Runtime AI Control for high-risk workflows.
Clairify is designed for AI use cases where regulators care about the output and mistakes can create customer harm, regulatory breach, or audit exposure.
Customer Support and AI-Generated Communications
What Clairify checks
Clairify can assess outputs against product facts, customer context, authentication status, conduct rules, and internal communication policies.
What evidence is created
The record shows the authentication status at the time of the output, which conduct rules were evaluated, whether vulnerable customer flags were present and acted on, and the intervention decision made, supporting Consumer Duty fair outcomes evidence and complaints handling review.
What can go wrong in Customer Support
- Incorrect product, pricing, or eligibility information
- Misstated customer rights
- Inappropriate handling of vulnerable customers
- Disclosure of personal or account data
- Confident but unsupported regulatory explanations
What can go wrong in Financial Promotions
- Misleading product descriptions
- Unsupported performance claims
- Unapproved financial promotions
- Blurred lines between information and advice
- Missing risk disclosures
Sales Copilots and Financial Promotions
What Clairify checks
Clairify can evaluate draft communications against financial promotion rules, product documentation, approval requirements, and customer context.
What evidence is created
The record constitutes a pre-send approval trail: the communication assessed, the financial promotion rules applied, the product documentation used, and whether the output was allowed or blocked before it reached the client. This is the evidence a regulator would expect to see if a promotion were challenged.
Insurance Claims and Underwriting
What Clairify checks
Clairify can compare AI-generated explanations with policy documents, customer data, claims records, and conduct requirements.
What evidence is created
The record captures the policy version and exclusion wording in force at the time of the decision, the customer data used, and the conduct rules applied, giving auditors and the FOS a complete picture of what the customer was told and on what basis.
What can go wrong in Insurance Claims
- Incorrect policy exclusions
- Misstated coverage terms
- Unsupported claim explanations
- Inappropriate use of protected characteristics
- Inconsistent customer communications
What can go wrong in Complaints and Disputes
- Incorrect dispute or complaint timelines
- Misstated reimbursement rights
- Incorrect explanation of payment failure
- Confirmation of an action before authentication
- Disclosure of fraud or security logic
Complaints, Disputes, and Payments
What Clairify checks
Clairify can validate outputs against payment rules, customer authentication status, case data, transaction data, and internal process requirements.
What evidence is created
The record logs the authentication status at each step, the transaction data and payment rules used, and whether complaint or dispute timelines were correctly applied, providing the timeline evidence required for FOS referrals and regulatory reporting.
Collections and Debt Management
What Clairify checks
Clairify can evaluate the customer context, vulnerability indicators, proposed wording, available options, and required escalation rules.
What evidence is created
The record shows the vulnerability assessment at the time of contact, which forbearance options were available and whether they were offered, and the escalation decision. This is the chain of evidence needed to demonstrate appropriate treatment under FCA collections guidance.
What can go wrong in Collections
- Inappropriate language toward a vulnerable customer
- Failure to identify signs of financial difficulty
- Misstatement of repayment options
- Incorrect escalation or forbearance handling
- Communications that conflict with policy or conduct obligations
What can go wrong in KYC and AML
- Incorrect summaries of adverse media
- Disclosure of investigation details
- Tipping-off risk
- Misclassification of screening outcomes
- Weak evidence for onboarding or escalation decisions
KYC, AML, Sanctions, and Adverse Media
What Clairify checks
Clairify can assess confidentiality requirements, process rules, customer context, source provenance, and whether AI-generated summaries are supported by authoritative information.
What evidence is created
The record captures the risk classification applied and the escalation decision, including whether tipping-off constraints were correctly observed. This supports MLRO review, internal audit, and regulatory examination of the firm's AML control framework.
Wealth and Investment Communications
What Clairify checks
Clairify can assess whether an output is informational, promotional, or advisory; whether required disclosures are present; whether product and portfolio information is supported by authoritative data; and whether the communication is consistent with the client context available to the firm.
What evidence is created
The record captures the client profile and investment objectives in force at the time, the disclosures present or absent, and whether the output was informational, promotional, or advisory, providing the suitability and disclosure evidence regulators expect when reviewing investment communications.
What can go wrong in Wealth Management
- AI presents a portfolio change as suitable without the required assessment
- Investment risks, fees, or limitations are misstated or omitted
- Generic commentary becomes personalised advice
- Client circumstances or investment objectives are ignored
- Performance claims are unsupported or missing required context