In this AI progress series, we have discussed at length the use of AI throughout the insurance value chain for time saving benefits, but for brokers, the greater value of the technology may lie instead in data insights to better match risk to capital.
This instalment of Insurance Insider’s AI Progress series will discuss how brokers are using AI, in particular agentic AI, to enhance their offerings and support their broking teams.
It will also discuss how the insurance value chain, including brokers, is unlikely to change or see key players in the chain disappear as a result of AI, which has been a key talking point and item for concern.
The first instalment of the series commented on AI uptake within the insurance sector, the second tackled AI in claims and the third looked at the way AI will impact the underwriting department.
Most recently, the series looked at how AI will impact talent for the insurance market.
AI to minimise data toil
As this series has previously touched on, AI can be revolutionary in its ability to take on and minimise the time it takes to perform administrative or repetitive tasks, including data collection.
For brokers, this includes the potential for AI to assist in submission processing, as it allows the client to provide risk information in any format that they would prefer, be it an application, a phone call with an agentic agent, an email or a PDF submission.
Another senior broking source told Insurance Insider that AI, particularly agentic AI, is “removing toil”.
They explained that a broker’s day-to-day workflow, including tasks such as performing know your customer and sanction checks, will become “hugely automated” with AI.
An AI agent could extract relevant data from these formats and in a very short amount of time present that data to the broker, alongside relevant information and key considerations that the broker should be aware of.
Without AI, brokers deal with a volume of data in a way that can be very “labour intensive” according to Bill Pieroni, global head of insurance strategy and AI at DXC.
These data processing wins provide an efficiency gain. However, Pieroni said the higher value-add for brokers comes from analysing the captured data to understand and help place the risk.
Similarly, Anthony Siggers, digital leader at Marsh, told Insurance Insider that while the full potential of AI continues to unfold, “we are already witnessing significant opportunities through practical use cases that enhance both colleague productivity and client service”.
According to McKinsey & Company, gen AI use cases in commercial P&C sales and distribution functions include hyper-personalised customer outreach, request for proposal streamlining and automated pre-filled forms, among others.
AI for data analysis
The next step from efficient collection of key data points is pulling out granular detail of a risk profile, which can be invaluable to brokers through dynamic insight into ever-changing risk.
Unlike generative AI, which only responds to prompts, Agentic AI can work autonomously to understand the risk and how it is changing and the full continuum of options a client has, meaning that the broker has access to consistently up to date information which will allow them to offer the best service as a result.
The recently launched AI platform Aon Broker Copilot is one such example of a broker leveraging the data capture and analysis capabilities of AI.
The Copilot is an integrated placement technology, data and analytics platform, Aon’s head of placement technology and trading analytics, Clyde Bernstein, told Insurance Insider.
Joe Peiser, Aon’s commercial risk CEO, explained that the platform’s most significant use is to capture all of Aon’s trading data around the world.
“Every product, every industry, every geography. Not just what we bind, but all the quotes, and every iteration of quotes, so that we have real-time visibility into how the market is pricing risk.”
Bernstein added that the primary objective of the platform “is not efficiency. The goal is to make us better”.
Peiser said that brokers would be able to input client profiles and exposures and receive insights on what the market is for that risk, who is broking those kinds of risks already, and what kind of pricing can usually be seen on those types of risks.
Furthermore, sources made the case for using AI for wordings analysis, comparison and understanding.
Bernstein said that a “really good use of AI” is picking up on potential gaps or ambiguities in the policy in real time.
Another source agreed with this use case, adding that a broker could upload any form of wording to generative AI and ask it to look for certain clauses, or compare it to wordings that are market leading.
Marsh’s Siggers said that the firm’s internal tools now handle around one million inquiries weekly, driving efficiency and automation internally.
“Externally, we are expanding our AI capabilities with client-facing solutions such as Sentrisk, our AI-enabled platform for supply chain risk assessment. We use our data as a global risk leader to improve how we deliver insights with AI and analytics, helping our colleagues and clients handle a more complex world with confidence.”
Similarly, a spokesperson for Gallagher told this publication that the broker is “using AI to more efficiently and productively support the service we provide to our clients and partners”.
Digitisation or AI
During discussions on how AI can be used to enhance a broker’s day-to-day workflow, the question of what gains came from digitisation vs AI often arose.
Sources worried that the term AI is being applied to “everything,” including any digitisation software, that may or may not use artificial intelligence.
“We are blending many, many things into this overarching statement [of ‘artificial intelligence’]” one senior broking source worried.
This is a concern because the conflation of any automated or digitisation technology with AI only makes the technology more difficult to understand.
Peiser told Insurance Insider that the term “AI” is often misunderstood.
He clarified that while digitisation creates new opportunities for data collection, AI will facilitate this opportunity at scale. In Aon’s case, Peiser said that has led to transforming the firm’s own proprietary trading data into a powerful tool for market insight on a very granular level.
In turn, that will “help our brokers navigate the global insurance marketplace more effectively”, he added.
The brokers of the future
One concern about the use of AI is whether it will render any parts of the insurance value chain redundant. Brokers are often perceived as one of the nodes on the value chain that is more at risk of being replaced by AI.
In some cases, brokers are starting to use AI technology to assist with facilitisation and placement. While this automates part of their placements, it also brings in additional revenues and commissions.
Many facilities do not require AI and are set up with more simple rules-based systems, but those that have involved more algorithms and digitisation include McGill’s Auton and Gallagher’s Evolve, which uses InsurX technology.
However, as discussed in the talent instalment of the AI Progress series, AI will not replace the broker altogether. There will always be a human in the loop.
Those who can harness the data insights and in turn offer the best coverage to their clients will be the ones using the technology to come out on top.
The philosophy seems to be that AI will make the powerful intermediaries more powerful. They should have the best ability to exploit their data assets and pay for the cost of large-scale AI implementations, and they have more proprietary data as opposed to third-party data, a senior executive said.
“But just as radio has not been eliminated by television, and television has not been eliminated by streaming, you're not going to see the whole elimination of any stage of the value chain.
Bernstein said that individual brokers are the ones that have the intuition, experience and relationships with both markets and clients, “but if we can empower them with the knowledge of what the rest of the firm knows about market conditions, client buying habits, etc, they will become even more equipped to do better deals.
Ultimately, he explained that the Aon Broker Copilot is allowing the broker to keep the emphasis on the art of broking, while “bringing the science alongside it”.
Peiser concluded: “Our goal is not to digitise broking. Our goal is to digitise everything around broking.”
Using the technology, Aon said it has achieved a 54%-win rate when it had led with its client-facing analytic tools in its requests for proposals, the highest win rate Aon has ever had.
Though AI implementation in broking is still in its early days and often confused with the shift to digital, it seems clear that the technology is not being implemented in a way that will take over the role of a broker, but rather will give the brokers the tools to take their role to the next level.
Got a tip for us? Reach out confidentially here.