Given the developments in AI over the past year, it’s no surprise that the hype around artificial intelligence in general, not to mention generative AI capabilities, is at an all-time high. Regardless of where any technology is in a “hype cycle,” it’s important to be able to understand what is real and worth our attention and budget, and what isn’t.
This is the focus of last week’s conversation on the Insurance Unplugged podcast. Host Lisa Wardlaw spoke with Steven Abel, Partner at Oliver Wyman, who, with 25+ years of experience in consulting and executive roles, is known recognized for his pragmatic approach to tech transformation.
Read on for an excerpt from Lisa’s conversation with Steven as he provides guidance for navigating the current AI hype.
Distinguish Between Real vs. Distraction
What’s real is the tooling that solves business problems today. The ability of modern tooling to consume extremely large data sets and generate predictive analytics to solve business problems in novel ways is groundbreaking. These are techniques, tooling and data that Oliver Wyman is applying on a daily basis with clients to solve real problems and make real money.
The capabilities of generative AI to write your PowerPoint presentation or create a poem are also real, but they can be a distraction. Generative AI has provided access to really powerful tooling. However, this focus on exciting capabilities and edge cases is distracting from the types of real problems that companies are dealing with and the solutions that are already available to solve them.
The idea that AI is going to solve everything is like trying to build a house out of nails. It’s an interesting concept, but not the best way to build a house. AI is a very useful tool that can do some amazing things, but it is not the only tool or material that you need to build your house. The distraction is the thought that AI will do everything.
Be Intentional About Vendor Selection
While the popularity of ChatGPT is making some people aware of AI, it’s important to remember that AI and machine learning have been around for decades. What has changed is accessibility, the ability of your average citizen to be able to compute at scale.
The practitioners that have been doing this a long time think very carefully about bias, about governance. Every practitioner I know that uses generative AI in their daily world knows about things like hallucinations, and they think about validation. What concerns me about some of these tools is the lack of an idea of governance to say, is this real? Is this not real? Is there bias? Is there no bias?
So, when we think about evaluating vendors or even solutions for our clients, we think almost exclusively about governance and bias because it’s incredibly important, but it’s something you can’t see. You can see a screen and you can have one of these solutions generate a nice poem for you, but bias is very hard to see. It’s very hard to detect. I worry that the casual user or even the casual InsureTech or FinTech won’t be doing adequate work around rooting out bias around governance.
For companies evaluating these technologies: as you’re looking at your vendor ecosystem, do it with intent.
If someone says, “I have AI,” you need to know what it’s doing. You need to be aware of the underlying calculation and how you’ve gotten to your conclusions. It’s not a bad idea for everyone to be thinking this way.
Start with the Problem You Need to Solve
If we think about insurance, the idea that you can use large data sets to generate hyper-personalized experiences, both from a risk standpoint and also a revenue generation standpoint: the tooling and the analytics are real. It exists now and companies can use this to not only look at your business from a loss perspective, looking at the past, you can now look at the future. You can consume extremely large data sets to make predictions that aren’t possible with the smaller data sets that existed even just five years ago.
Consuming large data sets and being able to do amazing predictive analytics is incredibly useful to so many types of business problems. We have to get a very clear picture of the tangible business problem and find a way to generate the result at speed.
One thing that we want everyone to take away from this conversation is this: Think about your problems, think about the technology, in a way that it can solve today’s problems.
Listen to the full episode with Lisa Wardlaw and guest Steven Abel on the Insurance Unplugged podcast.
