Most organizations don’t fail at AI because the technology is wrong. They fail because the structure is.
MIT Sloan’s three-lens management framework positions the strategic design lens as the organizational architect — the layer of management responsible for ensuring that processes, reporting lines and performance metrics actually serve the company’s goals. When AI enters the picture, this first lens gets tested in ways that traditional restructuring never anticipated. The org chart was built for human workflows. AI doesn’t care about your org chart.
AI adoption is less a technology initiative and more a structural interrogation. It forces leaders to ask: Does our current design support what we’re trying to become — and do we have the data to prove it?
Here are three ways companies operating at the top of their field are answering that question through the strategic lens.
Redesign Roles Before Deploying Tools
When Microsoft CEO Satya Nadella announced the company’s return to hiring in late 2025, he was explicit about what had changed: “We will grow our headcount, but the headcount we grow will grow with a lot more leverage than we had pre-AI.” That single statement captured the structural logic Microsoft had spent the better part of two years executing. The company didn’t simply cut roles to reduce costs — it systematically eliminated positions that AI had absorbed and rebuilt its hiring architecture around functions that AI amplifies rather than replaces: AI infrastructure, safety research and developer tooling. The org chart followed the strategy, not the other way around.
The lesson is unambiguous: AI doesn’t slot into existing roles. It demands that organizations ask what those roles exist to accomplish in the first place.
Align KPIs to AI-Driven Outcomes — Not Legacy Metrics
JPMorgan Chase’s Contract Intelligence platform — COiN — reviews commercial loan agreements in seconds, eliminating work that previously consumed 360,000 hours of attorney and loan officer time every year. But the more revealing structural challenge wasn’t the technology itself. It was that every performance metric evaluating those legal and operations teams had been built around volume of manual reviews completed. AI rendered those metrics obsolete overnight.
JPMorgan restructured its performance framework to measure accuracy rates, exception handling and process improvement contributions. Structure followed strategy. Metrics followed capability. That sequencing matters enormously — and most organizations get it exactly backwards.
Build Cross-Functional AI Governance Into the Hierarchy
One of the most persistent structural failures in AI adoption is ownership ambiguity — IT believes it belongs to the business unit and the business unit believes it belongs to IT. Cleveland Clinic tackled this head-on by establishing interdisciplinary teams that meet monthly to review AI performance and make real-time adjustments — with representation drawn from frontline clinical staff, IT, compliance, patient experience and finance. No single department owns AI. Accountability is shared, structured and visible across the hierarchy.
The Architecture Has to Come First
AI will not fix a misaligned structure — if anything, it accelerates the dysfunction already baked into one. The organizations gaining meaningful ground are not asking “Where can we apply AI?” They are asking “Are we structurally built for where AI is taking us?” That question belongs squarely in Lens 1 — and answering it with discipline is where the journey toward full AI adoption must begin.
