The gap between using AI and getting value from it
Brian opened the topic with two numbers. McKinsey reports that 88 percent of companies use AI in at least one business function, yet only 7 percent say they have scaled it across the organization. Boston Consulting Group found that 60 percent of companies report little or no material financial value from their AI investments. Adoption is widespread, and results are not.
The usual explanations are technology, funding, strategy, and fear of job loss. Brian pointed to a new Harvard Business Review article, “Middle Managers Will Make or Break AI Adoption,” that names a different cause.
Middle managers decide what happens to AI
The article calls the layer of managers who turn executive ambition into daily behavior the most common failure point in generative AI. Even when senior leaders have approved the budget, picked the platform, and announced the mandate, middle managers decide whether AI becomes a workflow, a side experiment, a compliance headache, or what the article calls a quiet casualty of organizational avoidance.
Andy agreed with the finding as it applies to companies as they are structured today. The C-suite issues directives and frontline employees follow direction, so the people deciding what gets prioritized are the middle managers. Many companies are removing that layer at the same time. Andy noted that the number of people each manager supports has roughly doubled over the last five to six years, which leaves frontline staff trying to make AI fit their workflows without a boss to help.
Where adoption breaks down
Brian summarized three problems the article identifies. First, treating AI as only a technology project. Handing out licenses is not business transformation. Second, uniform mandates. Middle managers have different profiles, risk tolerances, and support needs, and a single rollout plan ignores that. Third, nobody measures adoption. Very few managers tie AI use to formal performance evaluations or KPIs.
The recommended fixes are practical. Shift the focus from vendor features to manager enablement, give managers structured playbooks, align incentives so corporate AI mandates translate into team behavior, and coach managers according to their specific profiles.
Is the structure the real problem?
Andy raised a different question: whether working within the existing structure is the right approach at all. Managers manage the work, so removing them and expecting AI to take hold is not realistic. He added that the underlying issue is that companies have not rethought the work.
Too many large organizations in particular have not rethought the work. That’s what needs to happen.
Andy
He pointed to the Microsoft study the hosts discussed a few weeks earlier. When Microsoft rolled out AI internally, very little changed until the company rethought its workflows, focused on the end goal, and structured the work around it.
Brian pointed to Amazon, which has been flattening its organization and cutting middle managers while investing heavily in AI and implementing it across the company. He also played devil’s advocate: if every manager ends up with 50 direct reports, who manages the work, and does the CEO become accountable for everyone? He said he would like to hear from someone at Amazon about what the impact has been. Andy agreed that balance matters.
The West Coast offense and AI
Andy used a football analogy from a conversation with producer Jared. Before Bill Walsh, teams either ran the ball or threw long to a wide receiver, and when the passing game failed they went back to running. Walsh, working with a weak Cincinnati Bengals roster, built the West Coast offense around short passes to players in the slot. The players, strategy, and results all changed.
Andy’s view is that AI is the West Coast offense of corporate America. Companies swap in new technology, keep the same structure and roles, get the same outcome, and conclude the technology is at fault. The problem may be how they are thinking about it.
Who are the slot receivers in B2B?
Brian asked Andy to finish the analogy. Andy’s answer: people who are comfortable with technology, make decisions quickly, and show good judgment, but do not necessarily have deep domain knowledge. AI shortens the learning curve, so a determined, capable person can get up to speed on an industry quickly and manage projects with AI. Senior-level expertise still matters, and Andy was clear that very experienced people remain valuable at higher levels. The slot receivers, he said, are the hires companies have not made yet.
Brian acknowledged the hard part. Focusing on outcomes will put people with deep expertise, such as someone with 25 years managing channels, in a difficult position if their role changes.
You can’t start with how many people are we going to continue to employ. Start with how are we going to win.
Andy
Closing thought and the Mindshare Summit
Andy closed by urging B2B leaders to define the outcomes and the work first, then redesign how they accomplish it, because AI allows them to change the game. Brian and Andy said the conversation will continue at the Master B2B Mindshare Summit, March 16-17, 2027 at Convene in downtown Chicago. Registration is open at summit.masterb2b.com, and attendance is capped at 300.