By Andrew Harrington, CIO at Ripe
Artificial intelligence has evolved rapidly in recent years, and has now become a part of daily business operations. It is transforming all sectors – including the insurance industry – at a remarkable pace. The potential of AI is enormous. But with all this rapid change, many companies are rushing to adopt new technologies simply because they feel they should, rather than having a clear plan or understanding of what they want to achieve. The result can be an incredibly expensive folly – attractive but not delivering sustainable business value.
I am a firm believer that any technology solution should only be adopted if it solves a meaningful problem. There’s no long term-value in launching tech for tech’s sake. The best results come when you start with the end goal and work backwards to find what solutions can best help you get there. This process means asking the right questions from the outset: ‘What is the problem we are trying solve?’ ‘What is the opportunity we are trying to create – for the team, the customer or the business?’
The hurry to shout ‘AI’
The scale of the rush to adopt AI is staggering. 99% of insurers reported they were either already investing in generative AI capabilities or were planning to do so according to EY’s 2024 generative AI survey, yet more than half (57%) expressed uncertainty about inadequate ROI, data accuracy and integrating with legacy systems. In addition, insurance is ranked highest in the ‘nervous or sceptical’ about AI category (25%) among financial services sectors.
This disconnect between investment enthusiasm and underlying scepticism risks half-baked AI implementation. In the race to prove they are cutting-edge, many insurtechs talk a good talk in terms of what they have achieved in AI but dig a little deeper and there is no substance behind these claims. For example, I have heard many companies simply buying CoPilot licences for their employees and calling it an AI strategy.
The pitfalls of ‘AI FOMO’
Too many insurtechs suffer from “AI FOMO” – desperate to implement artificial intelligence because competitors are doing it, rather than based on a clear business case.
Without a defined purpose, these companies often find themselves stuck with fragmented systems and tools that make things more difficult and less streamlined. Some businesses implement chatbots that frustrate customers more than they help or deploy machine learning models that produce insights nobody knows how to act upon. The real value of technology only appears when it’s closely tied to commercial goals and has been integrated properly.
But there’s a better way. Rather than asking “How can we use AI?”, we ask “What problem are we trying to solve and is AI the right tool?”. Often, a simple automation or improvement delivers better results than any complex AI implementation would.
Starting with the outcome
This ‘build backwards’ process has proven incredibly valuable in practice. Consider a common challenge: customer service agents spending excessive time on routine verification tasks. Rather than at once implementing an AI solution, working backwards from the desired outcome – agents having more time for complex customer issues – reveals multiple solution paths.
The answer might not be AI-first. Streamlining processes and automating verification tasks can deliver immediate value. AI tools can then be introduced where they add genuine benefit. For example, transcription tools that convert call centre conversations into structured data, offering real insight into customer contact drivers, or intelligent systems that help customers get quick answers to policy questions.
The goal should be augmenting human capabilities rather than replacing them. For instance, we’re embedding agentic AI into our customer journeys, not just to surface information, but to complete tasks and reduce friction. Crucially, this adoption doesn’t play into the false stereotype of reducing headcount but supports our agents to upskill and become experts in assisting customers with more complex tasks, giving customers enhanced confidence in our knowledge and ability.
The importance of foundations
However, having the right method is only half the battle. What truly decides whether AI initiatives deliver value is something far less glamorous but infinitely more critical.
Any successful AI implementation depends on the quality of your systems, but more importantly the quality and integrity of your underlying data. For instance, at Ripe we’ve built our platform, Juice, to keep clean, structured data – ensuring 99.9% data accuracy through monitoring tools. This foundation enables easy integration with both internal tools and third-party services, giving us a strong base to build and scale AI across our business.
Data quality is paramount. Without high accuracy and sophisticated monitoring to show improvement areas, even advanced AI tools will produce unreliable results. Many insurers struggle here because legacy systems have evolved over decades into patchwork solutions. While these have valuable business logic, they’re often difficult to support and notoriously tricky to change.
However, effective transformation doesn’t require tearing everything down. It’s about evolving your existing foundations, layering modern architecture on top of stable core systems where you can add flexibility without disrupting day-to-day operations.
Preparing for what’s next
Looking ahead, the industry will continue to shift. Trends like usage-based insurance, real-time pricing and embedded products will only grow in relevance. The rise of embedded insurance – where coverage is seamlessly integrated into apps and websites – means insurers must develop flexible, API-driven products that can be offered now they’re most relevant to customers.
But chasing every trend isn’t the answer. The insurance companies that succeed won’t be the ones who adopt the most tools or shout the loudest about AI. They’ll be the ones that are able to use AI to differentiate, delivering frictionless solutions to problems that customers did not realise they had.
Smart implementation requires deep understanding of your data, clear goals and human oversight. Done right, AI enhances customer experience and operational efficiency. Done wrong, it creates customer frustration and regulatory headaches. Governance is still crucial with strict guardrails on data access, clear human oversight protocols, and regular auditing of outputs being key to success.
Truly understanding the risks associated with each AI implementation means you can predict what the worst-case scenario might be. This enables you to build protections around those processes to ensure no false decisions or actions from AI slip through the net.
The bottom line
The companies that will thrive are those that view AI as one tool in a broader technology stack, not as a silver bullet. They’ll be the ones that start with clear commercial aims, work methodically to find the right solutions, and implement technology with discipline and purpose.
At Ripe, we’ve learned that it’s essential to develop AI capabilities that provide real value. At the end of the day, innovation isn’t about showing off the latest tech. It’s about doing the right thing, with purpose, for your customers. FOMO might drive headlines, but it won’t drive sustainable business value. Strategy will.