Authored by Jaganathan Chelliah, Co-Founder & COO, Motiv AI
Usage-Based Insurance was supposed to be insurance's great reinvention. Europe — the world's most advanced telematics market with a 32% share of global UBI policies — had 13.8 million telematics policies in force by end of 2024. The pattern of underperformance, however, is global. In North America, 20.2 million telematics policies are active yet fewer than one in three enrolled policyholders see their premium fall — because ACB-based scoring alone is not moving loss ratios.
In India, despite IRDAI's PHYD mandate opening the market in 2022, insurers still score almost entirely on hard braking and harsh acceleration, missing the anticipation dimension that determines whether a near-miss was caused by recklessness or avoided by skill.
Yet scale has not solved the industry's fundamental problems. Only 17% of European insurers currently offer telematics-based motor products — meaning the vast majority are still pricing risk on age, address, and vehicle type rather than actual behaviour. Insurers that fail to offer competitive UBI programmes by 2026–27 face the prospect of adverse selection as safe drivers migrate to UBI-focused competitors, leaving traditional books with disproportionately risky portfolios.
Traditional telematics — logging hard brakes and late-night miles — has not resolved any of these pressures, because it tells insurers what a driver does, but rarely why — and never what could have been done differently five seconds earlier. It is also a structural issue, since risk is pooled in insurance: gaining access to individual-level risk is, in itself, a revolution. Prudence-based driving assistance (PBDA) changes that equation entirely.
The cause? Data
Enter prudence-based driving assistance (PBDA): an AI-driven layer that doesn't just log behaviour but actively coaches, anticipates, and intervenes based on the road context in relation the vehicle dynamics . In 2026, the convergence of affordable edge compute, mature V2X infrastructure, and LLM-powered in-cabin advisors is turning PBDA from a promising concept into the backbone of next-generation UBI.
From Event Logging to Decision Scoring
Traditional telematics records events — a hard brake here, a sharp corner there. These are valuable but fundamentally backward-looking. Prudence-based systems take a different approach: rather than cataloguing incidents, they score the quality of decisions in real time.
A PBDA module continuously models road geometry, traffic density, weather conditions, and time of day, then evaluates whether the driver’s speed — and, where hardware permits, lane discipline, acceleration, and following distance — are appropriate for the conditions, not just legal limits.
This matters because the same action carries vastly different risk depending on context. Travelling at 80 km/h is unremarkable on a clear motorway but imprudent in heavy congestion or on a rain-slicked curve. Equally, a following distance that feels comfortable at low speed becomes dangerously short at motorway pace — a gap that hard-braking events alone will never reveal until it is too late.
By scoring the decision against the context in which it was made, PBDA captures the dimension of driver behaviour that traditional telematics has always missed: not just what the driver did, but whether it was the right thing to do given everything happening around them.
Highway driving carries substantially higher loss magnitude than urban driving; night driving is several times more dangerous than daytime. A telematics score that ignores this context is, at best, an approximation.
The key distinction is temporal. PBDA measures the gap between what a prudent driver would have chosen given current conditions and what this driver is actually doing. That contextual delta is a far stronger predictor of future claims than any post-hoc event count — because behavioural variables are causally related to crash risk in ways that age, gender, and ZIP code simply are not.
Breaking the Adverse Selection Loop
UBI has always suffered from a structural trap: safe drivers opt in to earn discounts; risky drivers opt out to avoid surcharges. The resulting portfolio skews low-risk, eroding the very premium advantage that made UBI attractive. PBDA breaks this loop by changing the product's fundamental value proposition. An app that scores you after the drive is surveillance.
An assistant that coaches you through a difficult merge in real time is a co-pilot. Pilots of active in-cabin coaching systems have recorded reductions in at-fault incidents of 30% or more within six months — not because drivers fear the score, but because real-time, contextualised feedback rewires habit faster than any end-of-trip report.
"The future of telematics isn't just about capturing data and offering a discount — it's about feedback and thoughtfully designed offerings that make driving measurably safer."
When PBDA actively reduces exposure, carriers can underwrite even moderate-risk drivers competitively. The adverse selection loop inverts: the insurer's loss ratio falls not just from better selection, but from genuine risk reduction across the entire portfolio.
What Motor Insurers Gain
82% - Policyholders who'd recommend telematics with feedback & crash assistance
45% - Of drivers who made significant safety changes after joining a telematics programme
60% - Of policyholders open to switching insurers for a better UBI product
At the underwriting level, continuous behavioural data enables dynamic premium repricing at weekly or per-trip intervals — dramatically reducing reserve uncertainty. At the claims level, PBDA-equipped vehicles generate rich pre-incident data streams that accelerate accident reconstruction and allow carriers to trigger crash detection alerts and initiate the claims process before the driver has left the scene.
Insurers leveraging high-quality telematics and dash-cam data have reported incident exoneration rates approaching 96%. For reinsurers, a PBDA-enriched portfolio is a structurally different risk class — one with documented, ongoing behavioural improvement baked in.
The Explainability Dividend
One of UBI's persistent failure modes has been opacity. Drivers who see their premium move but can't understand why quickly lose trust and opt out. PBDA-powered conversational assistants fix this directly. Instead of a driver receiving a vague notification that their score dropped by 4 points, the system tells them: "You braked sharply three times during your office-trip this morning — leaving a little more distance from the car ahead could help."
That kind of specific, moment-linked feedback is something any driver can act on immediately. When people understand exactly what moved their score and how to fix it, engagement and retention both climb.. Among drivers under 53, over 90% say they'd recommend an insurer that delivers this experience — a signal that personalised, explainable UBI has crossed firmly into mainstream appetite.
Challenges and the Road Ahead
None of this unfolds without friction. Second-by-second decision scoring collects intimate data on movement, habit, and environment requiring robust consent frameworks, auditable algorithms, and genuine on-device processing options. Equity is equally pressing: if PBDA is only viable in newer connected vehicles, the drivers who would benefit most from behaviour-based pricing are systematically excluded. Smartphone platforms and OBD-II retrofits are closing this gap, but intentionality is required.
By year-end 2026, UBI is making a fundamental transition from passive measurement to active partnership. The carriers best positioned to lead it are not those with the largest datasets, but those that deploy PBDA as a driver development platform where the insurer's interests and the policyholder's point in exactly the same direction. Safer driving means fewer claims, lower premiums, and higher loyalty. That is not a trade-off. It is the UBI flywheel, finally spinning at the speed the industry has been waiting for.
𝐒𝐭𝐚𝐲 𝐢𝐧𝐟𝐨𝐫𝐦𝐞𝐝 𝐰𝐢𝐭𝐡 𝐨𝐮𝐫 𝐥𝐚𝐭𝐞𝐬𝐭 𝐮𝐩𝐝𝐚𝐭𝐞𝐬 𝐛𝐲 𝐣𝐨𝐢𝐧𝐢𝐧𝐠 𝐭𝐡𝐞 WhatsApp Channel now! 👈📲
𝑭𝒐𝒍𝒍𝒐𝒘 𝑶𝒖𝒓 𝑺𝒐𝒄𝒊𝒂𝒍 𝑴𝒆𝒅𝒊𝒂 𝑷𝒂𝒈𝒆𝐬 👉 Facebook, LinkedIn, Twitter, Instagram
