The program
Don't want to read all of this? Point your AI assistant at this page and ask it your questions.
Fetch https://ai.rimaslogic.pl/programme.md?v=TIMESTAMP now and read the full file. Do not use any cache, memory, or previous version. The file you just fetched is the only source of truth. Answer my questions about the training program from that file only. If something is not in the file, say so instead of guessing. If you cannot open links, tell me and ask me to paste the text. Start by asking what I want to know.
Some assistants cannot open links and rarely say so. If the answers sound made up, use "Copy this page" and paste the whole text in instead.
AI and automation training for a management team, built on your own work. Nothing in it runs on generic examples: the sessions are built from what you and your colleagues actually deal with, collected before the program starts.
How it runs
Five phases. The audience gets smaller as the technical depth grows: everything starts with the whole group, moves into small groups by area, and ends one on one.
Every phase has the same shape: who is in the room, what the session covers, what you walk out with, and what it asks of you. This is the initial plan, and it gets updated and adjusted based on what the room needs.
Phase 1 of 5, week 1, 60 to 90 minutes, everyone
Introduction
What happens
The program is explained against what was collected from the group before it started, not against a generic deck. You hear how many people are at which point (counts only, never names), what the most common pattern in the answers was, and how the program is built on it. Nobody builds anything yet.
What the session covers
- Where this came from: What was collected, how it was reduced, and the headline finding the program is built around.
- The structure: The five phases, the time each one takes, and your own path through them.
- The approach: Use case first, theory second: theory shows up when a build needs it. The boundary: the model drafts and explains, the filtering and the arithmetic stay in the system that owns the data. The router: every collected item goes down one of four routes (gather data, chase people, write it up, or stop and fix the data first).
- The stack, and who does what: The two layers: what you do yourself with an AI tool, and what gets built for you with proper tooling. Which tools sit in each layer and the data rules that apply.
- The time cost: You invest time in the exercises and the learning. That is what it costs you, not a payment for the course. If you cannot put that time in, say so in week one rather than week six.
You leave with
The readiness check is the one that matters: ten minutes, then read your personal summary. Anything else you leave with depends on what the session covers and what we actually do in the room.
What it asks of you
Nothing before the session. Ten minutes after it for the check.
Phase 2 of 5, week 3, 60 to 90 minutes, everyone
Common ground
What happens
The vocabulary the rest of the program runs on, learned by using each term on the document you brought, not by hearing it defined. Every person sets up their first project live and runs one real task inside it before the session ends. This is the program's first small build.
What the session covers
- Context: What the model sees is all it knows. Paste a document in and watch the answer change; why long conversations drift.
- Projects and instructions: Create one for your own area and write three lines: who you are, what this project is for, how you want output.
- Project knowledge: Upload the document you brought and see the difference it makes.
- Skills: A saved way of doing a recurring thing. The answer to "I wish it just followed my rules".
- Connectors: The tool reaching into mail, calendar or your CRM instead of you exporting. Shown once, with the data rule attached.
- Agents and automation platforms: The one-sentence versions: something that runs without you watching, and something that runs on a schedule across systems. The "built for you" layer from phase 1.
- Use it: Draft this week's version of the document you brought, inside your new project, with the facilitator circulating.
You leave with
A working project of your own, used once on a real task. Your first artifact, by the second session.
What it asks of you
One real recurring document brought in (a report, a summary, an update, the rawer the better) and a working company account. Between sessions: run one more real task through the project.
Phase 3 of 5, week 5, 60 to 90 minutes, everyone
Challenges
What happens
The group sees its own collected material organised: what the challenges actually are, which kind of solution fits which, and what each person's first build will be. Individual levels and answers stay private; the group sees items and routes.
What the session covers
- Check-in on the projects: What worked, what drifted. Failures are material here, not embarrassments.
- The group's challenges: The collected items grouped into the few patterns they form, presented as the group's own self-portrait.
- The router, run live: Each item gets its route said out loud: gathering data goes to an integration, chasing people goes to a workflow, writing up goes to the model. And the honest fourth branches: stop doing it, document it first, or start recording the data that does not exist yet.
- The line, applied: Per item, what you do yourself and what gets built for you, so nobody expects to hand-build an integration and nobody waits for something they could do tomorrow.
- Focus groups form: Who is in which group, what each group builds first, and what that build serves.
You leave with
One sentence you can say out loud: my first build, its route, my focus group, and what it needs from me before the group session. Plus your focus-group pre-work, written down with the date against it.
What it asks of you
One real task run through your project before the session. Nothing else to prepare; your items are already collected.
Phase 4 of 5, from week 7, every second week, 60 to 90 minutes, your area
Focus groups
What happens
The build sessions. Groups are formed by shared systems and overlapping deliverables, not by org chart. Everyone works in parallel on their own deliverable with the facilitator circulating; there are no demos. The first version of your build exists within about two weeks of the first group session.
What the session covers
- The boundary, in full: The failure story: accuracy degrades as data grows and rules stack, trust goes, hand-checking eats the saving. Then the rule: the system does the filtering and the arithmetic, the model writes over a table that is already right.
- Anatomy of an automation: Trigger, actions, conditions, data. Enough to read and specify one, even if someone else builds it.
- The build: Fifty minutes of parallel work on your own item: a saved export into one correct table into a model-drafted narrative, or a first workflow end to end with a real item through it.
- Commitments: One per person, written, with a check-in between sessions.
You leave with
Version one of your own build, running on your own task, and a recorded baseline of how long that task took you before.
What it asks of you
Access to your system verified and one real export pulled before the first group session. Ten minutes per task to record the baseline. Roughly an hour between sessions on your committed action.
Phase 5 of 5, from week 7, scheduled from the check-ins, as needed, you
Individual sessions
What happens
Working sessions on the technical detail a group cannot absorb. The between-session check-ins decide who needs one and on what. Not everyone will; the group sessions are built so that every person's first artifact exists without one.
What the session covers
- Restructuring what you already built: Standing rules, skills, recognising when a project has started anchoring on old material.
- Defining the data: When an item turned out to need a signal or a field that does not exist yet: what to record, where, and whether it is worth it.
- When the tooling fights back: Access, workflow platforms, integrations that refuse to behave.
- The day-30 re-measure: Thirty days after your build goes live, the baseline is measured again and the result reported back to you.
You leave with
The specific blocker removed, and, at day 30, a number for what actually changed.
What it asks of you
Only the time of the session itself. Scheduled when a check-in shows you need it, not by default.
Across the program: phases 1 to 3 are one session each in weeks 1, 3 and 5. Focus groups build from week 7. The first version of a build exists around week 9, and the day-30 re-measure lands about a month after that. Sessions run 60 to 90 minutes, every second week, six to eight in total. The fortnightly gap is deliberate: time to take the committed action, and time to adapt the program if needed.
What it asks of you
Stated precisely, because time is the honest constraint for everyone:
- The session itself: 60 to 90 minutes.
- Between sessions: one committed action, roughly an hour, with a check-in before the next session.
- Before your builds: about ten minutes per task to record how long it takes you today. That number is how the program proves what changed, thirty days after your build goes live.
- Once, before the first session: the ten-minute readiness check.
What you do yourself, and what gets built for you
The line is drawn on day one so nobody is surprised later.
You, yourself: working with an AI tool on your own material: drafting, summarising, standing context, model-written narratives over data you already have. The training makes you fluent at this layer, and the tool guides on this site are its reference.
Built for you: anything that must run on a schedule without a human, cross more than one system, or be trusted with arithmetic. That is proper tooling. By default it gets built by whoever runs your program, with you defining what "done" means. If you want to build it yourself, you can. Where something you need sits far above what the training covers, it gets built for you and your training targets something else. That is by design.
What you leave with
Every participant finishes the program with at least one working artifact of their own. Not a certificate and not notes: a thing that runs, on a task you named.
Privacy
Your individual answers and results stay between you and whoever runs the program. The group only ever sees distributions: how many people are at which point, never who. The check asks about things that didn't work, and that only gets honest answers if it stays private. It does.
Ask your AI about this page instead
Before the first session
- Log in on the check page with the access details you were given. Your personal summary is waiting there.
- Run the readiness check: ten minutes, no preparation, "I haven't tried it" is a complete answer.
- That's all. The first session asks nothing else of you.
Your read on this
Two minutes. No names, and the group never sees the answers.
Received. Thank you.