75% of new code at Google is now written by AI
- Just last year — 25%.
- The same trajectory awaits every profession that works with a keyboard.
Before you adopt AI
Before you adopt AI — see what your team actually does. In 7 days.
In a week we capture plan-vs-fact per role and hand you a map of real processes. Then automation decisions are made on data — not on slide decks.
Founder — Forbes · $1M+ startup exit · author of AI Implementation Pro
For B2B teams of 30–500 people
Why AI rollouts fail
AI gets deployed on top of processes nobody can see
Top teams buy GPT-4o, Claude, Gemini, hire AI consultants — and six months later have a Slack channel full of ChatGPT memes and zero impact on the P&L.
No baseline
Nobody knows how much time the team actually spends on repetitive work — so it's impossible to measure where AI helped.
Automating the imaginary
The workflow on the diagram ≠ the workflow in real life. AI optimises a process that exists only in the founder's head.
Silent sabotage
People keep working the old way. The dashboard shows 80% adoption; real usage is 12%.
Before automating anything — you need to see how the work is being done today. Without that step, every AI investment is a guess.
What we do in 7 days
Capture primary data on how the team actually works
Not an audit. Not a slide deck. A 7-day data-collection sprint from every workplace in your format — and a full process map at the end.
- 01
Opening session (day 1, ~30 min with you)
We align on roles, key processes, expected outcomes. We set up the data-collection channel (Telegram bot or Slack — your choice).
- 02
Plan/fact capture (days 2-6, ~5 min/day per person)
Each person logs daily: what they planned in the morning, what they actually did, what got in the way. AI helps make it specific — it won't accept «worked on project X».
- 03
Process map + recommendations (day 7)
You receive: a role × process × actual-time matrix, top-10 repetitive tasks (automation candidates), list of blockers that came up three or more times.
What you get
A real process map — not slides, a working artefact
Sample output after 7 days of capture in a 40-person team. Greyed cells were previously invisible to the executive.
| Role | Plan (stated) | Fact (captured) | Gap / automation candidate |
|---|---|---|---|
| Sales manager | 60% client calls | 32% calls, 28% manual CRM updates | CRM auto-fill from email/calendar |
| Customer success | Ticket response | 41% time spent searching across 6 sources | AI knowledge base over Notion + Intercom |
| Marketing | Content + campaigns | 55% — copy alignment in Slack threads | Templates + AI-first draft |
| Operations | CEO reports | 3h/week — copy-paste from 4 dashboards into slides | Auto-summary from BI to MD/PDF |
From this table you immediately see what to automate first — and how many hours/week it returns.
How we do it
Methodology — borrowed from sports teams and dispatch rooms
The «Plan → Fact → Gap» principle is not our invention. It's worked for a century where the cost of error is high: aviation, ops, pro sports. We adapted it for knowledge teams.
Low barrier for the human
5 min/day, via a bot in a messenger they already have open. No new interfaces, no «log in to the system».
AI makes the data clean
When a person writes «worked on project» — AI asks back: «which step? how many hours? what blocked you?». Without that, 80% of the data would be junk.
Anonymisation for the team, aggregation for the executive
Each person sees only their own log. You see role-level aggregation. This is not surveillance — it's telemetry.
Zero integrations in week one
We don't touch your CRM/ERP/Jira. The way we collect data: people briefly write what they did — and AI asks follow-up questions until the answer is concrete. This is intentional: we want the delta between what the systems say and what people actually do with their hands.
Who this is for
It fits if…
✓ Good fit
B2B knowledge team, 30-500 people
Sales, Marketing, Customer Success, Ops, R&D, Product. Anything where the output is decisions and communication, not stamped widgets.
You've already thought about or tried AI adoption
And you have the feeling «something is off — we invested but nothing moved». This pilot will show you why.
Ready for 30 min opening session + 5 min/day per team member
For 7 days. Without that minimum time commitment from the team, the result won't be high-quality.
✗ Not a fit
Manufacturing / line workers
That requires different tooling (time sensors, MES, IoT). We work with knowledge work.
Team <10 people
At that size you already know who does what. The pilot won't surface new information.
Looking for an «AI magic wand»
We sell diagnostics, not a silver bullet. If you expect AI to replace people in a week — that's not what this is.
Who runs this
Yaroslav Maxymovych
Founder of AI Advisory Board. I'm not a blogger, not a theorist, not a coach and not a programmer — I have a leader's mindset and have never worked for a salary.
50+ startups, 10 of them taken to profit, 4 sold. First in Ukraine to sell a startup for $1M+. That means I'm deep in the business — on the side of owners and executives, not outside consultants.
My first AI project took me 500 hours to build. The second — only 30. This pilot is a synthesis of what I learned to do right (and what not to) across dozens of real rollouts.
- 50+ startups, 10 to profit, 4 sold
- First in Ukraine to sell a startup for $1M+
- 3,000 interviews, 300+ people managed
- PhD candidate — applied math + economics
- Forbes features: Auctioneer's Portfolio, Intellectual Owners
Pricing
Fixed price for the pilot
No subscriptions, no retainers. One payment — 7 days of work — your company's process map at the end.
7-day diagnostic
$1,500
one-time, VAT extra
- Opening session (30 min) with all stakeholders
- Telegram/Slack bot for plan/fact capture (we configure it)
- Daily AI processing with normalisation
- Process map (role × process × actual time) in Notion + PDF
- Top-10 automation candidates with impact estimate
- 1-hour result review and roadmap session
Why now
This isn't a marketing trend — it's already happening
85% of Ukrainians already use AI, but only 6% pay for it
- The perfect window to outrun your competitors.
88% of employees already use AI — but only 5% «transformationally»
- 88% use AI daily — mostly for basic tasks (search, short summaries).
- 12% report enough training to go beyond basic functions.
- Only 5% apply AI in a truly transformational way — fundamentally changing how they work.
EY Work Reimagined Survey, August 2025 (15,000 employees, 29 countries) ↗
What this means for you
The bottom wants it. The top wants it. The window is open.
The top wants it
CEOs and owners see how AI has already changed their competitors — and are looking for how to roll it out internally. Top-down pressure is there.
The bottom wants it
Employees are already using AI despite the lack of training — because it simply makes the job easier. Bottom-up readiness is there too.
This is a rare window: the market is ready, the team is ready, and the structures for adoption haven't been built yet. Whoever captures the data first — outruns everyone.
Frequent questions
FAQ
Yes. Each person sees only their own log; the executive sees role-level aggregation, not per-person. We tell the team explicitly that this is telemetry, not surveillance, and the goal is to optimise processes, not evaluate people. If you want a human performance review, that's a different product.
Ready to see what your team actually does?
I review applications personally within a working day. If your case isn't a fit, I'll say so directly — I won't try to sell.
Apply now