Zywave prepares to launch cyber hub at the end of Q1: Cohen

People and resources are “scarce and expensive” and so an artificial intelligence (AI) strategy for re/insurers begins with “deliberate selections”, according to Jeff Cohen, senior vice-president of brokerage software provider Zywave.

In an interview with Insurance Day, Cohen says Zywave’s insurer clients are “dabbling right now, by biting off small pieces of impact” for underwriting, reserving, claims handling and loss control.

“An obvious requirement is sound data, and those insurers who have organised their internal data have an advantage here. When combined with external data sources, the potential to learn quickly, greatly transcends what a single individual could learn over the course of years,” Cohen says.

For example, when AI is applied to the text descriptions of Zywave’s one million loss data records, a specific underwriter could immediately decipher factors that  trigger large loss scenarios.

The mid- and long-term scenarios of an AI strategy will likely involve an effort to figure out how to expand into other areas and across lines of business, as well as how organisations shift their data collection efforts.

“There are so many opportunities to improve our industry’s workflows, products and services through AI – and we’re just scratching the surface today”

^{Jeff Cohen
Zywave}

“Inevitably, this topic will force insurers to get a better handle on how claims and submissions are stored and utilised,” Cohen says.

The world really started to see the power and potential of AI in 2023. While ordinary consumers queried the likes of ChatGPT to respond by leveraging the knowledge base of the internet, insurers began to consider how their proprietary data and loss experience might yield greater insight with AI.

Cohen explains: “There are so many opportunities to improve our industry’s workflows, products and services through AI – and we’re just scratching the surface today.”

Zywave has been creating content for more than 20 years aimed at helping insurers and brokers explain risks and exposures to policyholders. It has built its own AI engine to mine this material for far greater insights. Zywave is also working with a subset of its clients who license its loss data.

“Given the nature of these one million records – low frequency and high severity – our clients are tearing this material apart,” Cohen says, “in an effort to uncover the trigger factors, which in turn inform loss control and also to determine early stage emerging risks to avoid.”

Clients' main concerns

Zywave’s clients are concerned about the impact of climate change and extreme weather, social inflation, and litigation in fines arising from factors such as General Data Protection Regulation and meta-pixel data.

The company’s brokers and clients are also concerned about cyber risk, and cyber loss events are an important part of its event data collection efforts. Privacy themes are an important topic of discussion with insurers and brokers, and they are very aware of their own privacy protection responsibilities.

“AI is a fast-changing emerging technology which will create an evolving range of risks for the insurance industry, and the regulatory response is very much in its early stages,” Cohen says.

“We are closely following loss events related to AI, such as the New York Times’ lawsuit against OpenAI and Microsoft over copyright infringement,” he adds.

The insurance industry is well-versed in technology trends that are meant for good, but might get re-deployed for sinister purposes.

Cohen says: “Look no further than the 75,000 malicious cyber breaches in our loss database. AI unfortunately will not be an exception. Criminals are already using AI not just to create malware, but also to distribute it. Previously, the only criminals who could create malicious programmes were those with outstanding IT skills, but now they are able to develop these devices in seconds using AI.”

Cohen rattles off a number of examples.

Criminals can deploy AI for sinister purposes James Thew/Alamy Stock Photo

AI can be used to generate deepfake YouTube videos that are highly convincing, he says, and it can increase the believability of this material by generating fictitious likes and comments.

Passwords are the first line of security for most computer users, Cohen continues, but criminals are now using AI to crack these really quickly. He says some AI tools can crack 51% of passwords in less than a minute, and 71% in less than a day.

Another negative use of AI is to create emails, texts and phone calls that are designed to fool people into revealing valuable personal information. The software can correct poor grammar, which makes the phishing messages even more convincing, Cohen adds.

Criminals are also using AI to detect vulnerabilities in computer systems. Once they have stolen data, they deploy the AI to rapidly analyse the information to exploit it for maximum effect.

Cohen stresses that one way of mitigating these risks is for businesses to regularly test the cyber health of their computer systems. Basic measures include having passwords of at least 12 characters, backing up key data on an external hard drive or the cloud, continuously informing staff about cyber threats, and having a thorough cyber response plan.

Strategies for London

The US market is a “great source of premium”, Cohen says, and describes the distribution strategies London needs to penetrate that market.

“Expanding distribution beyond an organisation’s natural reach, and without adding great expense, remains an elusive goal for most capacity providers,” he says.

Whether that distribution resides in the US or in other jurisdictions “becomes almost inconsequential”. However, since much of the premium dollars in the London market originate in the US, Zywave has focused its efforts there. As the excess and surplus and specialty markets continue to increase market share, distributors are attracting ever more submissions, Cohen says.

He continues: “The ability for technology to efficiently get those submissions into the capacity providers, triage them, and present the underwriters with the application details and a quote so that they can quickly evaluate the submission with pricing, takes some of the friction out of the process, and allows the underwriter to get a feel for whether they are willing to accept the risk at the price that’s been generated by automation.”

Cyber hub

Last September, Zywave announced its strategic partnership with Cowbell, a provider of cyber insurance tailored for small- and medium-sized enterprises. Their collaboration is facilitated by Herald, a digital infrastructure provider connecting software developers to commercial insurance carriers through a unified application programming interface (API).

Their partnership represents the first standalone cyber insurance offering on Zywave’s hub, which accelerates access to cyber insurance products for managing general agents (MGAs), wholesale brokers and retailers.

“We are currently finalising the development of our cyber hub, with an anticipated launch date at the end of Q1, and Cowbell is a key partner for this effort,” Cohen says.

Zywave has also integrated with At-Bay, Coalition and Axis Capital, among others, who – together with Cowbell – will be first to go live on its cyber hub.

Zywave will continue to onboard capacity providers to expand its offering from a coverage standpoint and reach more producers.

“We’re powering digital distribution and every stage of the insurance lifecycle by fostering communication, collaboration and connections between key stakeholders in the insurance distribution chain, from re/insurers to brokers and all the way down to insureds and policyholders"

^{Jeff Cohen
Zywave}

“Ultimately, our plan for the MGA hub and the cyber hub is to add capacity providers and producers – retail brokers and wholesalers – with a goal to adding additional lines of business as demand requires.

“We anticipate a workflow whereby a big broker enters information into a proprietary system, pushes that detail to our hub, obtains quotes, and binds the risk. For capacity providers, you’re increasing your marketplace reach by integrating to one environment rather than connecting unique point-to-point APIs to each individual producer environments,” Cohen says.

Zywave’s primary objective this year is to “continue to revolutionise” the insurance market through digital transformation.

“We’re powering digital distribution and every stage of the insurance lifecycle by fostering communication, collaboration and connections between key stakeholders in the insurance distribution chain, from re/insurers to brokers and all the way down to insureds and policyholders,” Cohen says.

He concludes: “To further this mission in 2024, Zywave will continue to architect products and partnerships that make the placement of insurance more predictable, efficient, and streamlined for all parties.”

Insurance Day will be publishing a special report on market modernisation next month