The Right Dosage Of AI For Business

AI can be a great game-changer in nearly any industry. It ushers in massive changes in how we do nearly everything, or in CEO-speak, it leads to “business process efficiencies.”

But AI doesn’t inherently do this: it does it if the applications are balanced and fitted to a business model. I often say that new tech (any new tech) can be either a help or a hindrance to a company depending on the fit.

There are also ambiguities that we have to address in BPE (business process engineering) with AI.

“When we don’t have clarity on how the predictions are generated, it becomes very difficult for us to build defensibility or accountability,” writes Lian Parsons at HBS, showcasing the idea of Harvard DCE Professional & Executive Development Mark Esposito. “The challenge that we have to address is deciding the role of humans and how we make the technology human-centric. Even if humans are largely relieved of a repetitive task, it doesn’t mean that we are removing the delegation of power to humans. If we’re unclear about that role, then we have to intentionally try to define it.”

And then, what about that fit?

Taking the Medicine

In a recent TED talk, Karim Lakhani, a Harvard Business School professor, discusses this with a different metaphor: in terms of a patient taking a drug.

“We actually don’t know the right dose to take, the efficacy of the product, the side effect, the toxicity, or even the right diet to follow when we’re using this drug, but this drug has been introduced across hundreds of millions of people,” he says, noting that his organization has been conducting a series of “clinical trials” of AI, sort of like you would with a new drug, to actually take the temperature of adoption trends and see what happens when people use the AI.

What emerges, he suggests, is a “jagged technological frontier” – a webwork of systems and processes that evolve at different rates, and all of the challenges that come along with an ecosystem like that.

“There are some things that AI is very good at,” Lakhani says. “When you use it for (those functions), AI performs incredibly well, and people get better. But when you use AI for a task where it’s not good for (that use), your performance drops, and drops dramatically.”

Productivity Gains Measured in AI “Clinical Trial”

In terms of productivity, he noted that in one study, AI use led to 12.2% more tasks completed, on average, tasks completed 25% more quickly, and tasks completed with 40% higher quality.

But, he mentions, it’s important for leaders to avoid “falling asleep at the wheel” when it comes to the use cases of innovation.

Procter & Gamble’s Cybernetic Teammate

Citing randomized outcomes, Lakhani talked about AI transforming itself from a tool into a teammate.

“The first part that we see is a massive productivity boost,” he says.

Then, he notes, things really take off.

“While the performance capabilities of AI models, that’s increasing exponentially on an every six months basis, every nine months basis, the absorption capability of most organizations is linear,” he adds. “It’s linear. What that means is that every six months or nine months, you’re going to be exponentially behind what the models can really do.”

Companies, he argues, won’t be able to absorb the activity quickly enough to tackle an exponential curve. Is that a problem? I guess, in a way. Partially just in competition.

Lakhani described the process of trying to stand out in the crowd. It was interesting when he talked about trying to differentiate your business based on $20 a month, or a basic model subscription price.

In other words, you have to use a bounded value to create enormous value for your organization…

The Four-Act Play

Another key concept that Lakhani brought up late in the game was what he called a “four-act play” as a roadmap for leadership.

The first stage is to learn all you can about AI – get to know the systems and the models, and what they can do. Understand the context for business. Look at how things are changing over time.

The second step is to do things with AI. This, he says, is where things often get mixed up. He talked about leaders delegating all of the activity to others, where they should be taking on certain tasks themselves.

The third stage is imagining what can happen in the business, and the fourth stage is making those things happen – actualizing your ideas,

Essentially, Lakhani walked us through a lot of the productivity that comes along with AI transformation, as tools become agents, and agents become human-like in their function.

Will we have agents as bosses, and what is that going to look like?

Stay tuned.