Targeted artificial intelligence apps such as chatbots or as personal assistants have proven their worth, but when it comes to AI on a larger, enterprise scale, we still have our work cut out for us.

There are some companies, still in the minority, that are learning how to employ AI to compress “the distance between customer signal, decision, and action,” according to an analysis published by a team of McKinsey analysts, led by David Schiff, partner in the firm’s Austin office.

At least 37% of companies in the study attribute any level of earnings to AI, while 90% of companies regularly use AI in at least one business function. “AI enables commercial decisions to happen within the flow of work – continuously analyzing signals, recommending the next best action, and increasingly executing routine decisions with human oversight. As a result, a handful of people working alongside AI systems can deliver levels of personalization, experimentation, and optimization that once required entire departments.”

The question is how to go about achieving AI at such scale, to the point where it does deliver revenue. The trap into which many companies fall is an inability to push AI beyond the pilot stages, according to Nate B. Jones , a noted AI expert recently featured on Michael Krigsman’s latest CXOTalk show.

Most AI pilots never survive the transition to production and scale, Jones stated. The key is to stop treating AI projects as pilots. “We envision the pilot as a way to de-risk AI, but what we find in practice is that by naming and defining it as a pilot, you end up putting less resources behind it than you should. You pick more fragile and less important goals than you should, and you don’t get the learnings you want.”

Ultimately, AI needs to facilitate whole organization transformation – “it’s not something that you can effectively and easily sandbox into a little pilot space.” Pilots worked fine with traditional software applications over the years, but AI is a whole different animal.

Dive into AI right away, and pinpoint an area of the business where it will begin to have impact, he urged. “Pick the right project to work on, the right scope, identify places where if this works, it’s going to be transformational.” Avoid areas or processes where the returns would be ambiguous, he advised. “Start instead with something that's high leverage, but the processes are at least understandable to people in the company, and you can define the inputs so you can start to map them in and start to transform them with AI very deliberately.”

The efficiency of the models being employed also play in heavily to the levels of return as well from an AI effort. Tokenomics, or the costs of tokens, has the potential to drive up AI implementation costs well beyond initial estimates. Avoiding this trap is a matter of straightforward budgeting and shopping for cheaper models, Jones said.

“I know that sounds unsexy, but if you’re setting $5,000 as your budget, and the frontier model is eating up a lot of that budget in a given month, just push people and push your stack toward cheaper models and open-source models," he advised. "Make people make do within that budget. You will be surprised at how much of the work still gets done. You can get 80% or 90% of that value without the frontier model these days, and you get a tremendous cost savings.”

Ultimately, AI success may be in the eye of the beholder, Jones explained. “The question of whether it’s a failure or not is really a function of the organization’s ability to learn. Can the organization look at something and say, ‘this is what we learned from this pilot?’ ‘This is our next move in this space.’ ‘This is the lesson we learned from a people perspective.’ ‘This is the lesson we learned from a technology perspective.’ 'And this is what we’re going to do differently.’ If you’re doing that as an organization, it’s absolutely not a failure. It's actually a great learning opportunity.”