The News
Matt Gjertsen, founder of BUILT and former Head of Training & Development at SpaceX, has published a new book titled “The Minimum Viable Manager,” which argues that effective AI adoption inside organizations depends less on tool selection and more on managerial capability. Gjertsen’s central claim is that AI now functions as a class of non-human team member, and that every employee who interacts with AI tools must therefore develop the skills of a manager: understanding AI strengths and limitations, setting clear goals, and creating feedback loops. BUILT, which Gjertsen founded in 2021 after a career as a U.S. Air Force instructor pilot, works with technical organizations on leadership development in high-stakes, high-change environments.
Analyst Take
The Real AI ROI Problem Is a People Problem
The conversation about AI underperformance in the enterprise has been dominated by discussions of model selection, infrastructure, and data quality. Gjertsen is pointing at something different, and frankly more uncomfortable: the bottleneck is human behavior. Specifically, the absence of structured management discipline around AI outputs. That’s a harder fix than swapping one large language model for another.
This framing has real teeth when you look at where AI investment is actually flowing. According to ECI Research’s 2026 Application Development survey, 53.5% of respondents selected “AI-enabled development tools” as a top investment priority for the next 12 months. That level of commitment reflects genuine organizational appetite. But capital deployed into tools without corresponding investment in how employees direct, evaluate, and iterate on those tools is capital that will underdeliver. Gjertsen’s argument is essentially that most organizations are skipping the management layer entirely.
What “Managing AI” Actually Means for Engineering Teams
For developers and engineering leaders, the practical implication is specific. AI-assisted coding, deployment automation, and AI-driven anomaly detection are not self-correcting systems. They produce outputs that require human judgment to validate, redirect, and improve. ECI Research’s 2026 Application Development survey found that 61.7% of respondents have AI-driven anomaly detection in place as an observability strategy, which suggests broad deployment. But deployment is not the same as effective use. If engineers treat AI-generated alerts or AI-suggested code as authoritative rather than as a first draft requiring review, the feedback loop that Gjertsen describes never forms, and the system doesn’t improve.
The “manager of AI” framing also has a direct read-across to security. In the same 2026 survey cycle, ECI Research found that 45.3% of respondents reported that AI-assisted development has moderately increased security risk. That finding is not an indictment of AI tools. It’s an indictment of unmanaged AI outputs. Developers who treat AI code suggestions as ground truth, rather than as proposals that need security scrutiny, are bypassing exactly the oversight function Gjertsen argues is now a core professional skill.
The Business Case for Leadership Development as AI Infrastructure
For ITDMs, the strategic implication is that workforce development is not a soft complement to AI investment; it is a prerequisite for AI investment generating returns. Gjertsen’s framing of leadership training as enabling AI performance should land differently than generic change management pitches. This is not about adoption curves or digital transformation messaging. It is about whether the people deploying AI tools have the skills to direct them, catch their failures, and improve their outputs over time. Organizations that treat AI tools as autonomous contributors rather than as managed assets will find themselves with high spend and unclear returns. That dynamic is already visible across the market, and it’s what Gjertsen’s book is attempting to address.
Looking Ahead
The “human as AI manager” model that Gjertsen describes is going to become a structural feature of how mature organizations think about AI deployment, not a fringe perspective. As AI tools take on more consequential tasks, including code generation, deployment decisions, and security triage, the cost of unmanaged AI output will become measurable and visible. Expect to see enterprise training vendors, HR technology platforms, and even AI tool vendors themselves begin building “AI management” competency frameworks into their offerings. BUILT is early to this positioning, and that’s a real competitive advantage if the firm can scale its methodology.
The harder question is whether organizations will actually invest in this capability before the costs of not doing so become obvious. History suggests they’ll wait. The pattern with previous technology transitions, cloud, mobile, DevOps, was that the human skills gap closed years after the tool adoption curve peaked. AI is moving faster than any of those prior waves, which compresses the window. Companies that treat AI management capability as a first-class investment alongside their tooling budgets will separate from those that don’t, and that gap will likely be visible within two to three years.
Stay Ahead of Application Development Trends
Get weekly analyst insights, research notes, event coverage, and AppDevANGLE updates delivered directly to your inbox.
Subscribe for Weekly Insights
Join technology leaders, practitioners, and GTM teams following the trends shaping modern software delivery.
Looking for deeper research access?
Explore ECI Research reports, survey insights, and market analysis through the ECI Research Portal.
