The method
Two passes, one system.
Structure tells you what the organization is. Flow tells you how work moves through it. Together they tell you what to change first.
We call the methodology the Helix: two strands twisted together, and four artifacts that compose it. A function under AI cannot be read through one lens, so we read it along two.
Two strands, four artifacts. Where they cross is where a decision can be made.
Structure
How the function is organized to deploy capacity. The function decomposes into teams, teams into roles. This tells you which teams are still viable as AI changes the volume and nature of the work, and where capability investment should go.
Flow
How work actually gets produced. Critical workflows, often crossing several teams. This tells you what the future-state workflow looks like, where humans hold decision authority, and where exceptions route.
The two cuts cross-check each other. If the structural view says a team is core but its workflow evidence shows the work is mostly AI-handled, that contradiction surfaces instead of being averaged away. That cross-check is the reason both cuts exist.
A workflow crosses teams. A team sits across workflows. The role is the square where both meet.
The four blueprints
What we actually produce.
Four structured artifacts, produced in a defined order. Together they are the evidence every decision traces back to.
The four artifacts as they are actually built. Values shown are from a sample engagement.
FunctionDNA
The function as a portfolio of teams: cross-team value flows, the operating model, where capability sits, and a viability verdict for each team.
TeamDNA
The team as a work system: what it owns, how work moves through it, what it can deliver, and where AI changes the load.
FlowDNA
One workflow end to end: every step, current and future state side by side, where the hours go, who decides what, and the ROI band with its assumptions.
RoleDNA
A role at task grain: outcomes owned, decisions made, every task scored, the capability profile that follows. The proof drawer behind a number, not the pitch.
The scale
Every task is scored on the Stanford Human Agency Scale.
The scale is Stanford's, from the WORKBank research — not ours. What is ours is applying it at task grain across four levels of an organization and driving decisions with it. Higher always means more human agency, not less.
Nor human nor machine
The more AI carries at the low end, the more valuable the fully human work becomes. That is what makes the scale a design tool rather than a scoreboard: it shows you which work to move, and which work to invest in.
In practice
What actually changes, step by step.
This is the output the whole method exists to produce: a workflow as it runs today, and the same workflow after the redesign, with every step scored. Press run and watch where AI takes the load — and where it does not.
Four workflows from our sample engagements. The score is the Stanford Human Agency Scale — the fuller the mark, the more of the step is human.
Order of work
The structural picture is finished last, not first.
The first pass starts with a scaffold of the function from leadership interviews — a starting picture, not a verdict. Then the workflow and team evidence is produced in parallel, and it revises the scaffold. Only then are the team verdicts and the future operating model locked.
The second pass produces role-level detail, and only for roles that sit on teams confirmed viable or transitional and that operate a workflow we analyzed. That rule keeps the analysis pointed at work that will still exist.
The matrix view of structure crossed with workflow is not novel to us — it is established in operations practice, and the recent consulting consensus has converged on it. Our contribution is turning it into structured artifacts that hold up under challenge, re-use across functions, and stay current as AI capability moves.