5 practical things businesses should do about AI right now

Almost all company leaders say they expect AI to transform their businesses, but many are not yet adopting it, according to UserTesting's AI expert, Michael Domanic. Michael notes that the holdouts give four main reasons:

• Security fears
• Accuracy fears
• Questions about ROI
• They’re waiting for the rate of change in AI to settle down before they act

My first reaction this was to worry about those holdout companies. Yes, there are uncertainties about AI, but I think at this point holding back is a much bigger risk than jumping in. You're letting your competitors, or maybe a startup that's stalking your category, have a huge head start. Are you certain that you’ll be able to catch up later?

But then I reminded myself how hard it is for someone outside the tech industry to pick out the important tech from the fluff. Most companies have lots of urgent needs, and if you jumped on every tech trend it would be easy to disrupt yourself out of business (can you say “Google Glass”? I knew you could).

So maybe I’ve been looking at the problem the wrong way. The question isn’t “should you go all in on AI?”, it’s “what’s prudent to do about AI right now?” When you look at it that way, I think the answer is pretty clear. If I were talking to a business exec who’s uncertain about what to do about AI, I’d recommend the following five steps:

1. Focus on what’s possible with the current mainstream AI tools

AI isn’t one thing; it’s a suite of technologies and approaches that are evolving at different rates. For the next few years at least, I think there’s always going to be some new leading-edge AI trend that promises to revolutionize our lives. 

Right now it’s agentic AI, next month it may be something else. I’d like to suggest that you not spend a lot of time tracking those leading-edge concepts, but instead focus on the elements of AI that are relatively mature and stable right now. That means generative AI chatbots and their siblings, like image generation. They’re not changing all that rapidly, and we understand them well enough that your company can probably put them to practical use today.

2. In that spirit, I think the usage to focus on is generative AI bots as a general-purpose assistant for your employees

In UserTesting’s research on generative AI adoption, the knowledge workers who report big productivity increases use AI bots as generalized info-butlers and thought partners. When they have a problem or question, they run it past the bot to see what it can do. They rapidly discover what works best, and evolve new workflows that use AI. Many say they save five or more hours of work time a week.

Chances are extremely high that you already have some of these people in your company. They are already figuring out how to adapt generative AI to your business needs and processes. You should seek them out, learn what they’re doing, and think about how to adapt it to the rest of the employees.

Some observers (including The Economist, a publication I respect deeply) have suggested that those AI-adept employees will be hard to find because they’re hiding their AI productivity so they can work shorter hours. I doubt that’s widespread. In my experience, most AI adopters are evangelical about the benefits and couldn’t shut up about them even if they tried. The problem is not that they’re hiding; the company isn’t listening to them. So listen. 

3. Calculate the opportunity cost of not deploying

There’s a cost associated with implementing generative AI broadly in in a company. You have to pay for employee access to the bots, invest in training, maybe do a security audit, and so on. It’s not a crippling expense, but when you’re dealing with tariffs and a possible recession, you may be tempted to hold back on AI. 

Before you give in to that understandable caution, I encourage you to build a spreadsheet in which you list out the costs of deploying versus the potential productivity benefits you discovered in your audit of current AI usage. 

Maybe you’ll decide that AI’s not worthwhile to deploy now, but you may find that the benefits, when scaled across all employees, are far higher than the cost. Until you build that spreadsheet, you won’t have a solid basis to make a decision about AI deployment.

4. If you’re ready to deploy, accelerate the adoption curve

Once the enthusiasts teach you how they’re using generative AI, spread the word to the rest of the company through meetings, training, and other communications. The science of innovation teaches us that most people hold back on adopting an innovation until it has been proved by others. The more you spread the word on what’s possible and celebrate the people doing it, the faster adoption will move in your company.

5. If you're not ready yet, think about how generative AI can change your business model

Finally, even if you don’t decide to push AI adoption now, I strongly urge you to set aside time to think about how generative AI can change your business model. This will help you avoid being blindsided later, and you may even come up with a scenario that’s so compelling that you have to fund it.

Let me tell you a true story to give you an example of what I mean. I was meeting with the lead user experience thinker at a high-end hotel chain, one you’ve heard of. We were discussing how the company might use AI in its customer interactions. He said he’d come to terms with the reality that people would not come to his AI bot to plan their vacations. They see social media and hotel-agnostic sites like TripAdvisor as the place to do that, while the hotel website is a place to reserve a room after the plan is made. So he’s focusing on the use of AI in the reservation system.

That’s a good start, but I think he’s missing the point. Generative AI isn't just about simplifying current processes. It allows you to have a personalized 1:1 conversation with every customer, automatically, at scale. Stop and think about that: a personalized, ongoing, engaging conversation with every single customer. How can that conversation change the way you engage with them? 

When you think of what makes a high-end hotel special, what comes to mind? Other than the price and the quality of the sheets, I think it’s the level of service you get: the concierge who can help you with any problem, the front desk staff, the people who arrange the room just for you. 

What if the hotel chain had a bot that incorporated the knowledge and recommendations of every concierge in the chain, so it could offer personal advice during and after the booking? What if it learned the customer’s preferences and reported them to the hotel staff, so they could (for example) set the temperature in the room just right before they arrive? 

What if it recommended personalized activities around town and offered discounts on them automatically during the reservation process? What if you could create a completely bespoke experience for every guest, rather than doing it only for a short list of VIPs? 

Then ask yourself how you'd feel if your most important competitor did it before you.