We review your AI decision from the buyer's side: whether the economics hold, where the risk is, and whether your team can run it. You leave with a clear verdict: proceed, adjust, or stop, in about 3 weeks*.
*10 to 15 business days from kickoff, subject to the agreed scope, access to information and key people, the contract, and availability. This is a reference, not a guaranteed deadline.*10–15 business-day reference. See terms
of AI initiatives delivered expected ROI over the previous two years.
IBM Canada · Jun 22, 2026 · Canadian n unpublishedof AI use cases in infrastructure and operations achieve full success and meet ROI expectations.
Gartner · Apr 7, 2026 · n=782of technology CxOs are accountable for AI systems they do not fully control.
IBM · Jun 8, 2026 · n=2,000An early review tests the assumptions before the first contract.
Three possible situations before committing budget to an AI decision.
Three initiatives compete for the same budget, and no one can explain with data which one should go first.
Every option is tested against the same criteria, including doing nothing. You know what to prioritize, what to reject, and why.
Every stakeholder is pursuing a different objective, and no one has written down what “success” means.
Agreed success criteria and a clear map of who decides, who executes, and who owns each risk.
The business case rests on vendor projections and a demo that did not use your data.
An independent economic model with sensitivities shows which assumption could change the conclusion before you sign.
A poorly framed decision does not fail on signing day. It fails six months later, with budget committed and the team tied up.
The method examines assumptions about success, data, and capability in sequence, informed by 2026 research.
57% of all respondents had at least one failure; the 38% causes apply to leaders who faced setbacks, with multiple responses allowed.
Gartner, I&O leaders · Apr 7, 2026 · n=782cite data quality as a barrier; 51% cite integration with legacy systems.
Deloitte Canada · Jul 28, 2026 · n=300
scaled AI across multiple business units or adopted an AI-first approach. Most had not reached that scale.
Gartner · Sep 1, 2026 · n=1,303
feel fully prepared to operate AI agents at scale.
IBM · Jun 8, 2026 · n=2,000
Here is the method on a sample decision: four options enter, two are eliminated on evidence, and one reaches your table with conditions.
Possible recommendations: proceed · adjust, meaning proceed with changes · stop.
The recommendation is not chosen first and justified later. An option advances only if it answers each question with evidence.
The decision actually being made and what success means.
Decision Charter & success criteria
Current state, stakeholders, evidence, and real constraints.
Decision rights map
Alternatives, economics, risks, and dependencies, including the counter-case.
Alternatives, economic model & risk
What to do, what to avoid, and under which conditions.
Executive Decision Memo & Readout
Decision conditions, owners, unresolved risks, and review checkpoints supported by the evidence.
Conditions and checkpoints; optional 90-day roadmap where the evidence supports it
Qualified 10–15-business-day delivery reference from kickoff for standard scope. If critical evidence arrives late, the delivery date is re-baselined; the evidence standard does not change.
Typically 2.5–6 hours of sponsor time over about 3 weeks, depending on which optional sessions and interviews the decision requires.
We make participation straightforward with an information checklist, interview guides, and decision templates.
Every complete Review produces the same seven deliverables. What varies is the depth of evidence your decision requires.
Which roles, skills, and responsibilities are missing before scale, and the order in which to address them.
Technology that works in a demo but cannot be sustained by the organization.
Discovering the risk after signing.
Paying for licenses, scope, or scale that does not change the outcome.
A sound decision that stalls because its conditions and owners were never made explicit.
A recommendation you cannot defend to the board.
Value opportunities are presented as evidence-based hypotheses or ranges, not as guaranteed savings.
Medium-high confidence3 conditions · Day 90 checkpoint
Adjust = proceed with changes. It is one of the three possible recommendations: proceed, adjust, or stop.
Fictional logistics operator with 800 employees. The committee had two weeks to renew and expand its AI customer service platform for three years, a USD 1.4 million commitment.
The vendor projected 40% fewer tickets. No one had measured the baseline or defined who owned the outcome.
Renew for 12 months instead of 36, with a pilot in one business unit and three measurable conditions before expansion.
Illustrative example created solely to show the deliverable format. The company and scenario are fictional; this is not a client result or a promised outcome. Every Review depends on its own evidence.
If you engage us for the Review, the contract protects your investment in three ways.
If the agreed deliverables are incomplete at close for reasons attributable to our service, we continue working without additional professional fees to complete them, for up to 10 additional business days.
Applies when your team provided the agreed information and access on time, fulfilled its responsibilities, and the scope remained unchanged. It covers the deliverables, not business outcomes.
Read the complete terms →50% at signing. The other 50% is due when you receive the Executive Decision Memo. No hourly billing and no surprise fees.
No vendor commissions, no success fee, and no implementation engagement to sell afterward. The recommendation does not depend on your buying a solution.
The conversation is free. If your decision does not need a Review, we will tell you.
I lead every Review as the sponsor's independent counterpart: I make assumptions explicit, challenge the alternatives, and translate technical and operating risks into a recommendation a committee can use.
I do not sell technology or implementation. That allows me to recommend stopping when the evidence points there.
Includes: 1 decision · up to 3 alternatives + doing nothing · up to 5 interviews · 1 entity · 1 economic model · up to 3 working sessions · 1 comment round · 1 clarification session.
The fee is set before work starts, using objective criteria. It never depends on the verdict or on what you end up buying.
Book the first call → Message Victor on WhatsApp about the investmentDoes your decision involve more alternatives, several countries or business units, specific regulation, or the board? Two levels with expanded scope exist, Strategic Decision Review and Enterprise Decision Assurance. The level is set before signing with objective criteria, not negotiation.
How complexity is defined →Our welcome message asks about your role, the decision, and the deadline. Do not send confidential documents. We reply on business days.
Write to Victor on WhatsAppA 30-minute first call, at no cost, to see whether your decision needs a review. If it does, we work with you for about 3 weeks* until you reach a clear verdict.