Know your customers. Prove the change worked.

A transformation is only real when it produces better outcomes for customers, and you can show it. Most programmes measure delivery: features shipped, milestones passed. Far fewer can answer the question a board and a regulator now ask, which is whether the change actually helped the customer.

The FCA's Consumer Duty makes that a standing obligation: deliver good outcomes for retail customers, monitor whether those outcomes are landing, and evidence both. We design for it from the start, by understanding customers, watching how a change shifts their behaviour, and producing the evidence as a by-product of the work rather than a report assembled after the fact.

Understand the customer, not just the transaction

Good outcomes start with how a customer's intent actually forms, not only what they finally clicked.

Every interaction carries two kinds of signal. One is the behavioural detail of how someone moved through a journey: the sequence, the hesitation, the point where they needed help. The other is the governed record of what was decided and why. We treat both as first-class. Together they show not just that a customer completed a task, but whether they understood it, whether it suited them, and where a vulnerable customer might have struggled.

Monitor whether the change had its intended impact

Decide what good looks like before you ship, then measure against it.

A change is really a hypothesis about customer behaviour. We make that hypothesis explicit: the outcome each change is meant to produce, the metric that would show it, and the customer groups it has to hold for. Measurement then runs against those targets continuously, so a shift in behaviour shows up as it happens and a poor outcome is caught early, rather than surfacing in a complaints trend months later.

From

  • Delivery metrics: features and milestones
  • Outcomes assumed once a change ships
  • Poor outcomes surfacing in complaints
  • Annual reporting assembled by hand

To

  • Outcome metrics defined before the change
  • Behaviour monitored continuously
  • Poor outcomes caught early, by customer group
  • Evidence produced as a by-product

Evidence by construction, not a report afterward

The architecture that authorises an agent's action also records why it was the right one for the customer.

In our Capability, Context and Consent architecture, every material action is checked against policy and leaves an evidence trail. That trail is what turns "we believe outcomes are good" into "here is the record, by customer group." A board report becomes a query over evidence the business already holds, not a project to reconstruct it. You cannot be accountable for what you do not control, and you cannot prove an outcome you never measured.

Built around the FCA's four outcomes

Consumer Duty is framed by the FCA as four outcomes. Each is easier to deliver, and far easier to evidence, when the data and controls are designed for it.

Products and services

Show that a product reaches the customers it was designed for, and works as intended for them.

Price and value

Evidence that what a customer pays is reasonable for the benefit they receive, watched over time.

Consumer understanding

Test whether customers really understood a communication or a decision, from how they behaved, not only what they were sent.

Consumer support

See where customers struggle or need help, and whether support resolved it, with particular care for vulnerable customers.

Consumer Duty is an FCA framework and the four outcomes are the FCA's. Our role is to help you deliver and evidence them, not to certify compliance.

Let's build something that ships.

Tell us what you're trying to do with AI. We'll tell you, honestly, whether and how we can help.

Book a call

Get the insights by email

Field notes on agentic engineering, on-device AI and transformation. No spam, unsubscribe anytime.