published describes liability, indemnity or recourse when generated material misstates a client's position or a prompted conversation leads an adviser to an unsuitable recommendation. The layered structure makes the question harder rather than academic: the model runs on a third party's artificial intelligence platform, inside a vendor's administration service, used by an adviser at an institution, and reaches an end investor who is four parties removed from the model and has no way to know it was involved.
Among the most specific infrastructure disclosure in this index, and it follows the pattern recorded against InvestCloud: the vendors that buy their model capability name their suppliers, and the vendors that build proprietary ones name nothing. The partnership is public and itemised, naming the hyperscaler's artificial intelligence platform at the centre of the stack, its hosted model service, its data fabric and its productivity agents. Off an A because no base model, version or hosting region is named for the adviser product, so an institution still cannot record which model version produced a given client communication.
The platform is the system of record rather than a layer above one, spanning custody, trading, advice, reporting and compliance end to end for 650 institutions, which is the deepest position in this pocket alongside the core banking vendors.
Integration extends outward as well: a five year partnership places a hyperscaler's artificial intelligence platform at the centre of the stack, brings its data fabric and productivity agents into middle and back office operations, and distributes modular wealth components through that partner's marketplace.
No region choice, hosting location, tenancy model, residency commitment or subprocessor list was located, despite the platform operating in more than 30 countries where residency and outsourcing rules differ materially. The hyperscaler partnership establishes which cloud the model layer runs on and says nothing about where, which is the distinction that matters to a regulated institution assessing an outsourcing arrangement.
No pricing, band or rate is published for the platform or the model product. The commercial model is an institutional platform contract typically priced on assets and transaction volume, negotiated per client over multi year terms, and nothing public indicates whether the model product is bundled into platform economics or charged separately. Group revenue is disclosed, which speaks to vendor scale rather than to cost.
More than 650 financial institution partners, over 26 million end investors and close to two trillion dollars of assets on platform, delivered by roughly 6,000 staff across more than 30 countries. Segments run from private banks and wealth managers to retail investment platforms, retirement schemes and institutional distributors, and the model product is offered globally rather than in one market first. Coverage is anchored to counted institutions, counted investors and an audited scale of assets under administration rather than asserted.
Alternatives to FNZ
The closest documented capability profiles to FNZ in the same categories, ordered by similarity across the same fifteen axes the index grades every vendor on. Closest documented profile, not a claim that either product does the same job. No vendor pays for placement.
Stronger documented coverage on Operational and Outcome Evidence
Documents Autonomy and Oversight Model where FNZ does not
Documents AI Centrality where FNZ does not
Documents Deployment Model and Data Residency and Security Certifications and Trust Center where FNZ does not
Documents AI Centrality and Commercial Transparency, among others where FNZ does not
Documents Autonomy and Oversight Model and Deployment Model and Data Residency where FNZ does not
Similarity is computed axis by axis from published grades, not from a composite score. The index does not aggregate grades into a total. See the fifteen axes and the methodology.
Pricing
Vendor-published figures are labeled as such. Figures labeled “Estimated” are derived from third-party sources and have not been confirmed by the vendor.
No pricing data has been verified for this vendor. Pricing information will be published here once confirmed through vendor disclosure or third-party estimation.