‘A’ for AI, 'I' for insurance (1/3) - how Aviva is betting on Machine Learning investments to deliver advantage in a tech arms race.

The global insurance sector has seen an 87% year-on-year increase in artificial intelligence deployments, according to research by AI intelligence and analytics Evident. And as we all look around for real-world exemplars of AI in practice, around 40% of insurers already report tangible business benefits from AI, with 77% of those benefits coming from productivity gains and five percent linked to revenue growth.

Those stats are based on research covering 30 leading life, property and casualty (P&C), composite and re-insurance groups across North America and Europe, a study which has also picked up on a significant interest in agentic use cases. According to Evident’s data, over a fifth (21%) of deployments in the last three months of last year  were agentic in nature, with the tech deemed well-suited to handling the unstructured data and complex workflows involved in claims management, which accounts for more than half (56%) of agentic deployments to date.

Elsewhere, McKinsey pitches that generative AI represents one of the most significant opportunities for value creation within the insurance industry with the potential to unlock between $50 billion and $70 billion of revenue, with the tech   particularly well-suited to document-heavy processes, such as policy issuance and claims handling.

The main benefits will be seen in Marketing and Sales, customer operations and software engineering, McKinsey’s analysis suggests, as well as enabling simplifying processes such as quoting, filling in application forms and processing endorsement on carrier websites. In the future, it predicts, agentic AI will be able to handle renewals for simple risks with limited human intervention.

So all good, yes? Well, mostly. Bank of America’s Global Research arm does throw some cold water on all this with its estimates that more than $15 billion in insurance industry commissions are ripe for AI dis-intermediation, being considered “low complexity” and that could mean cost-savings for employers on the one hand, but a less appealing set of prospects for trenches of insurance company employees…

Aviva's AI in action

So that’s the theory. What’s the reality on the front line. Let’s look at three exemplars from the sector over the next three days, beginning with Aviva, a UK-based multi-national serving 25.2 million customers across Britain, Ireland, and Canada.

AI is going to be transformational for the insurance sector, says Group CEO Amanda Blanc, a sentiment we’ll see repeated across all three of our exemplars, and Aviva is well-placed to take advantage of what’s coming, she argues:

There are key enablers that you actually need to drive the value. It is not enough to just have the technology. You need access to millions of customers, the ability to deploy and reuse at scale, capacity to invest, and most importantly, proprietary customer and claims data. Aviva has all of these in spades, and our diversified model is more resilient for any disruption.

AI is not something new for Aviva either, she says:

We have been using traditional AI capabilities for over a decade now. In fact, over 98% of retail business in UK Personal Lines is priced with Machine Learning. And we have been training over 150 Machine Learning models in claims with our own data for years. Generative AI and agentic are just the next steps on this journey.

As a result of previous investment, many of the necessary AI-ready foundations are ready place, she pitches:

We have built an in-house platform to deliver use cases at speed, and we are already seeing tangible benefits. We have halved the time taken to review each case in medical underwriting. And we have also reduced call wrap times by 20% for customer service agents in Direct Wealth, which we are now rolling out more broadly in IW&R (Insurance Wealth and Retirement).

All of Aviva’s employees have access to AI tools, she adds, but it’s still early days in terms of what the organization can achieve she suggests:

We are proud of what we've achieved so far, but we are aiming much higher and always balancing ambition with pragmatism.Our focus now is on prioritizing progressively bigger end-to-end opportunities where AI can transform areas like customer engagement and distribution, underwriting and claims, right through to back-office operations. This is the kind of change that will shape Aviva's future. And some of this is closer than you think.

She points to the area of UK General Insurance claims to make her case:

We have already saved nearly £100 million through our claims transformation and  agentic has the potential to unlock much more. Over the next few months, we will be testing an AI-enabled claims agent, built in-house and launching later this year. This will enable us to handle simple claims from start to finish without human support. And the best part is that this is voice-enabled. Most claims begin on the phone. So this will be transformative for customers, delivering faster, clearer and more consistent outcomes.

There is also the seemingly obligatory partnership with OpenAI to be pitched as well, which Blanc positions as “a really important step for us” (and seemingly every other organization known to man, but hey ho!). She argues:

Combining OpenAI's cutting-edge capabilities with our expertise and data will help us to deliver powerful AI solutions for our customers and our colleagues.

Dis-intermediating

So that’s all good news and upbeat prospects for Aviva. But what about the dis-intermediation impact of AI tech on the insurance sector business model, as alluded to by Bank of America? Blanc says:

I think we've seen dis-intermediation in many places before. Take price comparison websites. I mean that massively transformed the motor market and dis-intermediated to almost a whole extent where today, 95% of quotes come that way. As a mass affluent player, we are in a perfect position to be able to manage that any potential dis-intermediation.

I still believe that [human intelligence] advice will be there, I just think that the advisers will be given better information, more support, and they'll spend more of their time with the customers, where the customers want that face-to-face advice.

But I think for those many people, 91% of the population today that don't take advice. AI will facilitate the ability to be able to do that and mean that they will get better guidance. You've got 12.5 million people in the UK today that do not save enough for their retirement. I think it gives a real opportunity to be able to do that.

My take

I would say we're bordering on the sort of excited end of the scale in terms of the opportunity that [AI] provides Aviva.

A curiously low-key way for Blanc to sum up what Aviva is clearly viewing as a major competitive leverage weapon in what she describes as “an arms race” in the insurance sector.

Next, up - AXA.