Finspector

Nothing survives the removal test. Take the models out and what is left is an upload form, a comments panel and an activity log, none of which anyone would buy. Every function the company sells is inference: reading text, video and audio for regulatory risk, reading back through a social profile's post history, and adapting checks from a compliance manager's corrections.

The company argues the point itself from the opposite direction, positioning against general purpose assistants on the ground that those cannot process video or audio and were not built for financial promotion rules, which is an argument about model capability rather than about workflow. Placed with the agent native cohort.

Low autonomy by construction, with the human decision preserved at every step and mechanisms rather than assurances behind it. The model surfaces risks and suggestions when a promotion arrives, a person approves or rejects through the dashboard, assignment and notification route work to a named reviewer, and every interaction is logged so the compliance trail is reconstructable.

The feedback mechanism is itself an oversight control, since it gives the reviewer a formal route to record that the model was wrong rather than only to override it silently. Held at B because no threshold, confidence measure, recall figure or sampling audit is published, and because reducing unnecessary flags is advertised as a benefit, which is a threshold being moved without being described.

No accuracy figure, benchmark, validation method or error rate was located. One disclosure sits above the category norm and is recorded rather than credited: the company states publicly that feedback loops from its early pilots reduced false positive rates, which is an open acknowledgement that false positives existed and a claim about direction of travel. Most vendors in this pocket will not concede an error rate in any form.

It remains a direction without a baseline, a current figure or any measure of the failure that matters more, the promotion that breached and was never flagged. Reducing unnecessary flags, advertised as a feature, moves precision and recall in opposite directions, and only one of the two is ever mentioned.

No named customer, no customer count, no case study and no independent evaluation were located, which is the C bar met plainly rather than harshly. The company is roughly eighteen months old and its published outcome claims are conditional in its own wording, describing hours that could be saved rather than hours that were.

The strongest available material is a described early engagement in which two policy documents were converted into eighty two separate automated checks spanning several jurisdictions, but the customer is unnamed and the figure describes configuration effort rather than result. Participation in a Scottish financial regulation innovation programme is a cohort placement.

One point of genuine credit that does not move the grade: the company publishes an unusually candid account of its own development, including a count of feature requests taken from early users, which is more methodological honesty than most vendors of any size offer.

No boundary statement was located, and the product's central mechanism makes the question unavoidable rather than theoretical. Compliance managers correct the model where it was wrong, and the company states that the checks improve continuously from that feedback. Improvement from whose corrections, serving whom, is precisely what is left unsaid.

If a firm's corrections tune only that firm's checks, this is the isolation architecture that earned credit elsewhere in this pocket and would be worth saying. If they inform a shared model, then competing firms are training a common detector on their own compliance judgements, which several of them would consider proprietary. The vendor has an easy answer available either way and has not published it.

No privacy statement, lawful basis, retention rule or handling commitment was located. The scope reaches further than a marketing review tool would suggest and that is what makes the silence matter. Continuous monitoring of social profiles, including reading back through post history, means processing the public output of identified individuals, and the influencers and partners being screened are third parties who have no relationship with the vendor and in many cases no knowledge that a firm has pointed a compliance system at their account. Sales call review adds recorded conversations to the same platform. Nothing published addresses lawful basis for any of it, or how long monitored material is retained.

No information security certification, audit report, penetration testing statement or trust centre was located in this pass. Secure artificial intelligence appears as a section heading in the product material, describing the ability to enter custom rules, which is a configuration feature rather than a security property, and the mismatch between the label and what sits under it is worth noting because a reader scanning for assurance would take the wrong impression.

Recorded as an absence found rather than a proven absence. For a company of this age the honest expectation is that certification is a roadmap item rather than an omission, and a stated roadmap would itself be a disclosure.

Correct technology supplier posture with rule engagement that is specific rather than generic. The product is built around Financial Conduct Authority financial promotion rules and Consumer Duty, named directly and with the obligation stated correctly, that a firm must be able to demonstrate every customer communication is clear, fair and not misleading.

It names the concrete failure modes it screens for, including mis selling, greenwashing and inadvertently giving regulated financial advice, which is a sharper articulation than most peers manage. Graded at B rather than A because no formal admission process was located: participation in a financial regulation innovation programme run by an industry cluster is a cohort placement, not a supervisory engagement, and is treated the same way here as an insurance market accelerator placement was for a vendor rejected earlier in this sweep.

No governance framework, fairness position, testing programme or independent assessment was located. The product specific exposure is that the judgements being automated are openly subjective ones, and the company says so itself when describing the problem it solves: whether a risk warning is sufficiently prominent, and whether a communication is clear, fair and not misleading, are interpretive determinations that experienced compliance officers disagree about.

Automating a contested judgement makes it consistent without making it correct. A second exposure follows from the video and audio capability, since vision and speech models sit upstream of every judgement on that content and their accuracy varies with accent, production quality and language, with nothing published per format.

No accuracy warranty, service commitment or remedy was located. The consequence structure lands wholly on the firm: a promotion the system passed and the regulator later intervenes on is the firm's breach, and under Consumer Duty the obligation to evidence that a communication was clear, fair and not misleading cannot be delegated to a supplier.

A second party sits further out and is unusual for this pocket, the influencer or partner whose account is monitored and whose post may be flagged, who has no contract with either side and no route to contest a determination made about their content. Nothing published describes a remedy, a correction path or a notification duty in either case.

No provider, base model, fine tuning approach or hosting arrangement was located. The omission is pointed because of how the company positions itself. Its published argument is that general purpose assistants are inadequate for this work and that its own model is purpose built for financial promotion compliance, which is a claim about provenance and architecture, and it is made without disclosing either.

A buyer persuaded by the argument still cannot tell whether the underlying system is a distinct model, a fine tune, or a general purpose model behind a domain specific prompt layer, which is precisely the distinction the marketing rests on.

Material reaches the platform by upload or by entering a social handle, and no named integration to a marketing resource management system, a digital asset store, a customer relationship platform, a project tool or a governance system was located, nor any public application programming interface documentation or connector catalogue. Continuous monitoring of the major social platforms is real inbound integration and is the strongest element here.

What is missing is the outbound half: a compliance check that lives in its own window rather than inside the tool where content is created relies on someone remembering to run it, which is the workflow gap the better funded peers in this pocket spend their integration budget closing.

Delivered as cloud software with no hosting region, tenancy model, residency option or subprocessor list located. The gap is straightforwardly commercial for this vendor rather than abstract, because a British company selling to British regulated firms will meet the question in the first vendor assessment it is asked to complete, and United Kingdom data protection obligations make the location of processing and the identity of subprocessors a standard item rather than an unusual one. Nothing published answers it, and a young company has the least established practice for answering it in a sales conversation instead.

No rate card, tier ladder or billing basis was located and the route to a number is a demo request. The omission is more surprising here than for the enterprise vendors in this category, because the stated buyer is the small and medium sized regulated firm, and that buyer is the one most likely to abandon an evaluation rather than sit through a sales process to find out whether a product is affordable. A published entry price is a cheaper acquisition channel than a demo form for this segment, and the absence of one is a commercial choice rather than an enterprise convention.

One of the few genuine C grades on this axis, where C means narrow rather than undisclosed. The named buyer types read broadly on paper, spanning banking, insurance, wealth management, fintech, payment providers and digital banks, but the company is roughly eighteen months old, positions itself for small and medium sized firms, and no customer count, named institution or deployment at scale was located anywhere.

The one wide jurisdictional claim, checks spanning the United Kingdom, Europe, Australia, the United States, Asia and the Middle East, came from configuring a single early client's own policy documents, so it describes that client's footprint rather than the vendor's coverage. The rule engagement itself is squarely British, built around Financial Conduct Authority promotion rules and Consumer Duty.