In a softer market, the ability to demonstrate a clear, differentiated view of exposure is becoming a genuine advantage at the negotiating table. Reinsurance News sat down with Laurent de la Porte, CEO of Allphins, and Ed Messer, Head of UK Analytics at Gallagher Re, to discuss how real-time, granular data is reshaping decision-making, where the technology is headed next, and what it will take to bring alternative capital into the specialty space.
From annual aggregation to real-time visibility
Discussing how granular, real-time data is changing the way brokers and reinsurers make decisions, particularly in a softer market, de la Porte said this is one of Allphins’ main strengths.
“We’re always limited by the data, as any tool is. But we make the most of its detail and frequency: normalising it, enriching it, and turning it into complex exposure calculations very quickly. In a softer market, that speed lets brokers and reinsurers differentiate between risks rather than price on averages.”
“Allphins’ use cases aren’t so different in a hard or soft market. For reinsurers, it’s about portfolio management, typically by avoiding overexposure across perils, lines of business, geographies and industries, and live event tracking.
“At the broker level, we see more diving into the cedent level. In our discussions with Gallagher, for example, there’s a real need to fully understand what the cedent’s book is and how it’s evolved versus last year. They want to be able to tell a story about the change, what’s better, why it’s a good book, and that’s where granularity is very important for them.”
Two decades of specialty analytics
Messer traced how far the discipline has come. “I’ve been working in specialty analytics for about 20 years now. Back then, the state-of-the-art methodology was pretty crude market-share approaches. We’d collect annual aggregate statistics from Lloyd’s and other bodies, work out total market exposure, and take a market share of that for things like accumulations at ports or who’s on what energy facility. No one really knew; it was just a market view.
“We’ve closed the gap on property over those 20 years, though specialty is less advanced given the complexity of the risks. Today we’ve got satellite and vessel tracking, energy pipeline networks mapped out, terrorism targets and attack vectors understood, industry exposure datasets built, and software like Allphins to aggregate those exposures.
“What that’s done for reinsurance broking is change the unit of decision-making. Back in the day, we looked at accrued market loss percentage, generally after an event. Now we’re thinking about asset-level identification and pricing pre-bind: understanding what cedents are exposed to, and the marginal impact of adding a risk to a portfolio. We couldn’t have done that 20 years ago; it’s been a big inflexion point.
“In a hard market, adequate pricing can hide a multitude of sins within a portfolio. In a softer market, the marginal impact of adding or removing a risk becomes even more important when you’re trying to differentiate cedents. Looking forward, it’s about speed and accuracy: giving cedents a timelier view of their exposures into a renewal, and relaying that to reinsurers in the best possible light.
“If you look at that timeline, that’s probably been the biggest change over the last 20 years, and I think that progress in the technology we’re deploying will only continue.
“There’s been a fair amount of loss activity over the last couple of years, which makes differentiation even more important. Reinsurers are looking at which cedents are monitoring their exposures sensibly, both before an event and during a live one. We had the conflict in the Middle East during a live renewal period, for instance, and having tools like Allphins made identifying those exposures much easier than it would have been otherwise.”
The next frontier: casualty, cyber and cross-class exposure
Allphins initially focused on property, energy and terrorism. De la Porte pointed to casualty and other company-based lines of business as the next opportunity.
“Where we see ourselves going is where our clients take us. We’ve seen a clear trend over the last six months, confirmed at Monte Carlo: the need to finally manage exposure for casualty and other company-based lines of business. That means cyber, political and credit risks.
“Clients usually don’t have a good system for managing these aggregations and are doing a lot of manual work. These are non-modelled lines of business. You have some models, but nothing well established, so you look at aggregation instead, because that’s what speaks to the risk. Clients look at exposure per company, and at stress tests on how companies in a certain industry, geography and size are impacted by a given event. These stress tests are often built by the reinsurers themselves, as a kind of analysis grid. For that, you need good data and calculation power, and that’s where Allphins helps.
“Our added value is in digesting the data: standardising it, filling gaps like missing industry classification or employee numbers, and enriching it. Visualisation matters a lot too: it’s good to know you’re exposed to a certain company at 50 million, but where are you attached? That’s a complex analysis in itself.
“These lines also generate cross-class exposure. A real scenario might involve casualty, credit, cyber and even property, in the case of a data centre. Being able to operate across several classes and provide visibility across them is part of the value. We already have clients using it this way, and these lines of business react differently through the cycle too.”
Closing the gap between modelling and exposure management
Messer picked up on de la Porte’s point about the gap between modelled and non-modelled classes.
“I agree with his views on the extensions to other classes of business, but it’s worth picking up on one point Laurent made. Right now in the specialty market, aggregation of exposure is how we think about potential risk. But if you look to the future, at how these technologies could improve not just the risk selection process but the whole value chain: exposure is just one element of a risk model. Using technology to understand both the hazard- the frequency and severity of a potential loss event- and the vulnerability, what a loss could look like based on the asset’s characteristics, are probably the two natural extensions across all classes.
“It’s not as straightforward as cat modelling, because specialty deals with man-made perils, terrorism and political violence, where an adversary is choosing a target, so the hazard is intentional rather than meteorological. There are moving assets too, which adds complexity. But there’s a huge amount technology can bring, and you can see a shift, in the not-too-distant future, from pure exposure aggregation to asset-level expected losses.”
De la Porte agreed, noting where the two disciplines still leave a gap. “There’s modelling on one side and exposure management on the other, but there are areas that sit in between. I’m not pre-empting where the best approach will come from, but there’s a gap in between where technology could help.”
Where AI actually helps
De la Porte was clear that clients come to Allphins with problems, not technology requests.
“Our clients are very pragmatic. They don’t come to us asking how to use AI; they come with pain points, and for some of them AI is the best or only solution. You can only leverage AI if you have a very good system of record for your risks and policies, plus the calculation power to make something of it. Because we have that at Allphins, we’re an AI enabler. You can use AI once you have that foundation. Clients put all their exposure and policies into Allphins, so the data is captured and standardised, and you can build AI on top of it.
“The main use case today is digesting client data. AI is better than anything else for that. What we’ve found is AI needs to be optional, auditable and visible: you don’t want half of your processes performed by AI without you choosing it, you need to trace back what it did, and whenever there’s an AI output in Allphins, there’s a small marker that says so. I think it can add a lot of value in data digestion, and potentially soon in interrogating the data for faster insights.”
Messer said Gallagher Re thinks about AI’s benefits in two categories: growth and efficiency.
“Growth is about using AI to identify trends and signals in the interconnectedness of our clients’ portfolios across classes: things we might not have spotted, like how certain data centre coverages clash with other classes. If you can better understand that interconnectedness, you can potentially find new sources of capital to connect to the risk.
“Efficiency is more obvious: speeding up analysis and surfacing important data points to decision-makers faster, from submission to quote to bind, without diluting quality. The governance matters: AI applied to a poor process just makes a poor process faster, so you need structure and the right governance in place.
“One other thing: as a broker, we generally get datasets from clients once or twice a year, so overlaying real-time capabilities onto a fixed point in time dilutes the value slightly. I can see this extending into a direct API feed into clients’ underwriting systems, so we get much more real-time access, and our job would be to surface that to reinsurers in the right way to get the best deal for our clients.”
What Allphins delivers at renewal
Messer pointed to four areas of advantage: aggregation, choke points, visualisation, and event response.
“Exposure aggregation is a key one. What does adding a risk do to a portfolio, particularly when rate adequacy is strongly contested? That marginal impact assessment is important, and Allphins lets us look at that.
“We’re also able to look at choke points of exposure across our portfolio, where the peak aggregates are and how they stack up across classes. That’s particularly important when structuring reinsurance programmes, looking at retentions and limits relative to aggregated exposure. Visualisation turns an aggregation point into something a reinsurer can actually see and interrogate geospatially, which helps with negotiation.
“Event response is another: events don’t follow a cedent’s renewal calendar. The Middle East ran straight through a very important renewal season, and being able to tell a client, on the same day, what they could potentially be exposed to is a game changer.”
The next two to three years
De la Porte pointed to two directions: deeper, more frequent analytics, and AI-driven personalisation.
“Continuous exposure management, or continuously updated live analytics, is one direction. When we started Allphins, exposure management was a once-a-year, post-bind exercise for insurers. They’d receive the data pre-binding, of course, but didn’t have time to really look into it, and would aggregate it post-binding instead. What’s changed since then is that analytics are now done pre-bind, with brokers like Gallagher Re performing the analysis and reinsurers using it to understand the cedent’s book and the incremental impact of their decisions.
“From here, I see two directions. First, analytics will be deeper, more granular and live, always knowing your exposure, with data accessible in a few clicks. One limitation is how much data players are willing to share along the value chain; it’s not always obvious you want to disclose exactly where you’re attached. That’s normal. The market is a bit like us as individuals; there’s only so much we’re happy to share.
“The other aspect is AI. Maybe AI will change the world in 10 years, but being realistic about the short term, I think what you’ll have is analytics custom to your needs and strategy, adapting to your story and immediately available regardless of the quality of the data you receive. We’re starting to see this already, with some of us experimenting with agents and playing around with Claude on small amounts of data. But with big systems of record like Allphins, and AI built around them, people will be able to access the analysis they need to make decisions. I don’t think decisions will be made by AI in this market, but the right data will be provided so the decision-maker can make a better-informed, or even a fully informed, decision.”
Messer closed with three ways he expects analytics to reshape the market over the same period.
“Better exposure capture at the point of writing should feed straight through to pricing, and from pricing into reserving, so we stop seeing the deterioration after major loss events that we so often see today. Deterioration can happen when the market didn’t know what it was exposed to at the time it wrote the business, as well as how event definitions might play out.
“Specialty exposure software should become common currency for risk trading through the value chain, the way cat models did in property. What cat models really delivered wasn’t accuracy; it was a shared language: cedant, broker, reinsurer, rating agency and investor could all argue about the same numbers, and the argument became productive because everyone accepted the currency. If specialty gets that, the market could move from negotiating over whose spreadsheet is right to negotiating over the risk itself.
“More data and better insight should open more options for alternative capital to be deployed in specialty, because what investors need before they commit is a modelled, comparable view of the risk. Capital isn’t the constraint today. The cat bond market set records through the first half of 2026, and alternative capital reached around USD 147bn in H1 2026, almost all of it in property catastrophe. The reason is modellability, not appetite. If technology can give investors a defensible view of specialty risk, capital has somewhere new to go.”
The post From annual aggregation to real-time visibility: Allphins and Gallagher Re on the next era of exposure management appeared first on ReinsuranceNe.ws.
