The AI Advantage Loop
Four phases and ten questions, run as a loop rather than a plan — because the technology moves faster than any twelve-month roadmap survives.
One full turn of the Loop is the five-week Readiness & Roadmap Sprint. Staying on the cadence afterwards is what keeps the Decision Charter true as conditions change.
A plan assumes the ground stays still.
Most AI strategy work produces a document with a twelve-month horizon. By month four the model capability it assumed has changed, a vendor you evaluated has been acquired, and a competitor has shipped something the plan did not anticipate. The document is still on the shared drive. Nobody has reopened it.
The Loop is built for that. Four phases, run in order, then run again — with the fourth phase existing specifically to catch what changed and send you back to the first. One full turn is the five-week Readiness & Roadmap Sprint. Staying on the cadence afterwards is what keeps the decisions true.
The phase most companies skip is the one that returns them to the start. That is why their AI strategy is eighteen months old.
Orient
Questions 1–3What is AI actually doing to the economics of your industry, which of your advantages does it erode or amplify — and on what clock?
What happens
- Six to eight executive interviews, 45 minutes each, run individually — CEO, CFO, COO, sales, operations, IT. Separately, because the disagreements are the finding.
- A disruption analysis of your industry, built against your competitive landscape and your win/loss record.
- The Moat Stress-Test — two hours with your leadership team. Every advantage you claim gets sorted into eroding fast, under pressure, defensible if strengthened, or increasingly valuable, with a time-to-erosion estimate on the ones at risk.
What you leave holding
- A disruption thesis in one falsifiable sentence: AI collapses the cost of X in our industry, moving advantage from Y to Z.
- A moat map with erosion horizons — which advantage goes first, and roughly when.
- A read on where your customers' expectations are heading, not where they are.
Assess
Questions 4–6Which of your proprietary assets get more valuable when everyone has the same models, where does the money actually sit, and can this organization get from pilot to production?
What happens
- The Crown Jewels workshop — two hours. Every candidate asset goes through one test: why couldn't a two-person startup with the same model rebuild this in a year? No hard answer, no moat.
- Value-pool mapping across cost, growth, and new business models — with at least three growth or new-model candidates forced onto the list, so the portfolio does not quietly collapse into a cost program.
- A capability audit across data, talent, operating model and adoption, plus a pre-mortem on your leading use case and the post-mortems of any pilot that stalled.
What you leave holding
- A proprietary-asset inventory with the crown jewels named and their defenses written down.
- A sized value-pool map — the opportunities ranked by prize and by feasibility, not by enthusiasm.
- The three binding constraints on execution, stated plainly.
Commit
Questions 7–9What do you adopt, buy, or build; in what order; and inside what guardrails?
What happens
- Every candidate initiative routed through a build / buy / partner decision — is this a differentiator, do you need to own the asset, can you maintain it — with a loaded-cost estimate on anything tagged build.
- The Portfolio workshop — two hours. Initiatives sorted into no-regret moves, big bets, and options; bets resourced like bets; and a named first domino, the early win that produces the data, skills and credibility the larger bets need.
- A governance session that produces the bright lines — the we will never use AI to… list that has to survive the front page — plus decision rights and escalation paths.
What you leave holding
- A prioritized initiative portfolio with the build-buy-partner call on each.
- The sequenced roadmap logic — not just what, but in what order and why that order.
- A governance one-pager: light enough to keep speed, explicit enough to survive an incident.
Adapt
Question 10How does any of this stay true when model capability moves faster than your planning cycle?
What happens
- One leading and one lagging metric defined for every big bet — measured against your numbers, not an industry benchmark.
- A fixed review cadence with a named owner, and pre-committed triggers: the evidence that would make you change your mind, written down before you need it.
- The synthesis: every initiative sized into a cost band with its assumptions named, the roadmap phased into quarters, and an investment case that quantifies the cost of waiting two quarters against your own figures.
What you leave holding
- The operating cadence — metrics, triggers, rhythm, owner. This is what makes the roadmap a loop rather than a binder.
- The sized roadmap and investment case, pre-read with your sponsor so the read-out contains no surprises.
- The phase most companies skip, which is why their AI strategy is eighteen months old and nobody has noticed.
Ten questions, and where each one lands.
The diagnostic asks these in fifteen minutes. The sprint answers them at depth, against your evidence.
- 01
How clear is our thesis for how AI reshapes the economics of our industry?Orient
- 02
Do we know whether AI erodes or amplifies our sources of competitive advantage?Orient
- 03
Do we understand how AI is shifting what our customers expect and value?Orient
- 04
Have we identified the proprietary assets that become more valuable in an AI world?Assess
- 05
Have we mapped and sized where the AI value actually is?Assess
- 06
Are we actually capable of capturing the opportunity — from pilot to production?Assess
- 07
Do we have discipline about where we adopt, build, or partner?Commit
- 08
Do we manage a balanced, sequenced portfolio of AI moves?Commit
- 09
Have we defined our AI risk appetite and how we govern it?Commit
- 10
Do we have a living system to sense change, measure impact, and re-strategize?Adapt
Six artifacts, assembled into one board-ready document.
Plus the working files behind them. Your team owns all of it.
The roadmap
Initiatives phased by quarter, each with an owner, a cost band with its assumptions named, expected value, and dependencies. The first domino is explicit.
The portfolio
No-regret moves, big bets and options — with the build, buy or partner call and a loaded cost on each.
The right-to-win analysis
The disruption thesis, the moat map with erosion horizons, and the crown-jewel inventory with each asset's replication defense.
The governance guardrails
Risk appetite, bright lines, decision rights, escalation paths.
The investment case
Investment in bands, value by bucket, payback logic, and the quantified cost of a two-quarter delay.
The operating cadence
The metrics, triggers, rhythm and owner that keep the roadmap alive. The strategic one — it is the Adapt phase made concrete.
A loop only works if something turns it.
After the sprint, that job needs an owner. A fractional Chief AI Officer is that owner — a standing seat at your leadership table, turning a rhythm deliberately small enough to survive a busy quarter and specific enough that skipping it is visible.
Weekly · signal sweep
A scan for anything that touches a live thesis. Most weeks it is short.
Monthly · bet review
Leading and lagging metrics against each big bet. Triggers checked, not admired.
Quarterly · re-strategize
One turn of the Loop at reduced depth. What changed, what it means, what moves.
The cadence is not something you buy separately. It is what the seat does — and it is how your Decision Charter stays true as conditions change. See how the engagement is structured →
Answering all ten takes fifteen minutes. The disagreement is the finding.
Run the diagnostic with your leadership team — separately, then compare. Where your CFO and your head of operations answer the moat question differently is where the sprint starts.