An agent we are building to keep your strategy true
A prototype is built and the MVP with design partners is in build. It aims to flag the claims that stop being true when the evidence changes. A human approves every change.
Strategy you can trace, kept true by an agent we are building.
A human approves every change. Consulting is available now, delivered by people.
- Prototype: built
- MVP: in build
- General availability: later
Read the graph as text
- Enter mid-market lending, the decision and root claim. Flagged: it depends on a declared assumption that failed. Depends on Strength (S·1) and Credit gap (W·1).
- Strength (S·1), a claim. Traced to evidence: Open-data rule (PESTEL L·1), Data platform (I·1), Bank APIs (PESTEL T·1) and Buyer power (F·4).
- Credit gap (W·1), a claim. Flagged: it depends on Credit team.
- Credit team, a declared assumption, not evidence. Failed.
Reading the picture: a dashed outline and the word “flagged” mark claims that depend on the failed item; ✕ marks the failed item; A marks a declared assumption; a tick marks evidence. 2 claims flagged.
One method, two ways in
Pick the way that fits your timing.
Consulting, delivered by people
Start with the Problem Diagnostic: a fixed-fee engagement that ends in a one-sentence problem statement, a worksheet and graded evidence.
The problem
A strategy nobody can trace goes stale without anyone noticing.
When the board, an investor or your own team asks for the evidence, “we had a good feeling” stops being an acceptable answer.
The method
No output without a traceable input.
When an assumption fails, you can see exactly which claims depend on it.
- Enter mid-market lending, the decision and root claim. Flagged: it depends on a declared assumption that failed. Depends on Strength (S·1) and Credit gap (W·1).
- Strength (S·1), a claim. Traced to evidence: Open-data rule (PESTEL L·1), Data platform (I·1), Bank APIs (PESTEL T·1) and Buyer power (F·4).
- Credit gap (W·1), a claim. Flagged: it depends on Credit team.
- Credit team, a declared assumption, not evidence. Failed.
Reading the picture: a dashed outline and the word “flagged” mark claims that depend on the failed item; ✕ marks the failed item; A marks a declared assumption; a tick marks evidence. 2 claims flagged.
- 1 · TRACED
Every claim is drawn to its source
Each claim points to the evidence it was derived from, or is declared as an assumption.
- 2 · FAILS
One assumption turns out wrong
The declared assumption, Credit team, fails. It is marked failed.
- 3 · FLAGGED
Only what depends on it is flagged
Credit gap and the decision to enter mid-market lending are flagged: 2 claims. The rest still stands.
- Evidence
- Declared assumption
- Failed
- Flagged: depends on the failed item
Read the method in the white paper (opens the white paper in a new tab)
published · auditable · open to review. The white paper describes the method.
The agent
We are building an agent that aims to keep a strategy true as the evidence changes.
The analysis is the agent’s. The decision is yours.
It aims to do the heavy analysis work. A human approves every change.
Where it stands
- Prototype: built
- MVP: in build
- General availability: later
The prototype is built. The recorded run shows it on an illustrative company, with a seeded signal and a pre-written proposal.
The recording could not be loaded. Use the download link below.
Videos
Latest videos
Company teardowns on public evidence, and data cases that run the method step by step. Newest first.
-
Data caseSeven theories for one revenue drop. Only one survived the data.
Watch on YouTube (opens in a new tab) -
Teardown10 people, 1 AI agent: what has Vercel actually shown?
Watch on YouTube (opens in a new tab) -
TeardownCan anyone prove AI made Intercom money?
Watch on YouTube (opens in a new tab)
Worked example · public data, not a client
(opens in a new tab) Sample Problem Diagnostic: HubSpot
See a sample diagnostic (PDF, 3 pages)
Published method
(opens the white paper in a new tab) The white paper
published · auditable · open to review. It describes the method.
Article ·
(opens in a new tab) Three well-built products, three failures — and the strategy discipline I wish I'd had
Read the article
On the method
It's the first strategy material I've seen that treats strategy the way engineers treat code … Alex's series is built on one rule instead: no output without traceable input. Every claim cites the coded evidence it was derived from, so conclusions read like stack traces, and you can click any statement all the way back to a raw fact. … it converts strategy from a genre of persuasion into a discipline of evidence. A strategy that cannot show its trace is not a strategy. Rare work.
Who leads
Who is behind it
Aliaksei “Alex” Ivanouski
Fractional CTO & Founding Engineer
Creator of the Business Strategy as Code method. Started as a military engineer, then spent 10+ years building products across fintech, adtech, social platforms and large-scale backend systems.
- MBA with Distinction, Bayes (formerly Cass), City St George’s, University of London.
- Chief Technology Officer Programme, National University of Singapore.
- AI: Business Strategies and Application, UC Berkeley Haas.
FAQ
Questions before you get in touch
Is the agent live?
No. A prototype is built and the MVP with design partners is in build. General availability is later. The agent is not open to the public, and no date is promised.
The agent page lists what exists today and what is planned.
What does a design partner pay for?
A paid pilot, not a free trial. A fixed fee, agreed in writing before anything starts. This site shows no price.
In return, people run the method with you on a real strategy question of yours, you have direct access to the founder, and you help shape the agent as it is built. The design-partner page has the details and the application.
Is consulting delivered by the agent?
No. Consulting is delivered by people today, and the agent takes no part in current engagements.
The agent is being built on the same method. It aims to run that method inside a company's own organisation, and a human approves every change.
How does an engagement start?
With a call. You describe the problem; Alex confirms whether a diagnostic fits, and the scope and fixed fee are written down before any work begins.
Is the fee fixed?
The Problem Diagnostic and Strategy Sprint carry a fixed fee with scope written in advance. Other engagements run on a fixed fee or a monthly retainer, agreed on a call.
How is this different from hiring a large consultancy?
One named senior lead does the work personally, the scope is written down before it starts, and every output is traceable to its evidence and stays with your team.
A large firm suits a programme that needs a big team. This suits a leadership team that wants a senior person and an output it can check.
What is Business Strategy as Code?
Business Strategy as Code is a traceable method for building business strategy and digital-transformation strategy. Every finding gets a code, and every conclusion cites the codes it was derived from, so anyone can follow it back to the evidence. It brings the discipline of engineering to strategy work.
Next
Three ways to take part.
1 · Apply
Apply as a design partner: a paid pilot in which your company helps shape the agent.
Apply as design partner 2 · Stay informed
Join the waitlist to hear when the agent opens.
3 · Talk to someone
Bring the decision you cannot yet defend. A brief call is enough to tell whether a diagnostic fits.
Request a callLooking for the open cohort for individuals?
See Individuals