Telematics was meant to change motor insurance fundamentally.
The premise was simple. Motor insurance had historically relied on proxy indicators and pooled cohorts of risk such as where they live, age, claims history, vehicle details, and numerous other factors. Telematics offered the possibility of something fundamentally different: moving from predicting behaviour based on characteristics to observing behaviour directly. Instead of asking whether a driver looked risky, insurers could understand how they actually drove.
The attraction for insurers was obvious. Better information will create better pricing.
The attraction for customers was also clear. A fairer approach to pricing. Drive well, and you get a cheaper price, drive badly and you get a high price. A responsible driver should not necessarily pay more simply because they belong to a higher-risk segment.
Nearly three decades later, the reality is more complicated.
Telematics has unquestionably changed insurance. It has improved underwriting capability, created new approaches to risk assessment, enabled insurers to understand driving behaviour in ways that were previously impossible, and help in the event of an accident. But it has not produced the wholesale transformation of motor insurance or pricing that was hoped.
The question is not whether telematics worked. The more interesting question is why a technology with such obvious potential has delivered a limited transformation.
I think it is less about the quality of the data, and more about what the insurance market did with it.
The earliest telematics experiments focused heavily on the technology itself.
In the mid ‘90s, the US insurance firm, Progressive, began exploring usage-based insurance through products such as Autograph. It was a clunky device that needed professional installation. Similarly, General Motors started development with connected car with the launch of OnStar. Both programmes focused on safety and assistance rather than insurance pricing. Both were early proof that the hardware worked. Neither proved that anyone would pay for it.
The early challenge was practical. Hardware was expensive and installation created friction. Customers needed to understand why they should accept monitoring in exchange for a possible financial benefit.
The UK experienced this in 2006 when Norwich Union (Aviva) launched its pay-as-you-drive product. The concept generated significant interest. Customer satisfaction was high, and more than 90% of pilot customers renewed1.
However, the economics proved difficult. The cost and complexity of hardware installation, combined with limited customer adoption, meant the proposition could not achieve the scale required. Norwich Union withdrew the product in 2008.
The lesson was not that telematics did not work. The lesson was that insurance propositions succeed when the ‘holy’ trinity of technology, product, and commercial economics align.
Progressive continued developing its approach through multiple iterations, including TripSense and MyRate, before launching Snapshot nationally around 2010–2011.
By 2011, Progressive had analysed more than two billion miles of driving data and enrolled hundreds of thousands of drivers. By 2015, Snapshot had grown to approximately three million customers.
The significance of Progressive’s approach was not simply the volume of data collected. It was the strategic decision to embed telematics into the insurer’s broader pricing capability. Telematics became a way to improve risk understanding across the portfolio rather than a standalone product aimed at a narrow customer segment.
That is an important distinction. The data was not simply being used to determine whether a particular driver should receive a discount. It became part of the pricing and underwriting capability that could influence the whole motor portfolio going forward.
The investment therefore had the opportunity to compound. More data created better understanding. Better understanding improved pricing and underwriting. And the benefit wasn't restricted to the customers who had bought a telematics policy.
The UK market evolved differently. Here, telematics became strongly associated with young drivers. The problem was clear. Drivers aged 17–25 faced some of the highest premiums in the motor market. Insurers needed better ways to distinguish between higher and lower-risk customers within that segment.
Companies such as Insure The Box, Ingenie and Young Marmalade built propositions around monitoring driving behaviour and using that information to manage risk. Later, larger insurers including Admiral, Hastings and Direct Line introduced telematics propositions, while newer entrants such as Ticker explored more flexible approaches, including pay-how-you-drive and pay-what-you-use models.
However, the underlying market positioning remained different.
In the US, telematics became part of a pricing capability and the portfolio as a whole.
In the UK, telematics became a product subset.
Same technology. Two completely different approaches. The difference is pricing capability can influence the whole portfolio. But a product category must compete for resources, customer attention, distribution and profit.
Insurers wanted telematics for one reason: better risk selection. Fair enough — that is their focus.
But customers don't buy insurance to improve somebody's actuarial model. They buy it because they want a fair price and confidence their insurer shows up when it matters.
The first generation of black-box products solved the insurer's problem — better segmentation for a hard-to-price segment — while leaving the customer holding the trade-off: give up your privacy, get a better price (and only if you drive safely and accept the restrictions). This approach works when the alternative is a very high premium. If you're a 17-year-old facing an otherwise unaffordable motor insurance quote, sharing your driving data in return for a discount makes sense. But it is a difficult basis for building a mainstream customer proposition.
The ambitious and transformative vision is to flip the business model. Make the relationship dynamic rather than punitive. Rather than simply monitoring risk, insurers could reward behaviour, provide greater transparency and create a dynamic relationship between customer and insurer. The value exchange is better underwriting information for the insurer and a successful customer proposition with a clear economic utility value.
Several businesses explored this direction.
Ticker developed propositions around pay-how-you-drive and pay-what-you-use models, moving closer to a concept where behaviour and usage could directly influence the insurance experience.
Around this time, I helped launched Kudo Insurance, a telematics smartphone app proposition, that used underwriting selection as part of the product design, featuring rewards as incentives for safer driving. It was designed to create a transparent relationship between driving behaviour, insurance price and product value.
The technology could support the model, and there was a compelling reason for the customer to engage and stay engaged.
Insurers trialled it and showed the observed claims frequency among scored drivers was materially better than market standard, and customer feedback was electric.
What we couldn't solve fast enough was distribution and getting enough capital runway.
Investor feedback was that you cannot beat price comparison websites.
No investors, no funds, no more Kudo.
Vitality provides another interesting example. Vitality Car launched in 2021 with Covéa as the underwriter. Vitality had already spent two decades demonstrating that it could build a successful engagement-and-reward model in health insurance. It had the brand, the customer proposition, the behavioural data expertise and the credibility to make the model work. If anyone could make a genuinely customer-centric, reward-led motor proposition work in the UK, Vitality was surely one of the better placed businesses to do it.
Sadly, Vitality Car was withdrawn in 2023. Claims inflation was cited as the reason. That was probably a smaller part of the story. It is difficult to ignore the broader commercial problem. That a well-funded business, with a strong brand and a proven behavioural-reward model, struggled to get a genuinely non-price-led motor proposition off the ground.
Perhaps the wider problem with UK motor insurance is the market, not the proposition.
The UK motor market has a unique structural feature: price comparison sites.
These firms transformed motor insurance, making it easy for customers to compare prices and switch providers. This created significant consumer benefits and increased competition. However, it also changed the way insurance products are presented.
A comparison journey naturally prioritises variables that are easy to rank: Premium is the focus, then secondary features such as excess and other basic cover features.
Telematics propositions are more difficult to represent. Their value is often behavioural and long-term. The benefit may not be a lower price today. It may be improved pricing over time, rewards for safer driving, greater transparency or a different relationship between insurer and customer. The value exchange of a telematics proposition is difficult to communicate in a price ranked list.
The structural tension is clear. Insurers can develop differentiated propositions or services, but if the dominant distribution channel rewards price only, the commercial advantage of differentiation is lost to the customer.
I remember once, at a large UK insurer, being sat in a meeting with the marketing team who categorically told me: “Only price works.”
My response: “Well, in that case, why don't we remove (the seven-figure) marketing budget and invest the saving in price?”.
Fun meeting.
This is not a criticism of comparison websites, or how motor insurance marketing and product teams adapted. Every distribution channel creates incentives, and those incentives influence how marketing and product teams adapt.
For telematics, the challenge was not only creating better risk information. It was creating a market environment where that information could translate into customer value.
If the distribution channel rewards the cheapest acceptable proposition, it becomes difficult to persuade customers to choose something differentiated, particularly when the value of that differentiation is difficult to quantify at the point of purchase.
Agentic search is a genuinely interesting new development, and shopping behaviour is changing rapidly.
MoneySuperMarket have launched their own version, and Aviva followed suit with a live quoting app inside ChatGPT, built with OnMarrow, and they have started with home insurance. Two of the major players in the UK market exploring different avenues of a conversational AI type solution as a route to market.
OnMarrow, trading as Marrow, is building the compliance layer and infrastructure that lets AI agents quote, compare and bind a regulated insurance policy. If they are successful, then the customer’s personal agent can act like an ‘old school’ broker. Comparing products, policy wording for suitability, and trading pure price for the product that meets the customers’ needs, and what they believe represents value.
The consultancy market and market research firms are positively brimming over in anticipation to launch services supporting Agentic/AI distribution channels.
GlobalData have stated that UK consumer comfort with chatbot-generated insurance quotes climbed to 41.9% in 2025. The error margin on that research would have to be astronomical for this distribution channel to be accused of being a novelty that will not scale.
The price comparison website limiting factor is that it cannot easily compare wordings, service, or rated insurance. Industry professionals know that not all insurance paper is equal, but often the customer doesn’t have clear visibility of that.
An AI agent doesn't have that constraint. Ask it to find you motor cover and it can weigh price against claims service, reputation, financial strength, scale, and importantly detailed product features and benefits, any restrictions, and the value of expected payoffs, all in a single conversation and a summary with a personalised recommendation. Eventually without the hallucinations.
So, perhaps this is where reward-based telematics propositions may take off. Ticker, Kudo, and VitalityCar had to fight against a dominant distribution channel, built to compare one number. If customers start asking an agent to find them the right cover instead of the cheapest one, the whole basis of comparison changes, and the propositions that got squeezed out by price-ranked lists get a genuine second look.
Telematics has sometimes been described as a technology that failed to deliver on its promise. I disagree. The technology did exactly what it promised. It gave insurers better information about the insurance risk.
The problem is distribution, where the focus point is one variable.
Progressive worked out how to embed that data into a business built to hold it for decades. The UK spent that same period proving good products can still lose to distribution.
Agentic search could create opportunities for propositions that have historically struggled within comparison-led distribution.
What if telematics stopped being something that a 19-year-old buys because their conventional insurance premium is unaffordable? What if it became a proposition that a 45-year-old driver actively chose because the insurer gave them something valuable in return for their data and their behaviour? That would be a very different market.
It would move telematics away from being a niche product towards being a genuinely customer-centric, reward-based proposition that could appeal to drivers across the market and, ultimately, feed the behavioural insight back into the whole portfolio.
That starts to look much more like the Progressive model: telematics not simply as a product, but as a pricing and underwriting capability that compounds across the portfolio.
Understanding the data and understanding risk was never the hard part of telematics. Distribution is and always has been. For the first time in a long time, distribution is changing.
1Source: https://www.transportxtra.com/publications/parking-review/news/12095/norwich-union-abandons-pay-as-you-drive-car-insurance-scheme/