But, 58% of businesses say that less than 10% of their company’s digital budget goes towards AI.
So, it’s not surprising that only 15% of AI projects succeed.
This disconnect comes as no surprise to us. I was an analyst at Morgan Stanley and earned my technology stripes (and scars) at Apple under Steve Jobs, while my partner and wife, Helen, is a former CIO and corporate venture capitalist. We know the hallmarks of innovation window-dressing and transformation-by-outsourcing.
And after four years, deep in this “AI thing,” it’s clear to us where we are in transforming business with machine intelligence.
AI projects are failing because people don’t understand what AI actually is and don’t take the time to learn about it. If they did, the CEO would also be the CAIO.
This view is echoed by many others involved in enterprise AI transformations. At the VentureBeat Transform conference in San Francisco this week, John Freemont of Hypergiant and Heidi Messer of Collective[i] discussed the state of AI adoption in traditional companies; highlighting a deep misunderstanding by executives about what an AI transformation actually means—it’s complex, global across the enterprise and touches everything. “Do they have the guts to invest the time and money into actually doing this transformation,” Freemont asked. Messer quoted Tom Siebel, the founder of C3 IoT, saying that “the companies that adopt AI will acquire the ones that don’t.” By implication, the leaders that understand AI will take over from those who don’t.
AI has created a different need for knowledge at the top because AI has created a different kind of intelligence. Companies that are so-called “AI-first” are run partially by intelligent machines. In these companies, machines make more decisions in real-time than humans do. They should. AI is beyond human reasoning and beyond human response time so there is no choice but to trust the machines and have humans fix what gets broken afterwards. So far, it’s arguably been a better strategy.
In the past, strategic decisions could be “wait and watch, technology always gets cheaper.” Companies could afford to miss the first wave and still catch up. C-suite executives could be successful with only a cursory knowledge of technology, because that technology wasn’t intelligent in itself; its behavior was fixed by the designer and it just didn’t change that fast. But AI’s behavior is also dependent on its post-design learning experience so knowing some isn’t knowing enough. AI learns and changes after it has been deployed and C-suite executives need to know enough to manage their machine employees—just like they manage their human employees.
AI is a fundamental change in the brains that exist in a company’s competitive ecosystem. Now those brains can come from anywhere. The tech platforms are not going to stop researching, innovating and disrupting all other industries. People might be talking regulation, anti-trust and low tech living but the reality of the global geopolitical and economic system means that we need some big players developing AI. Not just the government and not just AI that’s optimized to get us to click on ads.
So, what will it take for you to succeed in AI?
STOP: Telling people to “get some AI” without spending enough time to understand it yourself.
STOP: Thinking that a few pilot projects are enough. Just doing something doesn’t mean it’s valuable.
STOP: Expecting results from AI initiatives in a quarter. AI takes time to get it right.
STOP: Delegating AI solely to your technical team. AI is an intelligence that needs to be led by the teachings of the humanities as much as by science and engineering.
START: Learning what AI is. At the level of the math. Not too much, but enough. How do you know enough? When you’re truly scared, but not because of the singularity or because a robot might take your job. You’re scared because you see how powerless your company will be against the combined intelligence of humans and machines that aren’t under your control. Do you remember the names of those companies that were too slow to adopt the internet? No, neither do I—and I was an internet analyst. No one remembers them because they are gone. They were outcompeted by the ones who understood the power of the internet and did something about it. You’ll be gone too if you don’t get serious about AI—now.
START: Getting everyone else educated in AI. That means everyone. Knowledge is created on the front lines and at the boundary of human and machine, which means that every person who makes a decision—whether in the field or in the board room—needs to understand it.
START: Building an AI strategy from the top. Start with your business’s entire ecosystem (because that’s what your competitors are doing) and spend face time with your senior team to determine where your greatest AI opportunities are. That means spending a day or two with your senior team building your AI strategy roadmap. This should be 100% of the agenda for your next offsite. Don’t have one scheduled? Do it now.
START: making every financial and technology decision through the lens of a future where partners, suppliers, customers, employees and competitors are human+machine, meaning they are masters of the unpredictable, operating at the speed and scale of the connected internet.
DO: Write a positioning statement to tell investors what your AI strategy roadmap really is.
THEN: Set the budget. Set the governance project and team. Set priorities vertically and stay personally involved.
Fire anyone who doesn’t think this is the future.
Reward anyone who makes good data available, while still protecting it. Promote people who understand how to lead people when a good portion of your employees are machines.
If you can’t teach it, you don’t know it. Get out there and explain to everyone just how AI will transform your business. And then give yourself a new title: CEO and CAIO.