The Consumer Duty asks you to show your customers understood what you told them. Dutified gives you that evidence before any letter, email, ad or web page goes out.
Why it matters
Firms have spent three years on rules checks, sign-off gates and plain language guides. All of that matters, and none of it answers the question the regulator is now asking.
A rules check confirms the letter holds the right warnings, costs and deadlines. It cannot tell you whether a widowed customer worked out what her deadline was.
An average score hides the customers the regulator cares about most. A letter can work for 9 in 10 and still fail almost everyone who is recently bereaved.
Panels are small, slow and expensive, so most communications go out untested. The first real test is the complaints that arrive afterwards.
The evidence has to come before the letter goes out, from customers who look like yours, at a scale a panel cannot reach.
What Dutified doesWhat we do
Works on anything a customer reads
How the engine works
Everything you test goes through the same four steps in the same order, and each problem we find is linked to the part of the FCA’s rules it affects, from the four Consumer Duty outcomes to the guidance on vulnerable customers.
First we check the letter against a fixed list of FCA rules, making sure it has every warning it needs, gives deadlines as dates, shows charges in £, makes leaving as easy as staying and uses words most people can read.
Next we read for what the letter really tells a customer, flagging jargon, bad news tucked out of sight, benefits shown without the risks and anything that rushes people or makes it hard for them to act.
Then 1,000 modelled customers, built to match your own, read the letter and say in their own words what it means for them, and a separate checker marks each answer against what the letter needed to get across.
Finally a panel of real people like your customers reads the same letter, and comparing their answers with the modelled ones shows how far the model can be trusted, a figure we print on every report.
Every problem points to the exact FCA rule behind it, taken from our own library of rules instead of written by the AI, so anyone can check it.
Meet the readers
Each square is an AI agent built to behave like one of your customers, so you can watch them read the letter and then see who they are, how far they got and what they took away.
Mapped to the Consumer Duty
Does the tone and complexity suit the customers it is actually going to?
Do customers come away knowing what they pay and what changed?
Did they understand it well enough to make a good decision? Tested on your own book.
Do they know how to switch, exit, get help or complain, without friction?
How it works
Four steps. The population is built from what you already know about your customers.
A short questionnaire draws on your complaints, calls and segmentation to set who your customers are, including how many are vulnerable.
We model your book as 1,000 individual customers, with vulnerability tied to age and circumstance the way it happens in life.
The rulebook check runs, every customer reads the letter, then the test runs again weighted to vulnerable customers.
The verdict, the failing passages, who it failed and the rewrites, walked through with your team.
The report
A plain English verdict first, the provenance your compliance team needs last, and in between, exactly where the letter breaks and how to fix it.
Every passage that confused customers is highlighted, with a count of how many named it unprompted as the moment they lost the thread.
It is usually the sentence added in the final legal review.
The Consumer Duty is judged at the floor, so every vulnerable circumstance is scored next to everyone else.
A letter that works for 9 in 10 can still fail most of the recently bereaved.
Customers who struggled explain what lost them, and every failing passage comes back with a rewrite that keeps your meaning.
Your team accepts or overrides each one, and every decision is recorded.
"I don't know what a valuation point is, so I don't know when my deadline is."
Aged 78, managing the family finances after a bereavement
Our reference panel
Alongside your own population, every test runs a fixed panel of the customers the Consumer Duty worries about most. Your team gets to know them by name.
Her husband ran the money. She reads every word slowly, twice, and still worries she has missed the part that matters.
Confident with numbers and quick to skim. Reads the first paragraph closely and the rest far less carefully.
Managing an account that is not hers, between work and hospital visits. Wants to know what she is allowed to do.
Substantial money, no experience and reluctant to admit it. Agrees with anything that sounds authoritative.
Loses the thread of long sentences and says she understands when she does not, because asking feels embarrassing.
Tired by evening and opens post in batches. Acts only on what looks urgent.
Vulnerability, modelled properly
Your most exposed customers are often the ones with the most money, so we model the circumstances that actually appear in an investment book.
Method and governance
Same letter, population and settings give the same result. Every report carries its run reference, date and model versions.
A human panel reads a sample so the modelled scores are calibrated against real understanding, and the agreement rate is reported openly.
Inputs, answers, findings and your team's decisions are stored against each run and export cleanly for audit or a regulatory review.
Book a demo
Bring a live communication. We will run it through both layers and walk you through what the report shows.
We will be in touch within one working day to arrange a time.