This article is by Tom Philp, Founder and CEO, Maximum Information
In September last year, sitting on a plane as it began its descent into Las Vegas for the 2025 ITC conference, I had the visceral experience of receiving an alert from my bank telling me that our investment had landed following the completion of our seed round. A dangerous alert to receive in Vegas, I know.
It is a feeling I will never forget and one year on returning to that same conference gave me the opportunity to reflect on what has changed. We are roughly one year on from our last round and, if everything goes to plan, roughly one year away from our next.
When we first sought seed investment, the narrative was strong but still early in its development. The proposition we had begun to perfect was catastrophe model middleware, owing to the role MagniPhi plays in the wider catastrophe-risk management chain between the model vendor and the people trying to use those models to make decisions. The problem itself was relatively simple. Catastrophe-risk teams have access to increasingly sophisticated models, but extracting maximum value from them remains unnecessarily difficult. Before an analyst can compare different views of risk or interrogate the assumptions behind them, there is often significant work simply sourcing, processing and standardising the information.
MagniPhi was designed to remove that friction, analysing model outputs consistently and giving users the tools to understand what is driving the differences between them. We were not trying to become another catastrophe model vendor but rather wanted organisations to be able to extract more information from the models they already had and, ultimately, make better decisions with them.
That proposition still rings true today, but over the past year it has grown up with us. The capital we raised gave us the ability to invest properly in the team, technology and ideas around that original proposition. It also meant we could spend more time alongside more organisations, working on a much wider range of problems across the catastrophe-risk landscape.
One of the interesting consequences of doing so is that the closer we have got to the original problem, the harder it has become to see it in isolation. This is just the nature of emergence. Individual parts of a system interact and in doing so reveal something bigger than any one of them alone.
Catastrophe models, after all, do not exist in isolation. They sit within a much wider system of data, science, assumptions and expertise that eventually has to result in a decision. The deeper we have gone into one part of that system, the clearer it has become to see the connections around it. And it is in those connections that we increasingly see the opportunity for Maximum Information.
The starting point
MagniPhi remains our starting point, but we have begun stretching the original proposition further in both directions. If our first question was how we could help organisations extract more value from the models they already use, the obvious next questions were what happens before and after that point. How do you make it easier to access the right model in the first place? And once you have done the analysis, how do you make sure that information reaches the people making underwriting, pricing, claims or capital decisions?
We designed MagniPhi as a vendor-agnostic layer between models, data and users, bringing model outputs into one environment for comparison, evaluation, view-of-risk building and adjustment. Our integration with Model Exchange has pushed that idea further by allowing models to be executed on a pay-per-use basis and accessed through MagniPhi. There is an important principle behind this. The question an analyst wants to ask should determine the model they use, rather than the models they happen to license, determining the questions they are able to ask.
Model evaluation and comparison remain central to MagniPhi. However, we are finding that solving it properly requires us to think more broadly about how models are accessed and how the information they contain eventually becomes useful elsewhere in an organisation.
Connecting expertise with product development
This has reinforced another belief that has been fairly fundamental to Maximum Information from the beginning. Building useful technology in catastrophe risk requires more than understanding the technology. You need to understand the models, the science and the decisions people are trying to make with them. It is why we spread our work across software, consulting and innovation rather than treating product development as something that happens separately from the market.
Consulting puts us alongside teams dealing with these problems in practice. Innovation gives us room to investigate questions that do not yet have an obvious commercial answer, often with academics and others across the industry. Both feed back into the software. The result is a useful cycle – we encounter a problem, understand it properly, test different ways of solving it and then ask whether technology can make that solution useful at scale.
Building alongside the problems
PsiClone is probably the clearest example of where following that process can take you. It started in a different part of the catastrophe-risk chain but with a surprisingly familiar problem. During a live catastrophe event we have at our fingertips sophisticated weather forecasts, catastrophe models, exposure information, vulnerability data and disaster management plans, yet those sources are often being viewed separately by different people trying to understand the same event. PsiClone
brings those perspectives together, combining ensemble weather forecasts with exposure and vulnerability information to ask a more useful question than simply where might this event go, but rather what might its actual impact be? And then, what practical actions can I take to adapt and mitigate the risk as the event unfolds?
The answer depends on who is asking. An insurer might want to understand potential portfolio impact, prepare its claims response or assess its reinsurance position. A government or humanitarian organisation looking at exactly the same cyclone might need to decide where to position resources or whether the evidence is strong enough to justify acting before landfall. These are quite different decisions, but they are decisions about the same underlying risk.
In one sense, that takes us a long way from where we started with MagniPhi. In another, it really does not. For my team, MagniPhi and PsiClone act as two sides of the same coin, built around different moments in the same catastrophe-risk journey. MagniPhi asks how I can better understand the risk within my portfolio, including the different views of that risk and the assumptions behind them. PsiClone asks what happens when that risk begins to develop in the real world. What is changing? What could it mean for me? And, crucially, what should I do about it?
One helps organisations build and interrogate their view of risk before an event. The other helps them understand how that risk is evolving and decide how to respond as it unfolds. Put the two together and we can start to see what it might mean to connect more of the catastrophe-risk journey, including parts of it that extend beyond the traditional boundaries of insurance.
Towards a shared ecosystem
Which brings us back to emergence. The further we have travelled along the catastrophe-risk chain, the less sense it has made to think about each part independently. Insurers, reinsurers, brokers, governments, humanitarian organisations and capital providers may be making very different decisions but, ultimately, they are trying to understand different manifestations of the same underlying risk.
Our ambition is to build the first shared ecosystem for catastrophe and disaster risk management. Not one enormous system that tries to make every decision for everyone, but infrastructure that makes it easier for risk information, tools and expertise to move between the people who need them. In simple terms, we want to make catastrophe risk easier to access, easier to understand and easier to act on. Insurance remains our core market, but we can increasingly see how the same infrastructure can extend into disaster risk management and wider industries.
Perhaps that is the biggest difference between the company that landed in Las Vegas a year ago and the one that returned this year. The original problem has not gone away. If anything, our conviction in solving it has increased. We have simply got closer to it and, in doing so, started to see the much bigger system around it.
There is still an enormous amount to build. But for the first time, I think we can see the boundaries.
Be the first to comment