Consumer Duty Claims Handling: How AI in Dynamics 365 Customer Service Can Help Insurers

Consumer Duty claims handling: relieved cartoon customer holding a document in front of a teal protective shield

If you work in insurance, you’ll know that a claim is the moment of truth. It’s when a customer finds out whether the cover they paid for really works, and it’s often when they’re at their most stressed. A slow reply, a confusing letter or a query that sits in the wrong inbox can turn a difficult week into a complaint.

That’s exactly where the regulator is looking. In its 2026 Regulatory Priorities report for insurance, the Financial Conduct Authority (FCA) makes improving consumer understanding, claims handling and service quality a headline priority, and asks firms to respond promptly, fairly and transparently to claims and queries. Getting Consumer Duty claims handling right is no longer only about doing the right thing; it’s about being able to show that you have.

In this article I’ll look at what the FCA is asking for, where service teams tend to struggle, and how recent AI updates in Dynamics 365 Customer Service can help you deliver and evidence better outcomes, with your people firmly in control.

 

What the FCA Expects From Insurers in 2026

In February 2026 the FCA published its new Regulatory Priorities reports, which replace more than 40 portfolio letters with one annual report per sector. The insurance report sets out three clear expectations under its claims and service quality priority:

  • Be clear: tell consumers clearly what their insurance covers and test their understanding.
  • Handle claims fairly: respond promptly, fairly and transparently, and deliver good outcomes when customers claim.
  • Test outcomes: track whether products, services and interactions with consumers are delivering good outcomes in line with the rules.

Following the Which? super-complaint on home and travel insurance, the FCA says it will continue its supervisory and enforcement investigations into home and travel claims handling, and expand its review of how firms oversee outsourced claims processes. Its post-implementation review of the value measures rules, which includes considering how claims outcomes are captured, is due to conclude in Q4 2026.

So if you’re planning for 2027, claims and service data is likely to be under more scrutiny, not less.

 

Data Alone Won’t Prove Good Outcomes

Cloud9 Insight colleagues on a green sofa reviewing customer service data on a laptop, discussing how to evidence good customer outcomes

In July 2026 the FCA published its good practice and areas for improvement findings on Consumer Duty outcomes monitoring. One message stands out: collecting data, listing metrics or reporting management information (MI) won’t, by itself, show that customers are getting good outcomes.

Instead, firms should be able to explain what their information tells them, how they use it to spot risks or issues, what action they take, and whether that action improved outcomes. The FCA also highlights good practice in monitoring specific stages of the customer journey, rather than relying only on high-level indicators.

For many of the insurers and brokers we speak to, that’s where the real challenge lies. Claims and service information is often scattered across shared mailboxes, spreadsheets, phone notes and outsourced partners. When it isn’t joined up, it’s hard to see where a customer journey is going wrong, let alone prove you fixed it. That’s why we always start with connected data in a CRM built for financial services.

 

Getting Every Claims Email to the Right Place

Email is still one of the busiest channels for claims and policy queries, and it’s also one of the easiest places for something urgent to get lost among newsletters, auto-replies and general enquiries.

Email classification in Dynamics 365 Customer Service, which Microsoft made generally available in August 2026, uses AI to read the subject and body of incoming emails and assign each one to a category you define. You set up your own taxonomy, with up to 20 top-level categories and up to four levels of subcategories, so it can reflect how your team really works. For an insurer, that might mean separating new claims, claim updates, policy changes and complaints.

The category is stored on the email record, so you can use it to:

  • Route emails: send each message to the right queue or team, so a complaint doesn’t wait in a general inbox.
  • Control case creation: stop marketing emails and notifications from creating cases, so your SLAs reflect real customer work.
  • Drive automation and reporting: use the category in automatic record creation rules, Power Automate flows and reports.

Before you switch it on, you can run a simulation against historical emails to test and refine your categories. Microsoft notes that pay-as-you-go billing needs to be enabled for your environment, so it’s worth checking your licensing and configuration first.

Classification also works alongside the Case Management Agent, which helps automate the case lifecycle, from creating and updating cases to resolving and closing them, so your team spends less time filling in case details and more time with customers.

 

Checking Every Reply Before It Reaches a Customer

Cloud9 Insight colleague working at a laptop, checking that customer emails meet policy and compliance standards before sending

Clear communication is at the heart of the Consumer Duty, and it’s one of the main questions insurers ask us about AI: how do you make sure a drafted reply says the right thing?

Governance in Dynamics 365 Customer Service, generally available since August 2026, checks outbound emails against policies your organisation defines before they’re sent. It applies to emails written by service representatives and to responses generated by autonomous AI. For each policy you choose how strictly it’s enforced:

  • Flag: silently record the violation for auditing.
  • Warn: show a warning, but allow the email to be sent after review.
  • Block: prevent the email from being sent until the issue is fixed.

There are ready-made system policies, including Email Template Adherence, Groundedness and Profanity, and you can write your own in plain language. Checks can also cover text-based attachments in Word, PDF and text formats, and you can set the strictness from Low to Very High. Microsoft suggests the stricter settings where regulatory requirements are tight.

You can test policies in a simulation before you publish them, and each policy evaluation is logged, so supervisors and compliance colleagues can see what was checked and why something was flagged. If you’re thinking about where AI fits in a regulated service team, this is the kind of guardrail that lets you move forward with confidence. We’ve written more about the risks of ungoverned AI in our article on shadow AI.

 

From Spot Checks to Consistent Quality Reviews

Many service teams still rely on a handful of manual call or file reviews each month. The FCA’s consumer support findings note that some firms use proactive tools, such as speech analytics, to assess the quality of support, while smaller firms might use periodic call listening or file reviews with a simple checklist, focusing on higher-risk journeys such as bereavement or financial difficulty.

The Quality Evaluation Agent in Dynamics 365 Customer Service helps you do this more consistently. Your supervisors define the evaluation criteria, including the questions, answer choices and scoring, and evaluation plans decide which interactions are reviewed and when. The agent then evaluates cases and conversations, with email evaluation currently in preview, and produces scores, summaries and recommended actions.

That gives you a more structured view of service quality across the claims journey and a clearer trail of what you found and what you did about it. Microsoft is clear that it’s designed to help managers improve team performance and customer satisfaction, not to make employment decisions, and that’s a principle we’d always echo. AI prepares the evidence; your people decide what happens next.

 

How to Start Improving Consumer Duty Claims Handling

Smiling Cloud9 Insight colleague at a desk, representing a confident service team improving Consumer Duty claims handling

You don’t need to switch everything on at once. If I were starting tomorrow, here’s where I’d begin:

  1. Map your claims journey. Write down what a good outcome looks like at each stage, from first notification to settlement, so you know what you’re measuring.
  2. Bring your channels together. Make sure claims emails, calls and case notes live in one place, so your MI reflects the whole journey.
  3. Start with classification. Define a small, clear set of email categories, test them in simulation, and use them to route complaints and urgent claims first.
  4. Add guardrails before automation. Set up governance policies in warn or block mode before you let AI draft or send replies.
  5. Review and act. Use quality evaluations and reporting in Power BI to spot where outcomes slip, act on what you find, and check whether it worked.

Our Inspirational Clarity approach is built around this kind of step-by-step progress, using out-of-the-box capabilities wherever possible. You can also see how we’ve worked with insurance businesses in our Acquis Insurance Management case study and our Direct Commercial case study, or watch our 10-minute Dynamics 365 for Financial Services demo.

 

Delivering Outcomes You Can Stand Behind

The FCA’s direction is clear: insurers need to handle claims promptly and fairly, communicate clearly, and show that their monitoring leads to real improvements. Strong Consumer Duty claims handling comes from connected data, well-designed processes and AI that’s governed properly, with people making the decisions that matter.

If you’d like to talk through how Dynamics 365 Customer Service and AI for CRM could support your claims and service teams, contact Cloud9. Or, if you’re already considering Dynamics 365 Customer Service, get a quote with our Clarity Tool.