--- title: The Middle Manager Gap Is Where Series B AI Rollouts Actually Stall slug: series-b-middle-manager-ai-gap date: 2026-04-20 excerpt: "Series B AI rollouts keep stalling at the same organizational altitude: the director, senior manager, and team lead layer. Here is what actually moves the adoption curve, drawn from fresh Grant Thornton, Docebo, and HBR research." featured_image: "https://bbtxujdxvidaghmhxkqs.supabase.co/storage/v1/object/public/generated-images/blog-1776664304105-series-b-middle-manager-ai-gap.webp" canonical_url: https://cerevisor.com/blog/series-b-middle-manager-ai-gap updated_at: 2026-04-20T05:51:45.376079+00:00 --- # The Middle Manager Gap Is Where Series B AI Rollouts Actually Stall TLDR Series B AI rollouts keep stalling at one specific altitude of the org chart: directors, senior managers, and team leads. Training spend does not fix it. Protecting 20% of their calendar for rollout work does. The layer most responsible for making AI stick is almost never the layer leadership funds. A Series B founder I know sent me a dashboard screenshot last Tuesday. Weekly active users of the internal AI assistant: 31% of eligible employees. Usage had been flat for six weeks. The frontline was trained. The tools were rolled out. The exec memo had gone out twice. Adoption was going sideways, and the board check-in was 12 days away. I have seen this exact dashboard six times in the last quarter. Same flat line. Same bewildered founder. Same underlying cause. I saw this week, in coverage of Grant Thornton’s 2026 AI Impact Survey, that middle managers were named as one of the two groups needing the most support to implement AI. And yet, in most Series B rollouts I review, middle managers are the one group with nothing in the implementation plan at all. > "Frontline employees (37%) and middle managers (30%), the people closest to AI in daily operations, were identified as needing the most support to implement AI." Grant Thornton 2026 AI Impact Survey, March 18, 2026 (n=950 business leaders across 10 industries) --- ## What Series B companies actually try Let me describe the standard Series B AI playbook, because it is remarkably consistent across the companies I have sat with this year. Exec team picks a flagship workflow, usually support, sales ops, or engineering-side tooling. Picks a vendor or builds internally. Runs a two-week pilot with a motivated pod. Pilot shows a [productivity](/blog/opus-4-7-first-week-productivity-check) bump. Exec team approves the rollout. Frontline gets a training session. Tools get provisioned. A Slack channel is created. A dashboard is set up. Then somebody, usually the COO or head of ops, is handed responsibility for “driving adoption.” In practice this means emailing VPs every Friday asking why the numbers are flat. The frontline training is often the most expensive line item in the rollout budget. And a striking share of it is wasted. Docebo released an enterprise learning report on April 7, based on a Centiment survey of 2,000 enterprise respondents across six countries. 85% of employees say their training does not help them use AI effectively. Docebo’s own framing, which I have kept coming back to, was that this is a change-management problem dressed up as a training problem. Here is what almost always gets left out of the rollout plan: the director, senior manager, and team lead layer. The people who decide, in practice, whether a team uses the tool. Whether 1:1s adjust to include AI coaching. Whether Monday standup still starts with the same three metrics or adds two new ones. Whether the person who just automated half their own job gets protected or quietly reassigned. Harvard Business Review ran a piece by Jeremy Korst, Stefano Puntoni, and Prasanna Tambe on April 8 about the executive-manager gap in AI adoption. The data cited: 92.4% of executives believed they had visibility into AI adoption inside their company. 84.7% of VPs. 76.3% of directors. The closer a leader gets to where the actual work happens, the less confidence that leader has in what is going on. --- ## Where it actually breaks Here is the binding constraint, and I don’t think I have ever said this out loud to a founder without getting a long silent pause in response. Middle managers at Series B scale cannot run an AI rollout on top of their existing job. Not in any meaningful sense. And almost no company has made explicit space for that work on the calendar. 6% of executives identify change leadership and workforce enablement as an essential skill for the AI era (Grant Thornton, 2026) That is the whole problem in two numbers. 6% of executives see change leadership as the core skill. 30% of the people who are supposed to drive that change say they need the most support. The layer most responsible for rollout execution is the layer least seen by the people setting the strategy. I watched a Series B CRO rebuild a stalled sales AI rollout over six weeks by doing one simple thing. They took four senior managers, dropped one quota-carrying AE from each of their teams, redistributed that pipeline among the rest of the reps, and handed those four managers 20% of their calendar explicitly for AI rollout work. Not “also.” Not “when you have time.” Protected and named. The adoption curve started moving within eleven days. Key Insight The manager layer is not resistant to AI. It is overloaded. Ask someone to learn a new tool, retool their team's workflow, rewrite a comp plan, and hold the anxiety of three direct reports who think they are about to be automated, all without giving anything back on the calendar. The rollout stalls. Every time. --- ## The pattern across the data Here is the pattern I keep seeing in both the research and practice. A piece on Rick’s Cafe AI on April 19 unpacked the widely-cited “95% of enterprise AI pilots fail” number and noted that vendor-led deployments succeed about 67% of the time, versus internal builds at around 33%. The common gloss is that vendors are simply better than in-house teams. I think the real variable is different. Vendor-led rollouts force a structure onto the organization: a dedicated project owner, explicit [change management](/blog/ai-adoption-trust-not-training), named decision points, protected time. Internal rollouts almost never do. The gap between vendor-led and internal success, in my read, is mostly a manager-capacity gap. Which means the Series B P&L line item most companies are missing is not “AI training.” It is “manager capacity for rollout.” Same dollar amount, totally different outcome. When an AI rollout usage line is flat, the next dashboard to build is manager-level activity. Are directors blocking time for rollout work? Are 1:1s shifting to include AI coaching? Are team leads rewriting how work gets assigned? If the answer to all three is no, the adoption number is never moving, and no amount of frontline training will change that. > Your middle managers are not the problem. They are the lever. --- ## What I’d tell you over coffee Pick four middle managers, by name. Carve out 20% of their week. Take something off their plate to fund that carve-out. Make each of them the AI owner for their function. Check in with them weekly, not with their teams. Then watch what happens. In my experience, that one move does more for the adoption curve than any tool in the stack and any training program on the vendor list. It also gives the board something it actually wants, which is a credible answer to the question of who owns this, named and funded. That is the move. Pick the names, protect the time, check in weekly. #### Sources - [AI News Digest, April 17: The AI Training Gap No One Is Solving](https://asanify.com/blog/news/ai-readiness-training-gap-april-17-2026/) - Asanify, 2026-04-17 - [The State of AI Adoption in the Enterprise (Q1 2026 Review)](https://cafeai.home.blog/2026/04/19/the-state-of-ai-adoption-in-the-enterprise-q1-2026-review/) - Rick's Cafe AI, 2026-04-19 - [Docebo Releases The AI Readiness Gap: The 2026 Enterprise Learning Wake Up Call Report](https://www.theglobeandmail.com/investing/markets/markets-news/Business%20Wire/1176098/docebo-releases-the-ai-readiness-gap-the-2026-enterprise-learning-wake-up-call-report/) - Docebo (Business Wire via Globe and Mail), 2026-04-07 - [Managers and Executives Disagree on AI and It's Costing Companies](https://hbr.org/2026/04/managers-and-executives-disagree-on-ai-and-its-costing-companies) - Harvard Business Review, 2026-04-08 - [2026 AI Impact Survey](https://www.grantthornton.com/services/advisory-services/artificial-intelligence/2026-ai-impact-survey) - Grant Thornton, 2026-03-18