Start with the organization, not the technology
The first session focuses on how your company works rather than on what AI could do. That ordering is deliberate because it reveals whether the organization can absorb the change before you fund the technology.
Book an assessmentLeave with a one-page picture of organizational readiness and one committed action with an owner and a date.
- 05Foundation questions
- 05Evidence questions
- 01Committed action
Why the first session starts differently
AI programs often stall after the technology works. These figures point to organizational alignment as the more useful place to begin.
- 70%
of AI transformations fail to deliver the value expected. Culture, not technology, is the most cited cause.
MIT
- 95%
of AI pilots never reach production. The blocker is almost never model capability.
BCG and MIT Sloan
- 3–5×
more value can come from the same AI budget when the organization is strategically aligned rather than opportunistic.
McKinsey Global AI Survey
The pattern we are trying to avoid
Discovery sessions that start with technology produce a use-case list and a roadmap. Both are easy to approve and easy to fund. Twelve months later, the tooling is in place and the pilots have been demonstrated, but almost nothing has changed in how the business actually runs because nobody examined whether the organization could absorb it.
The spend was real. The leverage was not. This failure is predictable in advance, which is the reason to spend the first session differently.
- Use cases approved
- Roadmap funded
- Tooling installed
- Pilots demonstrated
- Work unchanged
What the 7S framework does
The McKinsey 7S model makes one claim: seven parts of an organization must reinforce each other, and changing one puts stress on the other six. It is not a maturity score or a roadmap. It finds the parts that will resist before you commit budget.
- 7SAlignment
Strategy
What customers pay us for
Structure
Who owns what and who decides
Systems
How work and decisions run today
Skills
What our people can actually do
Staff
Who is in which seat
Style
How leaders behave, not what they say
Shared values
What we do when the rules run out
The first three are visible and budgetable. The last four decide whether the first three survive.
Why it fits AI specifically
Adopting AI properly is not a technology installation. It is an operating-model change. It moves who decides, what gets measured, which work disappears, and what a good employee looks like. That means it touches all seven elements at once. A technology-first discovery only looks at one of them.
- Who decides
- What gets measured
- Which work disappears
- What good work looks like
What the discovery changes
The difference is not a longer deliverable. It is better evidence about what should happen before delivery begins.
Technology-first discovery gives you
- A ranked list of use cases
- A vendor and platform shortlist
- A phased delivery roadmap
- Confidence, with no information about what will block you
7S-first discovery gives you
- A map of where the organization will resist and why
- The two or three constraints that would sink any roadmap
- A defensible sequence: fix the constraint, then fund the build
- A shared, specific answer among your leadership team
What happens in the session
A focused 60-minute leadership session moves from assumptions to evidence, then ends with one commitment.
Score the foundation
Five foundation questions are scored openly by the leadership team. Disagreement is the first finding.
Point to evidence
Five evidence questions require a record, such as a decision, promotion, or log, rather than an opinion.
Name the constraint
Read the result, name the binding constraint, and agree one action with an owner and a date.
You leave with a one-page picture of where you are and a single committed action, not a strategy document.
What this will not do
7S diagnoses alignment. It does not prioritize, sequence, or estimate. It will not tell you which processes to rebuild or what the return will be. That operating-model work follows, and it becomes faster and cheaper once the constraints are known.
If your leadership team is already aligned and every evidence question comes back yes, we skip this step and go straight to the redesign.
References
Framework: McKinsey 7S (Waterman, Peters, and Phillips).
Operating-model reference: World Economic Forum and Kearney, “The AI-First Operating System”, June 2026.
