Since the emergence of aggregators in the insurance sector, purchasing an insurance policy is relatively straightforward; however, the true challenge, as stated by Michael Cook, a partner and claims advisory lead at consultancy giant PwC, lies in what happens next. PwC’s advisory practice manages industrial liability claims, covering lower-complexity cases like industrial deafness and more […]
Since the emergence of aggregators in the insurance sector, purchasing an insurance policy is relatively straightforward; however, the true challenge, as stated by Michael Cook, a partner and claims advisory lead at consultancy giant PwC, lies in what happens next.
PwC’s advisory practice manages industrial liability claims, covering lower-complexity cases like industrial deafness and more complex cases such as cancers caused by asbestos exposure.
The consultancy has to prioritise its customers, their claimants, and, occasionally, the claimants’ families. Cook emphasises that in the context of claims transformation within the insurance sector, these crucial moments of support are what truly count.
“I say this as someone who has worked in the industry for 10 years in the claims department,” he says.
“It’s one thing making an easy-to-buy policy, but whether you are an individual or a company, when something goes wrong, you want to be dealt with helpfully and efficiently. And that post-buying piece is probably where the greater amount of improvement is needed,” he adds.
The challenge
Insurance claims processing is a critical yet complex and time-consuming function. As Cook highlights, claimants in the construction industry may have worked for multiple employers, making it more difficult to determine liability.
Assisting the families of deceased claimants and gathering the necessary information can also be challenging.
Fortunately for insurance firms, more information is now digitised, including records of issued policies and detailed corporate histories — including acquisitions and periods of government ownership.
“Given that these claims often span 30-40 years from initial exposure to diagnosis, having a robust system to trace coverage history is crucial. Ultimately, our goal is to ensure those affected receive the compensation they deserve,” says Cook.
PwC’s claims advisory practice has typically concentrated on advising insurance clients instead of processing claims.
However, in 2020, a client approached PwC with a challenge: their existing claims service provider couldn’t continue operations. Rather than simply taking over legacy systems in a conventional outsourcing approach, PwC saw an opportunity to transform the claims process completely.
“We didn’t want to do ‘your mess for less’ outsourcing,” Cook explains. “Instead, we wanted to build a scalable, efficient, and data-driven claims processing platform.”
PwC selected Appian’s low-code, AI-powered platform to achieve efficiency gains while maintaining strong human oversight and ensuring rigorous data privacy and governance measures remained in place.
Procurement choice
According to Cook, PwC chose Appian due to its speed-to-value, process orchestration capabilities, and ability to scale. “We built our first system from zero to release in five months,” Cook says. “That’s incredibly fast for an enterprise solution.”
He adds that Appian’s strength lies in its ability to reengineer processes and enable seamless data orchestration.
Cook says traditional claims systems act as records management tools, whereas Appian provides a dynamic and flexible platform that allows insurers to optimise workflows and extract deeper insights from data.
Michael Cook, partner and claims advisory lead, PwC
“For instance, we’ve probably got 3,000 active claims but a hundred times more closed claims. So, how do you unlock the power of this valuable historical data, which is a huge asset? You need a platform that can look at all those claims and provide insight and analysis on things like the length of the task, etc. Being able to create this kind of tooling is important.”
“If you look at those two areas, data and AI, the technology is moving so fast that you need a platform that will keep up with that technology and stick with that evolution.
“We don’t want to have to construct a whole different data warehouse and develop a whole different set of tooling to extract value from the data when we can extract it from the Appian platform.”
Efficiency gains and continuous improvement
The impact of PwC’s transformation was immediate, according to Cook. “From day one, we achieved a 30% efficiency gain,” he notes. “We improved task management, automated key processes, and unlocked greater value from our data.”
Crucially, PwC didn’t stop at the initial transformation. They established two key streams for continuous improvement to ensure ongoing efficiency and effectiveness.
The first stream focused on people-driven enhancements. PwC onboarded 65 claims professionals, integrating them into the process and encouraging them to contribute ideas for improvements.
These employees identified areas for refinement through dedicated workshops, drawing on their experience with legacy systems to shape a more efficient and user-friendly workflow.
The second stream centred on technology-led automation. PwC leveraged Appian’s advanced capabilities to introduce workflow automation, dynamic reporting dashboards, and AI-driven optimisations.
Cook says these enhancements streamlined processes and delivered an additional 10% efficiency boost, further refining the claims handling system.
Integrating AI: keeping humans in the loop
PwC also integrated AI within the Appian platform for claims intake and processing. Using an AI-based ingestion tool, PwC automates document handling: It reads incoming claims, determines whether they are new or updated, and categorises them accordingly.
However, human oversight remains a critical safeguard. “There are always quirks in AI,” Cook admits.
“For example, our system initially struggled to process emojis in legal correspondence. That’s why we always have a ‘refer to human’ decision step built in.”
AI struggled with processing emojis in emails
Beyond initial intake, PwC is exploring AI to summarise vast amounts of data — such as medical reports spanning hundreds of pages. Cook notes that AI provides consistent outputs and does not suffer from human fatigue, making it an invaluable tool for claims handlers.
Data privacy and AI governance
PwC enforces strict governance protocols when handling highly sensitive personal data classified as DC3 (the highest data sensitivity level). Cook emphasises that “AI must remain a tool for human empowerment, not a replacement.”
“Every AI deployment undergoes rigorous internal and client-side governance to ensure ethical use, data security, and bias mitigation.”
He adds that PwC is working closely with clients to align AI governance policies with their risk appetites, recognising that different insurers are at different stages in their AI adoption journey.
Read more: Lack of upskilling and AI training contributes to resignation spike, PwC survey
Scaling AI in insurance claims processing
Looking ahead, PwC sees AI playing an increasingly autonomous role in claims processing, but with trust and transparency as guiding principles.
“We could introduce agentic AI today,” Cook says, “but we wouldn’t be comfortable doing that just yet. Trust needs to build over time.”
Cook envisions a phased evolution in which AI progressively assumes more low-level decision-making responsibilities while high-complexity cases remain under human supervision.
“As people become more comfortable and trust the AI’s consistency and fairness, we’ll see thresholds shift — but this will take years, not months,” he says.
