A framework by Farid Mheir
Leading with AI
A framework for examining what adopting AI changes in an organization. Four sets connect strategic direction, organizational capability, enabling choices and execution.
I reply personally. No obligation. No mailing list.
The useful question
What changes here, why does it matter, and what should someone decide or do?
Use the framework to follow a concrete change through the organization. If AI accelerates coding, for example, the consequences may include new product possibilities, a different team structure, new quality controls and a revised investment plan.
A team needs all four sets at once: the mindset, the skillset, the toolset, and the actionset. Each set contains three elements.
Mindset
Where could AI create value, and what has changed enough to require a strategic response?
Identify where AI could create value and which opportunities deserve attention.
Business strategy
What matters to the business? Which assumption might AI change? What outcome would justify action?
What AI changes
What becomes possible or practical? Which constraint moves? Who gains or loses an advantage?
Strategic use cases
What work is delegated? What goes in and comes out? Why prioritize this candidate over alternatives?
Typical output: Strategic direction and an initial use-case portfolio.
Skillset
What organizational capabilities, ownership and governance does this AI portfolio require?
Determine what the organization must be capable of doing.
Organizational readiness
What must people learn or change? Who lacks capacity or confidence? Which gap must close before proceeding?
Operating model
Who owns the outcome? What may AI execute? Who accepts results, resolves exceptions and maintains the process?
Governance
What is permitted? Who remains accountable? When must the system stop, wait or escalate?
Typical output: Readiness assessment, capability gaps, decision rights and governance requirements.
Toolset
What data, solution, deployment, security and lifecycle choices are required?
Make informed choices about how a use case can work.
Solutions and acquisition
What pattern fits? What should we own, rent or buy? How could we switch supplier or recover our work?
Data
What information is required? Is it usable and authorized? Who keeps it current? Where do corrections go?
Security and operational fit
How does it fail? What can it access or change? How will we detect, contain and recover from problems?
Typical output: Data and technology conditions, sourcing questions, controls and operating requirements.
Actionset
What should be funded, tested, scaled, paused or stopped over the next 12 to 24 months?
Turn the direction into funded, measurable action and learning.
Business case and lifecycle economics
What outcome creates value? What changes relative to the baseline? What does success cost over time?
Experiments and decision gates
What assumption are we testing? Against what baseline? What result changes our decision?
Roadmap and management system
What happens first? Who owns it? What evidence will change the sequence, funding or scope?
Typical output: Business case, experiment design, roadmap and next actions.
The sets interact
A useful idea can expose a capability gap. A data constraint can change the use case. An experiment can change the business case. Return to earlier decisions as evidence improves.
Use the four sets to examine the same change from several angles. Not every use case has an equally strong implication in every element.
The course
Leading with AI is also the name of the course I lead at the McGill Executive Institute, built around these four sets. It is for leaders, not engineers. Participants learn to evaluate opportunities, ask better questions of their teams and vendors, and connect AI decisions to business value.
The book
My book AI Mindsets is a related source, but its thirty-one mindsets are not the twelve elements of this framework. The book is about how a person thinks, works and leads with AI. The framework is about what an organization must decide and be able to do.
Leading with AI and the Fog Index
Leading with AI looks inward, at your organization. The AI Fog Index looks outward: it helps identify how much uncertainty remains around a use case and what could reduce it. It does not replace judgment.
Together they are the two frameworks of my own that anchor the AI work I do through FMCS. See the consulting work →