AI rollouts are difficult to get right for a lot of reasons – among them: employee anxiety and resistance to change, the rate of AI’s evolution, and the friction of rethinking work from first principles.
But OpenAI’s former Head of Go To Market, Zack Kass, suggests that these reasons might just be convenient scapegoats for overwhelmed executives. The biggest barrier he sees as an AI consultant is actually leadership, and the fact that very few leaders have the vision, clarity, and gumption to do this right.
The real reason enterprise AI is stalling
Zack has worked with hundreds of executives in the board rooms of Fortune 250 to Fortune 10,000 companies, and he says the determining factor in whether enterprise AI efforts succeed is executive conviction. He defines that as the belief, at the highest levels, that adopting AI is not optional but necessary.
Without that belief, no amount of selling, persuasion, or tactical planning from consultants like him will make a difference in the outcome. And he says these board rooms are suffering from a critical lack of executive conviction, for a few reasons:
- Leadership turnover. Executive teams are being swapped out, which stalls strategic decisions.
- Aging leadership are avoidant of change. Older executives have no intention of making large, disruptive decisions so close to retirement.
- Misaligned incentives. In some companies, leaders aren’t rewarded for taking big risks or pursuing bold innovation, so they avoid it.
“Courage is not actually cheap,” Zack says. It requires real personal and political capital to act decisively on AI. Many leaders aren’t willing to spend that capital.”
As a result, executives hide behind excuses, like not having the right talent.
“But if you believe this is going to change everything, unless you are truly an under-resourced company, you can make things happen,” he says.
The current shortcomings of enterprise AI leadership
This lack of executive conviction is ultimately what creates the roadblocks that become excuses for failing deployments. Here’s what Zack sees:
1. A climate of both urgency and confusion
We know all about IC anxiety around AI, but that could largely be a trickle down effect from executive panic.
“Right now, the friction point for leaders is this terrifying realization that something important is happening, mixed with this sinking feeling that they don’t know enough,” Zack says.
This leads most executives to analysis paralysis – and to outsource strategy to others who seem to know what they’re doing.
Zack says many executives are behaving like “non-economic actors” – acting from urgency, emotion, or perceived importance rather than rationality.
2. Borrowing perspectives vs. forming a clear vision
Zack has asked leaders across industries to imagine success with AI over a 10-20 year horizon, and very few can. Without this definition, he says leaders don’t have a “panacea” – a compelling, collective vision of what they’re working toward.
“If you can’t do this as a leader, as an industry, as a business, then the market will do it for you and it’s going to turn out really weird.”
The result is rudderless industries taking cues from outside voices like mainstream media, market analysts, and podcasters. They might hire people to teach their teams how to leverage AI based on these external opinions, but they can’t articulate what they want people to be able to achieve at the end.
Without that context and without goals, employees don’t have an anchor point to know what they should expect and what’s expected of them.
“Businesses have an obligation – not even an opportunity, but an obligation – to tell themselves and their industry what great looks like.”
3. Concealing plans, which breeds fear
Because leaders don’t have confidence in their path forward, Zack sees a lot of executive teams keeping their AI plans quiet to avoid spooking their team.
“But as far as your team can tell, you just want to observe how they work with AI so you can ultimately fire them.”
All of these strategy shortcomings create the symptom that leaders blame their stalled deployments on: employee anxiety.
How to right the ship
The biggest piece of advice is the obvious one – commit to doing this or don’t. There is no low-risk way to adopt AI that yields the big results you want.
Here’s what Zack recommends for leaders who are trying to figure this out:
Step 1: Run a conviction check
Ask yourself and your executive team this “yes or no” question point blank: “Do you think we need to invest in AI right now?”
If yes, proceed to step 2. If no, don’t waste anymore time or resources on AI right now.
Step 2: Convene the top for strategy workshops
Zack recommends starting with a few long sessions that include the CEO and executive team to get serious about the ‘why’. In these sessions, you need to define a couple things:
- What does great look like? What is your vision for what can be markedly better for your customers and your industry 10 years from now if you start today?
- What are you accomplishing besides cost-cutting? Zack says you need to be thinking about expanding access to your service or product (who gets it, how easily, how affordably etc.), rather than margins-only gains.
“The single most interesting thing that a leader can talk about is improving access,” Zack says. “Most leaders describe a world that is simply less expensive to operate in – that’s pretty lazy and also going to win you no favor with your employees.”
Step 3: Own your narrative
Don’t allow what you create in that session to get warped by outside perceptions and fear of reception. Commit to it and then live by it.
Share the answers to the previous questions with your team clearly and definitively. Make your confidence in this vision contagious. And give them real goals and context to help them understand why you’re asking them to shoulder your AI initiative.
Three imperatives for leaders
No one is saying leading through the age of AI is easy. But this is an opportunity, if you’re willing to take the risks. Here’s the directive to leaders:
1. Lead confidently even through uncertainty. Zack references Captain Shackleton, who led his crew through two years of being stranded in the ice in Antarctica. He remained calm and yet honest – and they followed him until he brought them home.
Confidence doesn’t mean pretending to have all the answers. It means creating stability through your demeanor and communication so your team feels secure enough to act.
2. Stay humble and adaptable. Acknowledge that your current strategy may not be the right one, and be willing to explore other avenues. This is not the time to double down blindly – you need to be willing to pivot, experiment, and learn rather than defend a potentially obsolete plan.
3. Know your limits. Identify the point where automation should stop – both in operations and in customer experience. Some processes, touchpoints, or services have intrinsic human value or strategic importance that shouldn’t be replaced, even if automation is possible.
P.S. If you loved these insights, you can pre-order Zack’s upcoming book The Next Renaissance: AI and the Expansion of Human Potential.
