Credit: Ar_TH/Adobe Stock
For all its transformative promise, AI isn’t working out for a lot of organizations. In fact, a reported 95% of generative AI pilots are failing.
Those misfires send many organizations back to the AI drawing board, which kicks off a cycle of rollout, regroup, and repeat that often leads to pilot fatigue. Staff members resist new projects or tools and start asking: “Why do things seem so much more complicated when AI was supposed to make them better and easier?”
That kind of reinforced skepticism can be particularly damaging in an industry like insurance, where many organizations already tend to approach AI with a “prove it” attitude. For these cautious insurers, failed pilot projects certainly don’t do anything to build AI’s case. But AI’s potential to speed up and optimize areas like policy and claims review is too great to write off the technology over a few misses. Here are three ways to avoid pilot fatigue and make sure it doesn’t become the defining trait of your organization’s AI journey.
1. Take a methodical approach to AI pilots
What are you looking to get out of AI?
Believe it or not, many organizations jump into AI without having a firm answer to this question. It’s not just about giving your team a flashy new tool and hoping for the best. That’s a recipe for failure. Taking a methodical approach to AI means considering that question carefully and coming up with a solid business case to justify the spend for any pilot projects.
In a presentation, that case needs to include information like:
● Which job functions would be impacted by the AI tool.
● How the tool would be rolled out and implemented.
● What the expected outcomes would be.
This business use case should also detail the software purchase process, including who in the organization will lead negotiations and how they’ll evaluate vendors. This discussion will also include a rough timeline of the project and outline the steps to come beyond the pilot period.
2. Choose the right AI pilot testers
Who are your AI champions? The quickest way to sink any AI pilot is to try and force the technology on staffers who aren’t excited about it or open to its potential. Your AI champions are those who are looking to change their workflows, even if that means navigating some unfamiliar and possibly uncomfortable new processes.
Once these users have been chosen, it’s important to keep them in the loop with open and honest communications at all times. Make sure they know the expectations around things like:
● How they should be using the tools, which workflows they should be testing, and what kinds of things they should be reporting back.
● How the pilot will be expanded if the initial experiment goes well.
● What you’ll measure to determine whether and how to continue beyond the pilot period.
Being upfront and transparent about this last point is particularly important. Because nothing inspires pilot fatigue in a staff quite like a project without a plan that ends up seeming like a waste of everyone’s time.
3. Build enthusiasm for AI throughout the organization
Staff adoption is a key factor in how an AI pilot will perform in your organization. After all, even the most useful AI tool doesn’t stand a chance if you can’t convince people to use it.
Maybe they can’t see how AI is going to make their work easier or better. Maybe they don’t believe AI has what it takes to do their highly technical jobs. Maybe they’re stuck in their old way of doing things. Maybe they’re worried that AI is going to be too good at their jobs and eventually replace them. Whatever the reason behind their reluctance, the organization should be working hard both before and during any AI pilot to tout the technology’s benefits and help calm any fears.
For example, if you’re piloting an AI tool that helps speed up and double-check the complex policy review process, you should be making its case to your staff by pointing out:
● How the AI is meant to be an accelerator for their work, not a replacement – and what that increased speed will mean for them (maybe fewer weekend hours?).
● How the tool may help them discover new insights in those reams of documents that they might not have come upon otherwise.
● The technology’s limitations – it’s not meant to be perfect, but rather to work as a second set of eyes to provide them with additional peace of mind.
Put an End to AI Pilot Fatigue
Not every AI pilot is going to work. That’s just a reality of the experimental nature of the technology. But given AI’s productivity and efficiency potential in the insurance business – particularly in areas like claims and policy review – those projects need to happen. That’s why insurers that build disciplined, transparent, and people-first AI pilots today will be the ones leading tomorrow’s transformation – not just reacting to it.
Opinions shared in this piece are the author's own.
Dan Schuleman is the co-founder and CEO of Qumis, a lawyer-built, AI-driven insurtech reimagining how insurance policies are read and interpreted.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.