The full session, for leaders of SMB companies: why most adoption stalls, the stages that actually work, and a real firm that walked the whole path. 58 minutes, with the Q&A.
Not a list of topics. You follow one real company through the whole sequence, and place your own on it as you go.
Most companies begin at stage two or three — tools handed out, a pilot somewhere, an automation built where it was easiest — without the stage that makes any of it hold. That single misordering is where most AI programmes quietly go, and it is the first thing we take apart.
Shared language, honest expectations, and enough literacy that the whole company can hold a conversation about what this is. Almost nothing is automated yet, deliberately.
Workshops for every tier. The first two words people put on the board were "robot" and "fear".
AI meets real work — processes mapped, the first assistants built by the people who own the work. Hours start returning, and the middle of this stage is where it hurts most.
Process and context mapping, the first assistants — and the dip where the whole thing nearly got cancelled.
Improvement becomes a rhythm rather than a project. People find the problems and build the fix without being asked, and the company gets redesigned around what it can now do.
20+ agents, internal apps, an offer checker built in 48 hours after a lost tender — and a platform swap taken in stride.
Not a stage you aim at. It appears once stage three is genuinely running: people spend their time on discovery, design and direction, and the company starts making things that did not exist.
Ahead of them, as it is for almost everyone. This is where the sequence points — not a finish line you cross.
What has to be in place before what. Most of the value is in knowing which moves are premature, not in knowing which moves exist.
The stages are described plainly enough that you can place your own company on them while you listen, and see what that makes the next move.
One company, told honestly — including the part in the middle where the value dips and the whole thing nearly got cancelled.
Everything the hour covers is written out in the Guide — the same sequence, at reading pace, sendable to anyone in your company. And if you want to place your own company on the stages with Cezary, that conversation is open.
Fifteen years in enterprise sales and go-to-market, including advisory work at Gartner across 400+ executive sessions, and execution roles at Scandit, Zowie and Droppe.
The four stages, and the reasoning for their order, were designed before they ever met a company. They then spent a year being proven inside a real engineering firm, in a full AI transformation run end to end — and that company is the one we walk today.