ForwardLane

Models are the product and they are the company's own. Proprietary natural language processing built specifically for financial services enriches aggregated data, a composite generative model answers questions over the resulting book of business, and the underlying capability is described as pre-trained on wealth management, asset management and insurance domains rather than adapted from a general model. The company states it has been applying AI since 2016 and holds patents pending on the insight platform. Remove the models and the virtual analyst reviewing every client record daily does not exist.

Output is recommendation rather than action, surfacing signals and next best actions that tell an adviser which clients to contact, what to say and crucially why, with the reasoning attached rather than the conclusion alone. Nothing is executed on a client's behalf and the adviser remains the decision maker throughout.

Held at B because no review threshold or escalation rule is described, and a system that ranks an entire book daily shapes which clients receive attention long before any adviser judgement is applied.

Two mechanisms support verification rather than being claimed in the abstract: a full audit log is generated for every interaction, and the service explains how answers are derived so the client has a transparent view of the process rather than an unexplained conclusion. Insights are created from the institution's own data through a point-and-click interface, which keeps the construction visible to the business user. Held at B because no accuracy measure, evaluation approach or validation method is published for the proprietary language processing or the generative layer.

Client institutions are described as large financial institutions holding 2.5 trillion dollars in assets under management, which is a substantial aggregate even without individual names, and the company appears in two independent industry rankings for wealth technology and applied AI in financial services plus two analyst research briefs. A major cloud provider granted it co-sell ready status, which requires a solution review rather than a marketing agreement. Held at B because no institution is named and no adoption, retention or outcome measure accompanies the assets figure.

An explicit boundary is published: data on the platform remains private and is not shared externally, and the deployment model reinforces it, since an institution can run the platform on its own cloud or take it white-labelled rather than sending client books to a shared environment. That is more than most peers state.

Held at B because the statement covers external sharing without addressing whether the company's own models learn from customer data, which matters given the platform is described as pre-trained on these domains and improved through accumulated question sets across clients.

The privacy position is stated as a property of the system rather than a policy, with data on the platform remaining private and not shared externally, and deployment available in the institution's own cloud so client records need not leave its control. The company frames data transparency, privacy and security as the specific enterprise problems its generative platform was built to solve. Held at B because no data processing agreement, retention schedule or subprocessor register was located.

No attestation, certification, trust centre or enumerated control set was located. Institutions holding trillions in assets have completed supplier assessment before connecting client books, and a major cloud provider has reviewed the solution for co-sell status, so assurance exists privately while nothing is published for a prospective buyer.

No regulator, statute or supervisory framework is named. Audit logging is described as generated to ensure regulatory compliance, which asserts the purpose without identifying the requirement, and for a platform driving adviser outreach and product recommendations across wealth and insurance there are suitability, marketing and record-keeping obligations that go unmapped.

No fairness testing or governance disclosure was located, and the adapted exposure is about attention rather than approval. The platform ranks and scores an entire client book by growth opportunity, wallet share potential and retention risk, which means it allocates finite adviser time toward the most commercially valuable relationships and away from smaller ones. That is a rational commercial design and it is also a distributional decision made by a model, and nothing describes how the ranking is constructed or whether its effects are examined.

No guarantee, indemnity or correction process was located. The end client is unaddressed and largely invisible to themselves here, since the person affected is an investor or policyholder whose position in an adviser's outreach ranking was set by a model they will never see, and nothing describes whether that scoring is disclosed, reviewable or contestable.

The dependency position is clearer than most because the intelligence is built rather than licensed, with proprietary language processing and a composite generative model described as the company's own and pre-trained on its three target domains, and the cloud platform it is built on is named directly. More than 300 data sources feed the aggregation layer.

Held at B because none of those data sources is identified, no base model family or version is disclosed behind the generative layer, and knowing a model is in-house is not the same as being able to document it.

The integration thesis is correct for this buyer: insights are delivered into the adviser's existing customer relationship system rather than a separate application, with a major customer platform named as a key partner and interfaces available for routing next best actions into workflow. The platform aggregates from more than 300 internal and external data sources and is listed on a major cloud provider's application marketplace. Held at B because beyond that one customer platform, no portfolio, custody or advisory system is named individually.

Two deployment paths are offered explicitly, with the platform running on the institution's preferred cloud or hosted by the vendor, and available white-labelled, which the company connects directly to giving the customer full control and meeting its data requirements. That answers the residency question by construction for institutions that need it. Held at B because no regions are named for the hosted option and no residency commitments accompany it.

No pricing, packaging or basis of charge was located. One procurement route is disclosed and is unusual enough to note: eligible customers can apply funds already committed under their cloud provider consumption agreements toward the platform, which changes how the purchase is budgeted without revealing what it costs.

Coverage spans three adjacent buyer types with genuinely different workflows, covering wealth managers and their advisers, asset managers and their fund distributors and wholesalers, and insurance firms and brokers, with the company stating it developed the question set separately for each. Presence is stated across New York and London. Held at B because no institution count, geographic breakdown or segment depth is evidenced beyond the aggregate assets figure.