It’s not hard to understand why more than three-quarters (78%) of insurers see customer experience (CX) as central to their growth strategy.

Consider Tom who reports a car accident through his insurer’s app and uploads the requested photos. When he checks the insurer’s website for an update the next day, there’s no record of his claim. Frustrated, he calls for help, only for the claims agent to ask him for the set of documents yet again.

This is a common industry challenge. 42% of insurance executives cite fragmented, multi-channel journeys as a top barrier to providing a great CX today. And 60% expect it to worsen over the next few years.

A discontinuity during a claim or when taking out a new policy can quickly damage customer confidence. And it shows up as cost: repeat contacts, duplicate data entry, and rework across teams.

But many leading insurers are beginning to close these gaps with AI and give their frontline staff the context they need to provide better experiences.

Remove friction, not the human touch

AI can make the moments that matter feel simpler. For example, at the first notice of loss (FNOL), AI can capture the essential details and ensure that all information is accurate and complete. In service and renewals, AI assistants can answer routine questions quickly and help customers with many tasks without waiting in a queue.

Over half (53%) of insurers say they’re already using AI for customer service, chatbots, messaging, and self-service portals. These interactions are concentrated at the FNOL and claims-intake stage, where automation lowers cost per contact and improves containment before a case reaches an adjuster.

But the more important shift is not simply putting a bot in front of the customer. It’s using AI behind the scenes to give claims handlers, underwriters, agents, and contact center teams a clearer view of each case. Insurers will continue to place high value on human expertise, especially in complex claims or renewal negotiations.

Our survey results show that 63% of insurance customers rank human support among their top three preferred channels. But at the same time, 59% also rank AI-integrated bots among their top three, reinforcing that a blended, human-in-the-loop model helps resolve more inquiries through self-service while maintaining effective personal support for complex cases.

Many insurers are already striking this balance. For instance, The Hartford, a US-based insurance and investment company, uses AI agents to answer common questions or provide real-time claim updates. But it also uses AI to offer real-time support to agents for better first-call resolution through call summaries and sentiment analysis.

Customers have a clear expectation: give us AI to simplify processes and give us frontline staff for when judgment, reassurance, and a more nuanced conversation matter to us. There’s a clear path to turn that balance from an ambition into a practical starting point.

Start where friction is highest

The best place to begin is not with AI itself, but with a customer journey where friction is already visible and costly. For many insurers, that means FNOL, claims intake, underwriting and quoting, or renewal – basically, all moments where delays, repeated questions, and unnecessary handoffs affect both customer confidence and the cost of serving them.

Choose one high-volume problem and define the outcome before choosing the tool. For example, an insurer might use AI to check whether a new claim contains the required information, answer routine status questions, or prepare a concise case summary for the claims handler. The goal should be specific and observable, whether that’s fewer repeat contacts, faster handling, less rework, or a smoother transfer to a person.

Then look beyond the individual task. If an AI assistant gathers information but the next team cannot see it, the customer still has to start again. That’s why insurers need a reliable, consent-aware trail of each customer’s interactions – what they have already shared, which channels they have used, and where they are in their journey – to recognize them and respond in context. That means connecting the channels, teams, and data around the chosen journey so that information follows the customer from self-service to the contact center and beyond.

One of the best things about AI is that, with minimum effort and cost, it can quality-check every customer interaction. And the same core standards can be applied to human and AI agents, while adding checks for risks specific to AI, such as drift and hallucinations. The resulting insight creates a feedback loop that shows where interactions break down and helps improve both service quality and self-service performance over time. The practical application is to learn from the first implementation and expand only when the improvement is clear.

Turn AI ambition into real outcomes

Capgemini combines deep insurance knowledge with the technology and delivery capabilities needed to create satisfying outcomes. We don’t just deliver a proof of concept. We help design, implement, and run the program. We’re known for our outcome-based commitments, which can be tied to specific, agreed metrics, with terms reflecting whether we meet or exceed them.

Download our latest CX report to see how customer expectations, AI adoption, and human interaction are evolving. Then schedule a personalized presentation to explore the data and implications most relevant to your organization. Let’s turn those insights into practical priorities for your CX strategy and ensure you only pay for measurable success.