The Hard Part Is Just Beginning: Senior Leaders Confront AI’s Next Test

Listen to an AI audio recording of this newsletter

One key reason I value our Outthinker roundtables is that they rarely stay theoretical for long.

We may begin with a framework. We may invite someone with deep expertise to discuss the issue. But quickly, the conversation moves into the real work of leadership:

• What happens when the technology is ready, but the organization is not?
• What happens when the opportunity is obvious, but the operating model gets in the way?
• What happens when employees can see what AI can do, but do not yet trust what leaders will do with the value AI creates?

That is what happened in a recent Outthinker roundtable with ​Scott Snyder​, a Senior Fellow at Wharton, adjunct professor at Penn Engineering, author of ​Your AI Life​, and co-author of ​Goliath’s Revenge​. We were joined by innovators and strategists across three continents.

Scott has experienced three eras of technological disruption: digital/ecommerce, mobile, and now AI. But what made the session valuable was not simply his presentation. It was the way leaders in the room translated his ideas into the problems they are facing inside their own organizations.

The conversation crystallized something I have been thinking about as the AI Adoption Paradox — the more powerful AI becomes, the more adoption depends on something distinctly human: trust.

AI gives organizations new capacity. But that capacity only becomes an advantage if people trust how leaders will use it.

Scott opened by naming the tension many leaders are feeling: “I’ve lived through a bunch of tech waves, probably like many of you, and I’ve never seen one with the level of excitement matched with the level of trepidation and fear that I have seen with AI.”

AI adoption is uneven because work is uneven

​Kevin Darbelnet​, founder of ​Strategy Matrix​ who has a background in the consumer goods and physical products world, pointed to one of the most important gaps in the AI conversation. Back-office use cases are moving quickly. Demand planning, supply planning, marketing, customer service, and other knowledge-work applications are obvious places to begin. But the manufacturing floor is different.

Kevin pointed out that AI seems to be moving quickly in office functions like planning, marketing, and customer service, but much more slowly on the manufacturing floor, where the opportunity may be just as big.

That matters because AI is not equally easy to apply everywhere. It is one thing to use AI at a desk. It is another to bring it into a factory, where work is physical, systems are older, and people need tools that fit the way they actually do their jobs.

It is one thing to give a knowledge worker access to a generative AI tool. It is another to redesign how a frontline operator, technician, nurse, warehouse worker, underwriter, or field employee does their work.

Scott added a useful distinction. Physical environments are harder to manipulate and adjust than digital ones. Legacy systems are more complex.

As Scott put it, “Most of these physical workers aren’t sitting at a desktop, at least most of the time of the day. So how do you deliver the AI intelligence to them at a form factor that works?”

That is why AI adoption is not an IT rollout. It is an operating-model challenge.

Trust is the hidden adoption infrastructure

Scott’s presentation framed AI adoption as behavior change, driven by both skill and will. On one side is ability: literacy, proficiency, and eventually mastery. On the other is mindset: openness, engagement, and excitement.

The interventions that move people along that curve are not just training modules. They include time allocation, peer support, real experiences, recognition, and rewards.

Organizations cannot simply announce AI transformation and expect adoption to follow. They have to design the conditions under which people will want to use AI, learn with it, challenge it, and trust it enough to change how work gets done.

This is one practical answer to the AI Adoption Paradox. If the message to employees is, “Use AI so we can squeeze more from you,” resistance is rational. If the message is, “Use AI so we can remove repetitive work, protect expertise, and redeploy time toward higher-value work,” adoption becomes much more likely.

The difference is leadership.

Leaders need quick wins and bold plays

​Atul Tripathi​, who has experience at PepsiCo, McKinsey, and P&G, raised another issue many executives will recognize. AI may be moving quickly, but large enterprises are not clean laboratories. Data is fragmented. Systems are aging. Integration is difficult. Foundational work can take years.

Atul asked how leaders can balance the urgency to move on AI with the slower work of improving data, systems, and infrastructure. More specifically, he asked how to “build like some credible value before the full data environment is fixed.”

That may be one of the most practical AI strategy questions leaders can ask right now.

Scott’s deck offered one way to think about that balance. Quick wins build confidence. Bold plays force the organization to imagine how AI could reshape a customer experience, an operating model, or even the business model itself.

He also framed this as becoming a “two-speed business.” One speed moves today-forward, finding productivity improvements and near-term use cases. The other works future-back, asking what a 10x change in operations, customer experience, or business model could look like.

Most companies are more comfortable with the first speed. They can identify use cases, run pilots, and find efficiencies. But the second speed is where advantage begins to form.

That is where leaders ask a different kind of question: What can we do now that we could not do before?

• Can we serve customers we previously could not serve profitably?
• Can we reduce cycle time enough to change the economics of a business line?
• Can we take work that was once trapped in functional silos and redesign it around outcomes?

This is where AI becomes more than a tool. It becomes an opening for new possibilities.

If leaders treat AI primarily as a cost-cutting mechanism, they may achieve short-term savings while damaging trust. If they treat it as a way to expand human capacity, they can begin to create a different kind of organization.

That real test of AI adoption

The organizations that win will not necessarily be those that adopt the most tools the fastest. They will be the ones that build trust, redesign work, and create a credible answer to the question every employee is quietly asking:

• What happens if this works?
• If AI saves time, where does the time go?
• If AI improves productivity, who benefits?
• If AI changes a role, will people be prepared for the next version of that role?

The answers cannot be buried in a slide deck. It has to show up in decisions, incentives, staffing, recognition, and communication.

That is why these roundtables matter. They bring together leaders who are not watching disruption from a distance. They are inside it, asking the questions that determine whether AI becomes another wave of tools or a genuine source of advantage.

The AI advantage is not simply about models, agents, or data.

It is about what leaders do next.

Join ​Outthinker​ to be part of the conversations where senior leaders pressure-test what comes next.