An Even More Powerful AI Looms. And Controlled Chaos Is The Key

If you can’t hear yourself think amid the cacophony of incoming AI advice, I have some bad news for you. The next wave is going to be even more radical, and as a leader, every day you wait to unleash your people in an orchestrated way, you are behind. We would like to propose a solution.

It was Ravi Mhatre, a leading global investor and co-founder of Lightspeed Venture Partners, who told me what’s fast approaching. Today, AI needs people to guide the process. In the very near future, what’s called “agentic AI” will be able to perform complicated tasks with hardly any human supervision. Couple that with the supercharged processing power of quantum computing and, wham, everything changes—especially how we think about our workforce.

“Artificial intelligence has arrived at this tipping point where automation of human cognition means that the world is going to be a very different place,” said Mhatre.

“Agentic AI will automate brain power just as the industrial revolution automated human labor and the computer revolution supercharged human productivity,” he said. “These are massive transformational periods, and the third is happening faster.” Think five to six years.

In this column, we want to give you examples of what is coming that shook me up and I hope will do the same for you. And we’ll give you a roadmap for getting your head—and your organization—around this seismic change. But we are only getting started. We will be gathering a group of CIOs and CHROs in Vancouver on April 15 to dive into this deeper along with some of the greatest AI and human-capital minds in the world. More to come on that.

How big and soon is it anyway?

Ravi asked me if I thought enterprise was ready. Well, one large-company president recently told me he was shutting down AI exploration because it was too chaotic, and his CIO was worried about the mass proliferation of technology that would later need to be undone. That’s not the answer. The fix will come from masses, but masses working in a new way to harness the broadest and most inclusive crowd for big bold change. Not chaos, but managed crowdsourcing. And who will lead? Wrong question. The ‘sherpas’ can be any invested leader. It can be you—if you are game for inclusive change management.

If I were a CIO, CHRO, or CFO today I would be partnering with a progressive group of peers to figure out how to harness today’s generative AI and tomorrow’s agentic AI. This requires close teamwork. If that doesn’t happen, your company will be left behind.

Consider what the brightest minds in AI see already: Companies are already implementing “lights out manufacturing,” where robots do it all and humans are only overseers. If you’re in software, you will be able to cut your engineering workforce by 80%, but you’ll also have to hire systems designers with a very different skill set. Ditto TV and movies, which will deploy digital actors indistinguishable from carbon-based people. Goodbye, armies of make-up artists. Hello, more graphics designers.

“AI and robotics might soon perform surgery, but for now it can be the one that tells you pretty quickly: Do you have cancer or some kind of benign growth,” said Mhatre.

Drug development for decades has taken three to six years to move from discovery to pre-clinical trials, said Nitin Mittal, a principal at Deloitte and one of the firm’s AI leaders. Generative AI can compress that time frame into three to six months. Future systems leveraging quantum technology and AI may take three to six minutes.

Gains from old-school digitization could be measured in percentages. From AI it will come in multiples. “At that fast pace of innovation, there are going to be winners and losers, and 90% of the world cannot keep up,” Mittal said.

Predicting consumer demand, evaluating suppliers, negotiating contracts and tracking supplier performance can be turned over to AI, says Suketu Gandhi, a partner at Kearney. An AI-powered transportation network could run autonomous vehicles, optimize the fleet, optimize loading and figure out the best delivery schedule. Employment may increase in some areas, such as product design, but routine work is vulnerable, and companies over three to five years should be able to shrink their workforces by 20% to 30%, Kearney estimates.

What does this mean for the future of the workforce?

If your job is to think about HR, training, or your company’s talent, the future of AI is going to blow up your world. In a widely cited study, Goldman Sachs predicted that about two-thirds of occupations in the U.S. will be automated in some way, and a quarter of what people currently do will be done by AI, on average. Legal and admin work is the most exposed. This is just with generative AI. Imagine the next phase of agentic AI plus quantum computing.

At the macro level, this means laying off a lot of people, doing a lot of retraining and also hiring new people with different skills.

“Of course, humans will still be part of the equation,” said Mhatre. “But there are these cognitive tasks that are really pretty important, that represent some meaningful piece of what people do in organizations that now computers can do.”

OK, now what?

How do we move ahead? I hope the prior section makes us all realize we don’t have a choice. This is our call to action.

The answer is to harness the crowd. Everyone in the company should be invited to join the movement to experiment with AI and lay a path to the new world. Leadership in this case is not about control—it's about enticing possibility, encouraging experimentation, shining light on success and embracing fast failure with curiosity.

Here is a suggested path. I hope you’ll comment on how you are building your own AI movements and together we will meet the challenges. This is one of the topics of the gathering on April 15.

The steps to harnessing the crowd might look like this:

• Invite to a movement: Start with an open challenge for anyone to experiment. The goal is to get ahead of what AI will do to you, and instead define what AI can do for us. Once a month, all those stepping up to join the movement, from any function or division, will meet virtually. This can all be done asynchronously.
• Each individual or group has five minutes to present their progress, laid out on only one page, including: 1) what they’re learning, 2) the indicators and metrics of success and progress, 3) where they are struggling, and 4) the plan for the next sprint over the coming month till we meet again.
• All those present open a shared Google document and give feedback to their peers who just presented. 1) the challenges and risks, 2) ideas and innovations to build on the presentation, and 3) offers of help and suggestions for followup. This is a five-seven minute process that combines presentation and feedback. In total, assuming a half dozen are presenting at any given time it’s over in about an hour. Everyone has exposure to best practices and company-wide shared learning. Amazing.
• When it’s done, the executive Sherpa group tackles the macro questions that arise: What are the overarching themes that are emerging, what are our worries and concerns, where are the big opportunities we may want to focus on as an organization. Each project will raise new and complicated questions. What does this mean for HR policy or financial modeling? For technology and tools?
• And then we keep going—and going. The same process should be repeated in monthly or quarterly sprints, because that’s how fast the technology is changing.

And remember, this is a team sport. The implications of AI are so dramatic and wide-ranging that it will touch every aspect of the organization. Let's have the CIO, CHRO, CFO and GC in the room. This will require the collective brainpower and trusting collaboration of the entire c-suite to conquer.