Companies Are Investing in AI Like They Expect It to Fail

As an investor in dozens of businesses, including early investments in companies like OpenAI, I’m used to backing ideas that may take years to prove themselves. But as a CEO responsible for 850 employees, I also must make payroll, protect margins, and deliver results. I don’t have the luxury of waiting years for every bet to pay off.

Lately, I’ve been struck by how differently these two worlds think about AI.

In venture capital, we ask what a business could become if the technology works and the founders get it right. In corporate boardrooms, the questions are more immediate. What’s this going to cost? What happens if it doesn’t work? And, of course, how soon will we see a return?

Any responsible CEO should ask those questions, but I’ve seen how easily they can become an excuse for thinking too small.

Look at where most large companies are putting their AI dollars. They’re buying productivity tools, automating routine work, and chasing efficiencies they can demonstrate within a budget cycle. Some of that will pay off handsomely.

But if the biggest thing AI does for your company is help you do the same work at a lower cost, you’ve missed the larger opportunity.

VC teaches you that you can’t always judge a new business by the economics of an old one.

When I look at an early-stage investment, I know there’s a good chance it won’t work out. What interests me is whether the founders have spotted an opportunity others have overlooked, whether they’re capable of building something around it and what the upside might be if they succeed.

Propose an AI project that will shave a few percentage points off operating costs, and you’ll probably find a receptive audience. Propose something that could change the business model, and suddenly everyone wants another spreadsheet.

Now compare that with how most big companies make investment decisions. We’ve built layer after layer of financial controls, approval processes, and reporting requirements. Much of that is necessary. But I’ve seen how it can work against us. Propose an AI project that will shave a few percentage points off operating costs, and you’ll probably find a receptive audience. Propose something that could change the business model, and suddenly everyone wants another spreadsheet.

As a CEO, I understand the instinct, and yet playing it safe carries its own risks. You can protect your margins today and still make decisions that leave you struggling five years from now.

I think companies need to get more comfortable placing bets they can’t fully quantify at the outset.

That’s not the same as throwing money around. In VC, we put money behind a team and an idea, see what they can accomplish, and decide whether they’ve earned more capital. Sometimes they haven’t. You take the loss, learn what you can, and move on.

Why should an AI investment inside a corporation be so different? Give a few good people the chance to pursue ideas that could move the needle. Set some milestones, make sure they’re accountable and resist the temptation to demand a detailed five-year forecast for something that hasn’t been built yet.

That only works, though, if the people involved believe they can take a risk without putting their careers on the line.

I’ve found that employees pay a lot more attention to what happens when something goes wrong than to what the CEO says about innovation. If a failed experiment becomes a black mark on someone’s record, people get the message. The next time they have an unconventional idea, they’ll probably keep it to themselves.

With AI, we don’t have years to sort this out. The technology is changing while companies are still debating whether to approve the budget.

There’s another lesson from the VC side worth remembering: We invest in people as much as ideas.

AI responsibilities shouldn’t simply be handed to the tech department with instructions to find efficiencies. Other members of the organization — including those who understand customers, products and operational problems — also have a stake...

AI responsibilities shouldn’t simply be handed to the tech department with instructions to find efficiencies. Other members of the organization — including those who understand customers, products and operational problems — also have a stake, and they may question why certain processes exist at all. They have a right to do so. Otherwise, companies risk spending fortunes improving what they should be eliminating.

I’m not suggesting that CEOs abandon financial discipline or run their businesses like venture funds. I wouldn’t run my own company that way. But established companies have advantages startups would love: customers, market knowledge, distribution and resources to bring new ideas to market at scale.

CEOs already understand AI’s potential. The bigger question is whether they’re prepared to change how their companies allocate capital, reward people and make decisions.

Will they fund ideas before returns are obvious? Reward smart risks even when they fail? Move quickly enough to capture opportunities? Those questions matter far more than which AI tools to buy.

The biggest danger isn’t that companies will waste money on AI projects that flop. Some will. It’s that they’ll become so determined to avoid bad investments that they never make one capable of changing their future.

Tim Ringel is Founder and Global CEO of next-generation international advertising group Meet The People.