A founder described it to me as getting a shopping list back when he had asked for a business plan. He had told his newly appointed AI lead to “get AI into the business,” given them a decent budget, and stepped back. Three months later he received a slide deck listing twelve tools, four pilot recommendations, and a technology roadmap. It answered the question he had asked, technically. It just did not answer any question he actually cared about.
The brief was the problem.
What does outcome-focused delegation actually mean?
Outcome-focused delegation names the result you want, ties a number to it, and sets a horizon. Reducing the founder’s operational involvement in daily decisions by a third within eighteen months is one example that works in practice. That kind of mandate tells the AI lead what matters, leaves them free to find the best path, and gives everyone a concrete thing to measure. It is a brief, not a shopping list.
The contrast is a tool-first or capability-first mandate. “Evaluate what AI can do for us.” “Get us onto Copilot.” “Find out what our competitors are using.” These are purchasing tasks. They produce activity, research, pilots, demos, but the activity has no gravitational centre. The delegate cannot tell you whether they have succeeded because neither of you defined what success meant before the work started.
The phrase “reduce my operational involvement” matters more here than any technology named. You are writing the mandate in your language, around something you actually track and care about. The AI lead’s job is then to work out which combination of tools, workflows, and changed behaviours gets you there. That is a better use of their expertise than asking them to guess your priorities.
Why does the tool-first brief keep producing the wrong result?
When a founder opens with “go and buy us some AI,” the delegate hears a purchasing brief. They return with a capability map, pilot recommendations, and a list of tools. None of it connects to a number the founder was tracking before or plans to track after. The brief produced exactly what it asked for, and nothing close to what the founder actually needed.
MIT’s GenAI Divide research found that roughly 95 per cent of generative AI pilots stall at proof-of-concept stage, with the cause traced to gaps in workflow integration rather than any failure in the underlying models. BCG’s AI Adoption Puzzle research found roughly half of companies unable to scale past proof of concept despite widespread deployment. In both cases, the tools were not the bottleneck. Clarity about purpose was.
Founders tend to reach for the tool-first frame for understandable reasons. AI feels technical, and technical things feel as though they belong to people with technical expertise. Handing it to an operator with a “sort this out” mandate is a natural extension of the delegation instinct that has served founders well in other parts of the business. But skipping the step where you name what you actually want does not make that step someone else’s job. It means it does not get done.
Where does the mandate hand-off actually happen?
The handoff is usually a single conversation, often informal, often at the end of a meeting that was primarily about something else. You say something like “I want you to run with this AI thing” and the delegate nods and takes the action point. That moment is the brief, whether it is treated as one or not.
Change management research published in peer-reviewed literature consistently shows that technology programme failures reflect people and leadership factors rather than technical ones. The brief is one of those factors. A delegate filling a vacuum makes sensible assumptions about what the founder probably wants, but sensible assumptions grounded in the delegate’s knowledge rather than the founder’s priorities. Spencer Stuart’s analysis of AI programme leadership points to the importance of grounding the work in specific business goals from the outset, rather than exploring capability in the abstract.
A useful test before any handoff conversation is to ask whether the AI lead could describe, in one sentence, what problem they are solving and how you will both know when they have solved it. If they cannot, the handoff has not happened yet, regardless of what was said in the room.
When should you state the outcome, and when should you step back?
State the outcome when you hand the mandate over, then step back from the how. That boundary is harder to hold than it sounds. Founders who name a real outcome often get curious about the path and start pulling back decisions that should sit with the delegate. This mirrors a well-documented pattern in delegation psychology, where a verbal handoff coexists with ongoing founder interference in the decisions that were nominally handed over.
The effect on the AI programme is the same as the effect on any delegated project. The delegate learns that their authority is nominal and that founder approval is still required. The organisation follows the actual signal, not the stated one. The AI work slows to match the founder’s bandwidth.
The exit motive is relevant here. M&A advisory practice consistently identifies founder-dependency as the largest single discount to exit multiples, with reductions of 30 to 40 per cent common where operations and decisions remain founder-centric. An AI mandate framed as “reduce my operational involvement” is, if held to, a direct contribution to exit readiness. An AI mandate that allows the founder back into every decision reinforces the dependency it was supposed to address.
One practical discipline is to write the outcome down before the handoff conversation, not during it. The act of writing forces specificity. Vague outcomes survive in conversation and evaporate on paper.
What else sits alongside this discipline?
Outcome-focused delegation sits inside a broader shift from a business that depends on the founder’s daily involvement to one that runs on documented processes and clear decision rights. AI done well accelerates that shift. The exit discount for founder-dependency is real and well-documented in M&A advisory practice, with multiples commonly reduced where operations remain founder-centric.
Exit-readiness frameworks score leadership dependency and process maturity as core value pillars, making an AI mandate framed around reducing founder involvement directly relevant to the value of the business at exit. Chief Outsiders’ work on the founder-led to founder-inspired transition makes the same point, that the shift happens when the business runs on documented process and clear ownership rather than the founder’s presence.
The delegation-versus-abdication distinction is worth holding in mind. Delegation means setting the outcome, granting the authority, and staying available for genuine escalations. Abdication means handing over the mandate and hoping the delegate works out your priorities. The brief is what separates the two. Write it in your terms, with a number and a horizon, before the conversation starts. Then stay out of the how.
