Photo by Campaign Creators on Unsplash
UNSPLASH.COM
Despite pouring billions into generative AI over the past three years, the vast majority of companies aren’t seeing a payoff. A recent MIT study caused a flurry with its findings that 95% of organizations investing in AI report no measurable return. Obviously, it’s not for lack of ambition—leaders everywhere are frantically experimenting with all manner of copilots, agents, and chatbots, hoping to land on the single solution that will unlock productivity and transform their business.
The problem is that many leaders expect this change to be immediate. They launch pilot projects with splashy announcements, hoping to woo customers and intimidate the competition. But all too often, these AI projects get stuck in proof-of-concept limbo.
That doesn’t mean AI is a dead end. It’s actually the opposite—the companies that are finding value understand that this is a long game. Like any breakthrough technology, the early returns are uneven. But strategic, targeted investment separates the winners from the rest, and the leaders willing to build sturdy foundations and practice patience will be the ones who ultimately come out ahead.
Why Quick Fixes Fail
Organizations are making a lot of promises about what their AI-powered processes can achieve. But the fact is that under the sleek marketing campaigns, the implementation is riddled with issues, including “brittle workflows, lack of contextual learning, and misalignment with day-to-day operations,” the MIT report concludes. Ultimately, only 20% of these tools are reaching the pilot state, and 5% are reaching production. As one CIO put it: “We’ve seen dozens of demos this year. Maybe one or two are genuinely useful. The rest are wrappers or science projects.”
The mistake here isn’t the investment in AI—it’s viewing it as a siloed solution instead of an integration. As Ramyani Basu wrote in Harvard Business Review, “leaders urgently need an approach to help them make smarter, value-focused decisions about AI—one that will help them avoid throwing money at new solutions and hoping something will stick.”
The takeaway? AI doesn’t fail because the models aren’t powerful. It fails because too many are looking for shortcuts instead of laying the groundwork for sustainable results.
The Power Of Building Strategically
A lot of press these days is focused on AI’s disappointing returns. But far more instructive are the lessons of the other five percent—the ones who are successfully scaling AI and achieving real structural change.
While it’s not too late to start implementing AI strategies now, organizations currently seeing returns had many of the essential building blocks in place early on. As Ben Lorica, the editor of Gradient Flow, explained in a recent interview, the companies experimenting with AI before its recent explosion are “generally doing better,” having likely already had data infrastructure, engineers, and other talent in place. “So now with generative AI, they can just extend those initiatives to another form of AI,” he said.
That doesn’t mean that those without a head start should scramble to retrofit their operations. It does mean that there’s solid evidence that when it comes to AI, haste makes waste. Rather than optimize yesterday, invest in tomorrow, advises HBR’s Basu. “Value creation is non-linear, so you must think tomorrow-first, then work backward to make the right investments in areas where your organization can leverage disproportionate value.”
To Build Or Buy?
I was a coder long before I was a founder, so I understand the impulse to make every tool from scratch. But when it comes to AI, this is not necessarily the right path. In fact, one of the key takeaways from MIT’s study was that even though more organizations opted for internal development, success rates favored external partnerships.
That’s not to say internal development is never worthwhile. If building a tool will give you a competitive edge and fit into the unique landscape of your business, that’s a sign to go for it. If it has an obvious place in the customer experience, and you have the bandwidth, talent, and systems to maintain it for the long haul, build on it.
But in many cases, the smarter path is to leverage the tools that already exist. There’s no need to compete with tech behemoths if you don’t have to, especially if they’re peripheral to your core offerings. If a good-enough solution already exists, use that. If it doesn’t, there’s a good chance it will soon, especially if it solves a widely applicable pain point.
The fact is, transformative technology takes time to reach its full potential. The cylindrical steam engine was invented in 1769. Despite this, water wheels continued to reign supreme for decades because of the difficulty of transporting the coal the engines required. The same halting, non-linear path was also followed by electricity and computers.
The lesson here isn’t that investing in AI is a waste. It’s that returns follow a long, uneven curve. Leaders who prioritize building fundamentals, have patience, and a clear sense of purpose will be the ones best positioned to reap the rewards when the technology matures.
link