Commercial property has changed shape over the past few years, but for brokers placing that risk in 2026, the underwriting conversation increasingly turns on how quickly, and how confidently, a decision can be reached. Andy Jones, head of underwriting at Commercial Express, a Dudley-based managing general agent (MGA), argues that the answer depends less on how much data an underwriter can access than on the quality of the information in front of them and the judgement applied to it.
The fundamental objective of underwriting has probably not changed. It is about understanding the risk, confirming it sits within appetite and making sure the premium reflects the exposure and controls in place. What has changed is the volume of information now available on everything from construction techniques and occupancy patterns to supply chain disruption.
More data, however, does not automatically mean more certainty. Broker submissions are not always presented in a clear, well-structured way, Jones said, and a good underwriter's job is to "distinguish between what is good and bad information" rather than simply processing everything that arrives.
The trade-off between a fast decision and a disciplined one can appear unavoidable. Jones disagrees. "The fastest underwriting responses," he said, "are those that operate within a really clearly defined framework," with clear appetite, authority levels and referral triggers that let underwriters act quickly without unnecessary escalation.
Judgement becomes necessary once a risk sits outside those parameters and the underwriter has to decide whether more information would genuinely change the outcome. That distinction matters because, in Jones's experience, cases can be slowed down by requests for information that ultimately have no bearing on price, acceptance or terms.
"Brokers want a quick response," he said, "but they want clarity of decision as well," including a clear explanation when the answer is no. That framework-first approach also helps explain why, in the SME segment, speed alone no longer differentiates one underwriter or MGA from another – it has increasingly become the price of entry. But a framework only tells an underwriter when judgement is required; it says nothing about how well informed that judgement will be.
External data sources, including geospatial flood modelling, rebuilding cost assessments and property characteristics, can genuinely help underwriters assess and price risk.
"They inform the underwriting decision," he said. "They don't take them on behalf of the underwriter." Where a risk falls outside the normal underwriting process, it is the underwriter's own knowledge and understanding that has to take over.
That makes the quality of the information particularly important on risks that cannot be decided neatly by a model. For Jones, the risks that stand out are not necessarily those with the lowest exposure, but those the underwriter understands best.
"It's not just around risk with the lowest exposure," he said. "It's around those that are best understood."
Maintenance regimes, prompt repair of defects and effective fire and security controls, he said, rarely show up fully in external datasets. That is where a transparent and consistent broker presentation can materially change an underwriter's understanding of the risk. Information that shifts significantly partway through the process, by contrast, can undermine confidence in writing it at all, however low the exposure might otherwise appear.
That reliance on the broker is significant. Brokers arrange around 94% of all UK commercial insurance business, according to the British Insurance Brokers' Association (BIBA), underlining that position at the centre of the information flow between clients and underwriters.
As underwriting becomes more data-rich, Jones does not expect the broker's role to diminish. He argues that underwriters will increasingly need to interpret data alongside their own technical judgement, particularly on more complex risks - a shift the CII's Underwriting New Generation Group flagged back in 2021, when it argued that future underwriters would need to combine core technical skills with statistical and analytical fluency rather than be replaced by them.
More underwriting data does not reduce the value of a strong submission; it arguably increases it. Models can tell an underwriter a great deal about a property, but they cannot always explain how well a business manages the risk behind those numbers. The broker who can fill that gap gives the underwriter something data alone cannot: the confidence to make a decision.
