Consumer Duty reporting, automated.
Check everything. Sample nothing.
Most firms evidence the Duty by spot-checking a sample of files each quarter and assembling the story afterwards. We build the data foundation and checks engine that monitors every client outcome, continuously - so reporting becomes a query, not a project.
Sampling was a workaround, not a method.
Four outcomes to evidence, a board report to stand behind, and the data scattered across CRM, platform, billing and inboxes. So compliance reviews 5% of files and hopes the sample is representative - while the evidence for "ongoing monitoring" gets rebuilt by hand every quarter. The firm doesn't have a Consumer Duty problem; it has a data access problem wearing a compliance costume.
A checks engine, not another dashboard.
To be clear about who does what: we're not compliance consultants- we're engineers and data people who understand how technology has to fit this environment. Your compliance team defines what good looks like. We make it run against every client, every day.
data_foundation
CRM, platform, billing and communications data joined into one governed model - the single version of each client the rest of the system runs on.
checks_engine
What "good" looks like for each outcome, codified as deterministic, versioned rules your compliance team approves - fee fairness, review frequency, communication cadence, vulnerability flags.
exception_workflow
A failed check becomes a routed task with an owner and a deadline - found same-day, not at the annual file review.
evidence_archive
Every check, result and resolution logged as it happens. When evidence is requested, it's a query - already in the shape a regulator expects.
board_reporting
The annual board report and ongoing outcomes monitoring drawn from live data - numbers your board can interrogate, not a snapshot assembled the week before.
self-serve_access
Compliance gets direct, governed access to the data - no more requests into the IT queue for questions the Duty expects you to answer routinely.
A worked example. Your diagnostic produces the real numbers.
A mid-sized wealth manager: ~2,000 client relationships, one compliance officer plus a part-time consultant, file reviews running at 40 per quarter. Phase one joins the CRM and platform data and ships the first ten checks; coverage grows from there. Illustrative before/after:
AI makes us faster to build. It doesn't make your compliance decisions.
Systems like this used to be a six-figure consultancy programme, which is why only the largest firms had them. We engineer with AI-accelerated tooling and ship production code several times faster than that era - and the saving is priced into the engagement. What ships is the opposite of a black box: deterministic, versioned rules a regulator can read, with AI used only where it survives an audit.
Asked and answered.
- Are you compliance consultants?
No - deliberately. We're an engineering and data practice. Your compliance team or external advisers own the interpretation of the rules and what 'good' looks like; we turn that into a system that checks it across every client, continuously, with evidence. We don't advise on the Duty - we make your approach to it operational. - Does automated checking replace our compliance team?
No - it replaces the sampling drudgery. Instead of manually reviewing a few dozen files a quarter, your team reviews the exceptions the engine flags, applies judgment where judgment is needed, and owns the rules the engine runs. The work moves up a level; it doesn't disappear. - Is AI deciding whether our customer outcomes are fair?
No. The checks are deterministic, versioned rules that your compliance team defines and approves - a regulator can read them line by line. We use AI to build the system faster and, where appropriate, for document extraction inside auditable controls. It never makes the compliance judgment. - What systems does it need to connect to?
Typically your CRM, platform or custody data, fee billing, and client communications. We've designed for the usual mix - mainstream CRMs, wrap platforms, document stores, spreadsheets that are quietly load-bearing. The first phase of any engagement maps exactly what you have. - How long until the first checks are live?
Weeks, not quarters. We phase it: data foundation and the highest-value checks first, then coverage grows outcome by outcome. You see flagged exceptions from the first phase - not after a year-long programme. - We already have dashboards. Isn't this the same thing?
Dashboards show you numbers; they don't check anything. A checks engine codifies what 'good' looks like for each outcome, tests every client against it continuously, and records the evidence. Your BI tool can sit on top of it.
What would checking everything change?
A 20-minute diagnostic: we'll map your data sources, estimate what continuous checking would cover from day one, and tell you honestly whether the build is justified. Part of the Find → Capture → Compound method.