“Should we approve this AI proposal?”
Pressure-test the claimed benefit, full cost, adoption requirements, and alternative uses of the budget.
01 / Independent decision review
Decision Factory
An independent review of one consequential product or AI decision. A clear recommendation, its supporting evidence, and the assumptions that could overturn it.
What you take into the meeting
A decision you can
explain and defend.
Useful when your team already has a recommendation—and the cost of getting it wrong deserves another look.
The call
Proceed, change the scope, gather specific evidence, or stop. With the commercial reason up front.
The reasoning
The relevant evidence, alternatives, unit economics, and risks. Including the case for using what you already have.
What would change our mind
The uncertain assumptions, the evidence that would resolve them, and the owner of the next action.
A focused second opinion
“Do we build this capability ourselves?”
Compare the specific need against existing products, internal capacity, integration work, and ownership costs.
“Is this the right next bet?”
Examine the proposed product investment against your goal, current evidence, constraints, and opportunity cost.
How it works
Small input.
Substantial scrutiny.
Bring the recommendation
Share the decision, deadline, existing proposal or brief, and the evidence your team already uses. No new presentation required.
We confirm the fit
We agree one decision, a bounded evidence pack, the output format, and a delivery date. If the question needs original research or a larger investigation, we say so before you buy.
Receive the decision brief
A one-page executive recommendation, a compact evidence and assumptions appendix, and a model when economics are central. We agree memo or presentation format before starting. A 30-minute discussion and one factual correction pass are included.
Inspect the work before you buy
The spreadsheet says
something different.
A fictional support-automation proposal looks attractive until eligibility, review effort, and first-year costs enter the model. Explore the assumptions and see what would change the call.
Illustrative analysis. Not a client result or an ROI promise.
Choose the right level of help
A good fit when the decision is ready to examine.
Bring us this
- A named decision owner and a real deadline.
- An existing recommendation and usable evidence.
- A commitment meaningful enough to justify an independent review.
Use a different scope for this
- Original customer research or a full technical audit.
- Reconstructing a business case from scattered, unverified data.
- Ongoing ownership of implementation. See Embedded Delivery ↗
Questions before you commit
A few useful details.
Why not ask our team or an AI tool?
You should. This review earns its place when an experienced outside view tests the assumptions, alternatives, and business consequences that an internal draft may miss. Inspect the example and judge whether that is valuable for your decision.
What information do you need?
A decision statement, the existing proposal or brief, your goal, constraints, and the evidence behind the recommendation. The standard scope is one decision and a compact evidence pack, normally up to 10 pages plus one relevant data extract. We confirm what is usable during qualification.
How quickly will we receive it?
Tell us the date of the decision. We confirm a delivery window after seeing the inputs and confirming delivery availability, before you pay. Tight deadlines work only when the evidence and scope are ready.
Is this AI-generated advice?
We use AI to accelerate research, analysis, and drafting. An experienced reviewer is accountable for the recommendation, checks the material claims, and makes uncertainty explicit. The value is the judgment you can inspect.
Can we work across countries and time zones?
Yes, subject to an agreed scope and availability. Start in English or Swedish. Prices are listed in USD; tell us about required invoice terms and meeting overlap. Working arrangements are confirmed before you commit.
How do you handle confidential inputs?
Start with a description or suitable redacted material. Before using project inputs, we agree the information, tools and providers, access, retention, and any processing restrictions. Your required approval process remains part of the engagement.
What if the evidence does not support an answer?
Then the recommendation may be to defer, narrow the commitment, or collect a particular missing fact. We will not manufacture certainty to complete a deck.
What are you about to commit to?
Tell us the decision and when you need to make it.