Funding for insurtech startups worldwide has gone through rapid changes, and another shift is underway, as insurance companies rush to take advantage of artificial intelligence. What would a unicorn in this new era look like?
In the zero-interest rate and Covid-accelerant boom years of 2019-2021, the industry’s unicorn cohort (those privately valued at $1 billion or more) included a lot of challenger companies.
In Asia, this would include general insurance tech players such as Acko General Insurance, CXA Group, Digit Insurance and ZhongAn Insurance. These firms targeted narrow segments such as cars and travel where they could use technology to create new businesses or win share in segments that incumbents didn’t prioritize.
It also included comparison sites (Policy Bazaar, PasarPolis, CompareAsiaGroup) that have tried to become marketplaces. Globally, embedded insurance, climate risk and cyber insurance became relevant. (What didn’t get funded by the VC world: life insurance.)
New funding paradigm
The changing macro paradigm starting in early 2022 led to an emphasis on startups helping incumbents manage their digital transformations, or on startups with an edge in health and wellness. These tended to be businesses that were larger, later stage, and able to demonstrate profitability.
This has been a painful period for most insurtech startups. Many leaders, even well regarded ones, have lost their unicorn status as valuations have been slashed: Waterdrop, for example, is no longer considered a unicorn despite its expertise in China’s healthcare marketplaces.
There are some exceptions: Singapore’s Bolttech, a global embedded-insurance/marketplace platform, has seen its valuation increase from $1 billion in 2021 to $1.6 billion as of May 2023. India’s Acko has also improved its valuation, to $1.5 billion.
Overall, the global insurtech industry has seen venture funding collapse (although not as badly as in other tech sectors, such as edtech or foodtech, or indeed fintech generally). For insurtech, the pullback has been especially severe in later stage investments, from Series B through to growth equity, reflecting pessimism over the ability of unprofitable companies to exit. Series A and seed funding has been steady, although always modest.
Most problematic was the drop in valuations, which has been steep for insurtechs: as of June 2023, the average insurtech revenue multiple had fallen below that of listed insurers, according to a study by Dealroom.co.
Funding has fallen off for challenger or full-stack digital insurers, and switched to digital brokerage or supporting agent networks, or back-end functions such as claims handling and payments. Even in successful niches, however, fintechs are likely now to consolidate.
Generative AI: gamechanger
Coming into 2024, artificial intelligence, including generative AI, is widely accepted among insurance tech executives as a gamechanger.
It is not the only theme that will drive insurtech: back-office automation and embedded insurance will remain important, the industry remains committed to building in the health space, and cyber protection and cybersecurity are evergreen needs.
But insurance companies get AI’s uses in a way unlike how they have embraced other forms of digitalization.
Earlier forms of narrow AI, such as natural-language processing and optical-character recognition, are being applied to areas such as claims processing, risk monitoring, and marketing.
According to OpenAI, the company behind ChatGPT, generative AI is going to have a much bigger impact. OpenAI has said it believes insurance and banking are the sectors with the highest potential for genAI-led automation, more than energy, capital markets, software, retail, media, automobiles, health or industrials.
Insurers leapfrog
This is a huge change, because until now, insurance was considered a digital laggard, not just behind banks but behind other industries. Comprehensive digitalization only began with the onset of Covid, when agents couldn’t conduct face-to-face meetings, thus imperiling revenues.
Insurance companies, like banks, have been generally poor at working with fintechs. Their legacy systems are mainframe-based and little strategic investment had gone into transformation. That enabled the rise of both challengers and B2B fintech partners, but it’s only now that leaders in the insurance industry are moving to cloud and automating most functions.
The second difference with genAI is that it allows insurers to begin by focusing on user experience, agents, and distribution, rather than just dabbling with back-office processes, claims, or risk management.
According to industry executives, the use cases under study are boundless.
Australian general insurer QBE is testing genAI for customer experiences, to make their agents and customer services teams more responsive. FWD Group is looking to use it as an internal productivity tool, to help employees query internal procedures and paperwork, as well as for marketing. FTLife wants to use it to not just support distribution but predict policyholder or agency needs.
And, as DigFin reported in our exclusive interview with AIA’s group digital and analytics head, Asia’s biggest insurer is testing genAI to hire and groom agents to become top sellers.
GenAI insurtech?
So where is the next generation of insurtech startups that are going to lead the charge? Perhaps the better question is whether there will be one.
Many insurtechs interviewed by DigFin say they are incorporating genAI into their offerings, but they aren’t genAI natives, and they aren’t on the cusp of becoming unicorns. There remains a need for the things these insurtechs provide; AI isn’t putting them out of business. But it’s unclear whether genAI will create the insurtech unicorns of the future.
There are three areas where generative AI is being applied first. It will be used in online search by customers looking for information, it will help agents or bancassurance salespeople to understand individual customer needs, and it will help insurers and their distribution arms to engage with customers.
Do insurance companies need insurtech partners to achieve these goals?
The insurtech world has already figured out that incumbents can’t be disrupted by technology, or that insurance can be sold like a Software-as-a-Service agreement. Successful insurtechs have learned to navigate the incumbent sales cycle, while insurers have learned to trust tech companies providing targeted services rather than reinventing the entire value chain. But is there a genAI service that incumbents want that insurtechs can deliver at scale?
This is a vital question for venture capitalists looking to back the next unicorn. They will struggle to find genAI insurtechs. Instead they will have to look to areas that have traditionally not attracted VC funding, where there remains a huge opportunity to automate.
Possible plays
In Asian emerging markets, where labor costs are lower than in developed markets, third-party administration and other unglamorous businesses have forgone automation. With genAI, they may be ripe for change. Similarly, insurance companies in Asia may not pay for software, but software that is bundled with devices or services to enable sales can be attractive.
These are niche areas that may not support a large insurtech industry, and they require people who know the guts of the industry and can think laterally.
But genAI solutions are not capital-intensive: they are compute- and data-intensive, which makes them suited for broad uses tailored by proprietary data. It may make more sense for an insurtech to leverage an AI company’s LLMs (language-learning models) for its existing services.
Moreover, for enterprises, genAI doesn’t require heavy multi-year lifts, like data warehousing or moving to cloud. Even within artificial intelligence, the time it takes to train neural networks to read texts is a lot more hands-on and lengthy than it is to curate genAI prompts around a proprietary data set. Insurance companies that have gone through digital transformation have the data and the data scientists; they just need a few APIs, compliance and governance frameworks, and a clever use case.
This means insurance companies can use AI themselves rather than rely on a fintech. It’s the big AI companies such as OpenAI, which are tied to Big Tech providers such as Microsoft and Google, that can meet most of these needs.
There’s still a need for insurtechs to help insurers with digital transformation, and there’s still a role for insurtechs to create new marketplaces. The genAI insurtech unicorn will need to rise above these tasks and become the glue of greater ecosystems.
Healthcare is the most likely arena.
Insurers are increasingly focused on health, healthcare, and wellbeing, and are hungry for solutions to help them prevent or predict chronic diseases, combat medical inflation, change policyholder behavior, and services that can be used for more narrowly segmented customer bases.
However, the paradigm may be changing from clunky insurers looking for help with digital, to insurers using generative AI (and other digital tools) to grow into new business lines.
Health convergence
In the US, some insurers are moving into the provision of healthcare services, not just protection. In Taiwan, Cathay Life’s parent company also operates hospitals and clinics; Bupa does the same in Hong Kong.
The direction of change can also go the other way, with some US hospital networks setting up their own insurance arms. Embedded insurance will also evolve, placing health and insurance into retail marketplaces.
If there’s a place for genAI startups, it may well be in helping knit together the disparate arms (and data, especially medical records) as insurers and providers of wellbeing blur all around customers. To whatever extent a health-related experience can be made to feel like browsing on Netflix, with personalization that covers a dizzying array of touchpoints, it’s going to take massive amounts of data – and specialized AIs.
Generative AI startups are less than ten years old, and for now they are grouped around broad categories: model makers like OpenAI or Anthropic, image or text generation (Midjourney), video, tooling (eg prompt engineering, data management), and code generation.
But a few sector-focused areas are emerging, notably legal, gaming and education. There’s no reason why fintech and insurtech can’t follow. But their value is likely to be based on transforming the entire insurance sector, rather than insurance functions such as supporting agents or optimizing internal processes. The incumbents can handle that on their own.
