Specialty claims are ready for automation – here’s why

The specialty insurance market has long been considered too complex, bespoke and fragmented for claims automation to make an impact. But that narrative is changing fast. From the Lloyd’s market to digital TPAs, forward-thinking players are showing that automation is not just viable in specialty lines, it’s already delivering value.

At InsTech’s May 2025 event, Claims Automation Platforms in Specialty Insurance: Dispelling the Myths, sponsors RDT and Synergy Cloud joined early adopters and innovators to unpack the state of play. Discussions focused on how platforms are being implemented, where AI is genuinely working and why insurers must act now to avoid falling behind.

The event opened with a discussion on what claims automation really means. Jon Mitchell, COO at RDT, emphasised that it’s not about removing human judgement: 

“When you say automate, it’s almost too simplistic because augmenting decision making and ensuring that the human is in the loop for where the value is added is the key. It’s not black boxing decision making, which traditionally insurers don’t like anyway.” RDT’s ACE platform, already proven in motor and commercial lines, is now being deployed in the specialty market. 

Scott McKenna at careless brought the consumer lens. His team built in-app claims filing and status updates, then chose people where empathy matters. “Tech and low cost might be best for the business, but not always for the customer.” 

Synergy Cloud CTO Tom Burroughs echoed the sentiment: specialty carriers are ready for smarter workflows and better experiences, not massive tech overhauls. This mentality has been sped up by the new breed of MGAs and coverholders who expect a joined-up customer journey from purchase to claim. That means API-first intake, flexible workflow, and clean hand-offs to handlers, TPAs, underwriters and finance. Burroughs mentioned:

“They’re demanding a different approach to how the whole customer cycle works, from buying the product through to the claims process… It’s great to see the expectation and the bar raised that much higher, which is great for us as vendors.”

A recurring theme was the shift from legacy platforms to modular, modern systems. “The retail space has always had pressure for process automation and operational efficiency,” said Mitchell, “there’s been, rip out your legacy platforms and spend millions of pounds. There was the dawn of the low code and no code era… but it proved really hard to scale and manage at scale.

From Tesco’s early data experiments to image analytics in motor – personal lines shows how scale arrives in steps. Andy Stevenson at PwC and Patrick Hayward at Altus pointed to two lessons that transfer:

• Move from document handling to data ingestion so handlers see the right facts at the right moment
• Build workbenches that remove re-keying and surface portfolio insights, not just claim-by-claim tasks

Chris Payne at EY noted the pattern in London. Underwriting has gone first with digital workbenches. Claims is next, with the same focus on collation, control and outcome.

This isn’t about rip-and-replace. Speakers called for agile, iterative rollouts – modernising in steps, not leaps. Structured data and open platforms are making this approach practical and affordable.

One of the strongest messages from the panel with Somerset Bridge and HiBRID was that AI isn’t experimental – it’s operational. AI is already helping classify documents, orchestrate workflows and handle unstructured data at scale. “Some of the things we’re doing now weren’t even possible three months ago,” said Mitchell.

Digital TPAs like HiBRID are combining human expertise with AI to provide scalable, high-touch service models. The result: faster resolution, reduced overheads and more transparency for clients. 

From a carrier perspective, Jonathan Brown of Somerset Bridge emphasised the operational gains:

“Our initial use case was efficiency… We’ve automated all of our customer communications, which is great from a customer outcome perspective as well. And we can standardise and just have a really rigorous process across the board.”

Simon White of Apollo was clear: “The train is moving. Get on now, then adapt as you go.” Apollo has processed over 23,000 claims with an average touch time of under one day.

Aidan O’Neill at DOCOsoft reminded the room that adoption is real. The firm now works with over half the managing agents at Lloyd’s, and carriers are running straight-through processing on defined workflows.

From the market view, Zoe Woods, Claims Improvement Manager at Lloyd’s highlighted that while not every managing agent is moving at the same speed, momentum is building. The key is to test, iterate and partner. Lloyd’s Lab, elective services, and new governance models are creating space for innovation within a regulated framework.

One of the most practical insights came from Apollo’s approach: get claim handlers in a room and ask them what tasks they do on every file. Then automate those. It started with simple checks – policy period, premium status, named insured – and led to meaningful time savings across thousands of claims.

This ‘low-hanging fruit’ approach means that firms can see immediate benefits without massive budgets or risky transformations.

A dedicated session for early-stage companies proved just how broad claims innovation has become. 

OpenOrigins is tackling AI-generated fraud with blockchain-backed image verification. Hannah Gormley, Chief of Staff described the process: “We are able to, in a completely secure way without taking your data out of your system, compare unique IDs for every image we’ve ever seen… Every time you see a claim, we’re able to tell you, have we seen this before? Have we seen part of this before?”

Quarto AI is building back-office automation tools for brokers and exploring AI-native consumer models in the U.S. Looking ahead, Charles Calzia, Founder, said: “We’re now thinking of becoming an AI native consumer insurance broker… What if we can not only build this price comparison site, but also offer the full value of a broker in the US?”

Meanwhile, nettle is rethinking risk engineering by automating inspections and reports, as Jack Co-founder & CEO explained: “Risk engineering, typically you have someone do a survey or an inspection of a property like a factory. That process, as you can imagine, is very manual today. 60% of their time is spent writing reports and 40% of risk engineers will retire in the next five years. So it’s already become a bit of a bottleneck for underwriting and stopping new business growth. Nettle basically automates all of that manual work.”

Lastly, Morph Cloud is building adaptable data tools for underserved insurance niches. Alex Dabell, Founder divulged that “The conclusion I came to at Price Forbes was that the industry was really interesting, and full of opportunity, but it was never gonna digitise with existing software development practices… Morph Cloud makes it radically easier to digitise complex insurance data and processes. The aim is to make it routine in this industry. So easy, it’s routine.”

Each company shared a common goal: reduce friction, move faster and deliver value without the overhead.

A final theme that cut across every discussion: claims isn’t just a cost centre. It’s a core part of customer experience, underwriting insight and competitive differentiation.

Whether it’s fraud detection, document processing, or dashboarding, automation is enhancing – not replacing – the people in claims. And it’s helping insurers deliver on the most important promise they make: to pay promptly and fairly.

Justin Murphy, HiBRID, stressed: “I believe that claims will move away from just being seen as a cost to control, but will be a strategic asset that can work alongside the underwriting team. You talk about algorithmic underwriting, dynamic pricing. Well, dynamic pricing doesn’t work unless you have clear claims, accurate data. So I think the two can work alongside.”

Andy Stevenson, PwC, highlighted the portfolio view:

“What people have been looking at now in specialty markets and very recently has been really this kind of claim by claim approach… I think there’s a different perspective now, which is to take a step back and think about what is the claims function really there to do? It’s the portfolio, not the individual claim.”

Chris Payne, EY, reinforced that investment cases must join up:

“You can’t have something that’s really sharp and digital at the front, and the underwriters are working so much more quickly, and then find that the guys in claims are still working in the “caves”, as it were.”

And Patrick Hayward, Altus Consulting, added the importance of framing benefits clearly:

“What is it that you are trying to achieve in terms of ‘what’s the problem that you’re trying to fix’ or ‘the opportunity you’re trying to grasp’?… If you could bring all that to life in a single view, I think that helps with that conversation.”

The myth that specialty claims can’t be automated is officially debunked. What matters now is action.

Insurers, MGAs, TPAs and brokers don’t need to wait for perfect systems. They need to start with what’s real: existing data, repeatable processes and partners who know the space. As AI and platforms mature, those who act early will set the new standard.

Claims isn’t catching up to underwriting – it’s leading the charge.