Show the regulator the call actually met the duty
FCA Consumer Duty moved the bar from selling a product to evidencing good outcomes on every advice and service conversation. Reading a sample of calls by hand cannot cover the volume. Semarize runs each regulated conversation through a Kit of typed checks and returns suitability, vulnerability, and complaint signals with the exact quote attached, so the evidence is on file before anyone asks for it.
Kit
Consumer Duty Kit
Structured signals from a regulated conversation
Output
{
"suitability": "review",
"capacity_for_loss": false,
"vulnerability_indicator": "financial difficulty",
"complaint_intent": true,
"redress_risk": "medium"
}
The problem
The duty is evidenced in the conversation
Whether advice was suitable, whether a vulnerable customer was handled well, whether a grumble was really a complaint - all of it is decided on the call and almost none of it lands in a system field. Sampling a handful of calls cannot prove good outcomes across the whole book.
Suitability is recorded, not evidenced
A fact-find ticks the boxes. Whether capacity for loss was genuinely explored, or just asserted, only lives in what was said on the call.
Vulnerability gets noticed but not acted on
A money worry or a sign of confusion is often heard and then handled the same as any other call. The duty is to adapt, and that gap rarely surfaces until a file is pulled.
Complaints hide inside ordinary calls
Customers do not announce a complaint. They mention a fee that felt unfair or a promise that was not kept, and it slips past unless someone catches the intent.
QA samples a fraction and hopes it is representative
Manual review covers one or two percent of calls. The regulator expects outcomes across the whole population, not a sample that happened to be picked.
Financial services examples
Regulated conversations, with evidence attached
Each advice and service call runs through the same Consumer Duty Kit and returns suitability, vulnerability, and complaint signals tied back to the moment they were said.
01 / Suitability
Prove suitability was assessed before any recommendation
Consumer Duty expects firms to show advice fit the client, not just that a product was sold. Semarize evaluates whether needs, risk appetite, capacity for loss, and rationale were genuinely covered on each advice call, and returns a pass, fail, or review field with the evidence behind it.
Suitability check
Advice call, one client
02 / Vulnerable customers
Flag customer vulnerability and how it was handled
Vulnerability shows up as a money worry, a health condition, or signs of confusion - and the duty is to adapt, not just notice. Semarize detects those indicators and whether the advisor responded appropriately, with the exact moment attached for the file.
Vulnerability handling
detected on callmoney has been really tight since my hours were cut, and I can't keep on top of the payments
vulnerability_indicator = "financial difficulty"
03 / Complaints and redress
Categorise complaints and catch redress risk early
A complaint is rarely labelled as one on the call. Semarize identifies complaint intent, the root cause, and whether fair treatment or redress is in question, then categorises it - so complaints teams see the cases that need action before they escalate.
Complaint pipeline
not labelled on the callHow it works
How Semarize turns regulated calls into outcome evidence
Connect the advice and service calls you already record. A Consumer Duty Kit of Bricks evaluates each signal, attaches the quote it came from, and returns typed fields ready for your QA, complaints, and compliance reporting.
- Step 1
Conversation captured
Advice calls, service calls, and chat transcripts are ingested.
- Step 2
Consumer Duty Kit runs
Bricks evaluate suitability, vulnerability, and complaint signals, grounded in your policy.
- Step 3
Signals extracted
Pass and fail flags, categories, and extracted values come back with confidence.
- Step 4
Evidence attached
Every signal carries the quote and timestamp it was drawn from.
- Step 5
Systems updated
Typed fields feed QA queues, complaints systems, MI dashboards, and reporting.
Example Bricks
The compliance signals Semarize returns
Each Brick is a typed check that evaluates one regulated signal and returns a structured value with confidence, reason, and evidence. Bundle them into Kits to score every conversation against the same standard. Knowledge grounding lets a Brick reference your own suitability and complaints policy.
Suitability
Client needs explored, not assumed
Attitude to risk discussed on the call
Ability to absorb a loss genuinely covered
Overall suitability verdict
Vulnerable customers
Driver of vulnerability named in context
Advisor adjusted the approach in response
Advisor confirmed the client understood
Call routed for a manual outcome review
Complaints and redress
Expression of dissatisfaction that meets the complaint bar
What the complaint is actually about
Customer treated fairly on the call
Likelihood redress is owed
Example output
Structured evidence, not a call note
The Consumer Duty Kit returns the same schema every run: typed values, confidence, and the evidence span tied back to the conversation it came from.
Consumer Duty Kit returned
Outcome fields written straight to the customer record
- Suitability
- Review
- Capacity for loss
- Not covered
- Vulnerability indicator
- Bereavement
- Advisor adapted
- Yes
- Complaint intent
- Detected
- Redress risk
- Medium
{
"run_id": "run_fs7841",
"source_call": "call_2c19",
"status": "succeeded",
"output": {
"bricks": {
"suitability": {
"value": "review",
"type": "category",
"confidence": 0.86,
"reason": "Needs and risk appetite covered, capacity for loss not explored",
"evidence": [
{ "quote": "we never really went through what a downturn would mean", "span": [1840, 1902] }
]
},
"vulnerability_indicator": {
"value": "financial difficulty",
"type": "category",
"confidence": 0.91,
"evidence": [
{ "quote": "money has been really tight since my hours were cut", "span": [728, 779] }
]
},
"complaint_intent": {
"value": true,
"type": "boolean",
"confidence": 0.82,
"evidence": [
{ "quote": "I was never told there would be a fee like that", "span": [2410, 2456] }
]
},
"redress_risk": {
"value": "medium",
"type": "category",
"confidence": 0.78
}
}
}
}The difference
Why a call note will not stand up to a file review
A call note is written for a colleague to skim later. To evidence good outcomes you need signals you can sort, count, and trend across the whole book of conversations.
Call notes and sampling
Structured compliance signals
FAQ
Conversation intelligence for regulated firms, answered
How suitability, vulnerability, and complaint signals are produced from real advice and service calls.
Keep exploring
Where compliance signals go next
The same structured outputs feed adjacent teams and the product layer underneath.
Score every regulated call against your standard, not a sample.
Track service quality and at-risk customers across conversations.
Coach advisors on suitability and vulnerable customer handling.
Feed typed conversation signals into your own models and MI.
Define the typed compliance signals Semarize evaluates.
Run every conversation through Semarize from your own stack.
Get started
Evidence good outcomes on every call
Run each advice and service conversation through a Consumer Duty Kit and have suitability, vulnerability, and complaint signals on file before the regulator asks.
No card required. Start testing in minutes.