It is not the most obvious career trajectory: ancient world studies at the University of Manchester, followed by a career building the digital infrastructure of a fleet insurance managing general agent. But for Paul Waring (pictured), IT director at Blagrove Underwriting Agency (BUA), the path makes perfect sense in retrospect.
Waring had grown up as something of a computing anomaly. His family owned a home computer at a time when few schools could say the same of their pupils. When it came to choosing an undergraduate degree, computing was the practical choice. “There was definitely going to be jobs in computing,” he said. “I was not going to go down a path that would not result in any work.”
History, however, remained a persistent passion. As his undergraduate degree drew to a close, Waring seized what he recognised as a now or never opportunity to pursue a master’s in ancient world studies. He finished the course with no regrets, though he freely acknowledges the limited commercial demand for expertise in the classical world.
His entry into insurance was, as it so often is in this industry, a matter of connection rather than calculation. A contact from his university days knew people at a company building an IT platform for commercial motor insurance, specifically fleet policies. Waring did some contract work, proved himself, and was offered a full-time role. Within 18 months he was a director. “It just all worked quite well,” he said.
After a brief return working at the University of Manchester on the software side, and 10 years running his own IT consultancy, Waring was asked last November to join Blagrove Underwriting as a director. It is a role that keeps him close to the coalface of commercial fleet technology, while leaving him enough time to continue the consultancy on the side.
Blagrove Underwriting is a motor fleet MGA, underwriting policies on behalf of its capacity provider, currently SCOR, and managing claims entirely in-house. It operates in the SME fleet sector, covering three to 15 vehicles, as well as a particular focus on non-standard risks: haulage, couriers, and fleets with more complex claims histories that larger carriers are reluctant to touch.
What makes the operation distinctive from a technology perspective is its reliance on a single, entirely in-house-developed platform. Accounting runs on Sage; email is handled via Microsoft 365. Everything else flows through one system that Waring and his team of three software engineers control entirely: underwriting, claims, invoicing, and data exchange.
“Everything in the business runs on this platform,” he said. “It is not some side product that we use to help us. All the insurance business runs through it.”
The benefits of that model are significant. There is no vendor dependency, no waiting in a supplier’s development queue, no negotiating over features with a third party whose priorities lie elsewhere. If the business needs something built, it gets built. The platform today is, in Waring’s words, “completely different” from the version that existed two decades ago, even if certain structural elements remain. The interface, he is candid, has never been redesigned for aesthetic appeal. It does not need to be. The audience is underwriters and insurance brokers, not consumers. Function, not form, is the governing principle.
Ask Waring about the tangible outcomes of his team’s work, and the answers come quickly. Three areas stand out.
The claims management system was built from scratch when Blagrove brought its claims handling in-house after a brief period of outsourcing. Every element had to be constructed: payment authorisation workflows, correspondence records, reserve movement tracking, the approval hierarchies required as the team grew from three handlers to more than 20. “We had to adapt to allow for that,” Waring said, “because when there are three people, there is no need for specialisation. Now there are separate teams for specific tasks.”
The bordereau process, once the responsibility of a dedicated member of staff producing reports manually, now runs entirely automatically overnight. No human intervention is required unless something goes wrong, which Waring describes as “extremely rare.”
Perhaps the most impactful change has been the overhaul of the accounts system. Previously, closing a piece of business on the platform meant someone then had to manually raise an invoice, verify the figures and enter the correct split between tax and commission. The process has been fully automated. “Now if you close something on the system, the invoice just appears,” Waring said. “It knows how much to charge, and it knows how much is tax and commission.”
The effect on headcount has been telling. The accounts function, which handled the company’s full book of business with one person at the outset, has grown to the equivalent of one and a half. The claims team, by contrast, has grown to over 20. Automation absorbed the volume in one area; skilled human judgement absorbed it in the other.
Underpinning Blagrove’s approach to automation is an unusually rigorous philosophy around data quality. Waring is unapologetic about it.
When integrating with a third party, whether receiving data or sending it, the process begins with a detailed specification document that sets out precisely what data is required and in precisely what format. Any submission that deviates from that specification is rejected outright. “If somebody puts an email in the wrong field, it will be rejected,” Waring said. “We do not half-accept it or be fuzzy about it.”
The rationale is straightforward. Clean, validated data is what makes full automation possible. The moment you introduce tolerance for inconsistency; you introduce a point at which a human must intervene and make a judgement call. That defeats the purpose. “As soon as you start introducing flexibility or possibilities, you cannot really automate it,” he notes.
The Motor Insurance Database submission process, now fully automated, is a case in point. Blagrove validates everything before it goes out. Rejections from the database are, in Waring’s experience, extremely rare. “We have already checked the data before we send it,” he said, “whereas some people just send stuff and hope it gets through somehow.”
It is not a glamorous philosophy, but it is a consequential one. The discipline required to maintain it, he acknowledges, can test the patience of external partners not accustomed to working at that level of precision. The payoff, however, is a system that simply runs.
Integrations with third parties remain, by some distance, the most challenging area of the technology operation. “The first 90 per cent is easy, and then the last 10 per cent takes you 80-90 per cent of the time,” Waring said. “That comes up time and time again.” Larger counterparties bring their own timetables and their own constraints, and no amount of internal agility can accelerate a partner’s development cycle on the other end of a data pipeline.
Waring is aware that measured positions on AI are not always welcomed in a moment of significant industry enthusiasm, but he does not appear unduly troubled by that.
Blagrove does not currently use artificial intelligence on its platform. The reasoning is layered. First, the nature of the business. Underwriting and claims at the non-standard commercial end of the market remain, in Waring’s view, inherently relational. “Brokers want to speak to an underwriter,” he said. “They have a difficult to place risk and they want flexibility and a conversation.” Claims, similarly, are not amenable to automation beyond a certain point. Policyholders managing large fleets “value speaking to a real person and knowing that a person is reading the email they have sent and that a person will respond.”
Second, and perhaps more fundamentally, the regulatory framework is unsettled. “It is not clear what the regulatory landscape really is for AI, particularly as we are in a heavily regulated industry,” Waring said. “If you use AI and it goes wrong, who is to blame? What does the regulator say?” He draws a deliberate historical parallel with the excitement around e-commerce in the 1990s, when predictions of universal online retail proved considerably more optimistic than reality. “It is safer to wait and see how this rolls out.”
The conclusion is not that AI has no future in insurance, but that timing matters. “At the moment, the risks outweigh the benefits for us,” Waring said. “There might be other insurers for whom it makes absolute sense.” If AI does find its way into Blagrove’s operations, he suggests it is more likely to begin in data analysis than in any customer or broker-facing function. The prospect of chatbot-generated quotes is, he said plainly, “not what people want, and not what people are paying us for.”
Waring is more willing to speculate on the broader trajectory of the sector than on his own firm’s AI timeline.
He anticipates a gradual push towards commoditisation in smaller commercial fleet: the three to ten vehicle, standard-use policies that carry no particular complexity. The price comparison model that has transformed personal lines could, over time, extend into this part of the commercial market. “It has not happened yet,” he said, “but I think it will.” More unusual risks, haulage in particular, will remain the domain of specialist human underwriting for longer.
Cyber risk and technology resilience, he argues, deserve significantly more regulatory attention than they currently receive. The dependency of financial services infrastructure on technology systems is, in his view, one of the more underappreciated systemic risks in the sector. “If the Motor Insurance Database goes down, the police cannot check whether you are insured when they stop you at the roadside,” he notes. He would like to see the regulator take a more proactive stance.
On the question of what transformative technology might emerge over the next five years, Waring is genuinely agnostic. Prediction in this field, he suggests, is an exercise in humility. Virtual reality has been the imminent future of computing for more than two decades without delivering on its original promise. Whatever does reshape the industry may be something no one has yet seriously considered.
When he steps away from the platform and the policy bordereaux, Waring’s interests reveal the intellectual breadth that his unusual academic background might lead one to expect.
He is a keen hiker and a member of the Ramblers, heading out regularly both with the group and independently. He is, by his own admission, a committed bookworm with a library that considerably exceeds his reading capacity, favouring science fiction and history. He gives history talks at the University of the Third Age in Warrington, where he has spoken on subjects ranging from the German occupation of the Channel Islands to the military campaigns of ancient Rome.
For a man who has spent his career building systems designed to run flawlessly in the background, the appeal of standing in a room and talking about Caesar or Guernsey seems, on reflection, entirely consistent. Both require precision, preparation, and an audience that trusts the work that has gone into them. In that sense, at least, ancient world studies and insurance technology are not so different after all.
