toni lacki_

The AI Operating Manual · presale

Build your own system for working with AI and get hours back every week.

Reports, documents and analysis in half the time, with a result you check in 60 seconds. You start with the task already sitting on your desk.

Full version launches: 10 November 2026

--days
--hrs
--min
--sec
Buy in presale · $99

If within 30 days you do not speed up a single task, I refund the full payment. One email, no questions asked.

A lesson from The AI Operating Manual open on a laptop: a brief for a weekly report with context, task, format and a check, plus the result-checking step.
A task-library lesson on the training platform.

Toni Lacki · over 10 years teaching companies technology adoption. First edition, no participant reviews yet.

Sound familiar

Get back the time recurring work still takes.

You have AI. And the recurring work is still waiting. Below are example client situations. These are scenes describing the problem, not quotes or participant reviews.

Marketing

AI is already at the company. And I am still doing the work by hand.

I want to put materials together faster and show my team a concrete result. I do not know where to start or what I am allowed to feed into the tool.

You start on the practice files. You use a brief and check the answer against the source.

Owner of a service business

Meeting is over. The follow-up email to the client is waiting for tonight again.

I already pay for AI. I want to organize what we agreed on faster, but I need to know the answer did not add a promise I never made.

In the sample brief you set the format and the rule: every fact has to come from the notes. Gaps get flagged, and you compare the result against the source.

Manager

The team uses AI. I want to know what that actually changes about the work.

Watching people use a chat tool tells me little. I need one concrete task, a result, and a time record.

You compare the same recurring task before and after AI. You count preparation, checking and corrections.

Why now

Most people have a licence. Few have a way of working.

Almost everyone has access to AI today. A minority uses it regularly and in a way they can verify. That is the gap this training is built for. Figures from research, with date and source.

1 in 5

Only 19.95% of EU companies used AI in 2025. Denmark: 42%, Poland: 8.4%. Most teams and most of your competitors have not started yet.

Eurostat, Use of AI in enterprises, 2025 data

19%

Only 19% of employees use AI at work a few times a week or more, and 8% daily. The rest have a licence or curiosity.

Gallup, US workforce survey, June 2025

71%

71% of leaders would rather hire a less experienced candidate with AI skills than a more experienced one without them. 66% would not hire anyone without those skills.

Microsoft Work Trend Index, May 2024, 31,000 workers in 31 countries

+40% / −23%

Consultants using GPT-4 delivered 40% higher quality where AI fits the task, and 23% worse where it does not. That is why every task in the course ends with checking the result, not just generating it.

BCG and Harvard, experiment with 758 consultants, September 2023

None of these figures are results of participants of this training. They describe the market you are starting in.

Before / after · example from the Start module

Check whether AI saves your time. Even after corrections.

Compare the same recurring task. Log the full time for the manual version and the AI-assisted version. Include preparation, execution, checking and corrections.

Before · by hand70 min
After · with AI and checking40 min
Example from the Start module. An illustrative time-log entry, not a participant measurement or a promise about your own result. You measure your own result with a time log on your own task.

How it works

Catch the guess before an AI result reaches your client.

Before you use an AI result, compare it with the source. Every task links source material, a brief, execution and a check. Below is a sample brief for summarizing meeting notes.

An editorial example on fictional notes. This illustrates the method and the check, not a real AI conversation or a participant result.

Source material

The team needs to collect feedback on the proposal. No owner or deadline has been set.
  1. 01

    Context

    Give AI a point of reference.

    I am preparing a meeting-notes summary for my team.

  2. 02

    Task

    Name the work to be done.

    Pull out the decisions and action items recorded in the notes.

  3. 03

    Format

    Specify what you want back.

    Table: action, owner, deadline.

  4. 04

    Check

    Set how you will judge the result.

    Every row has to come from the notes. Flag gaps as "no data".

Correct result

An accurate mapping of the information

Correct meeting-notes summary
ActionOwnerDeadline
Collect feedback on the proposalno datano data

The gaps are visible. You know what to confirm before using the summary further.

Result with a guess

A deliberately wrong variant for comparison

Summary with an invented deadline
ActionOwnerDeadline
Collect feedback on the proposalno datatomorrow

Error to catch: the deadline "tomorrow" is a guess. The notes do not give a deadline.

Check

Remove the invented deadline. Keep "no data" and set a deadline before you use the summary further.

Execution

You go from brief to work.

Hand AI the practice file and the brief. Then compare the answer with the source material.

Material + briefResultCheck

The 60-second check

You catch mistakes before you hand off the work.

  1. Do the facts match?

    Check names, dates and decisions against the source.

  2. Did AI add anything?

    Find information that is not in the source material.

  3. Does the result meet the brief?

    Check format and criterion. Fix mismatches before you use the result.

In the exercise you run a quick 60-second check. This is not a promise about how long it takes to check every piece of work. If you find an error, add time for the correction and a second check.

From login to a checked result

Your first 90 minutes

You finish with a completed exercise and a first entry in your time log. You do not have to go through all 52 lessons. These steps make up your first pass through the Start module.

  1. 01

    Pick a task

    You log in to the platform and pick a task from the list of 32.

  2. 02

    Open the practice file

    You start on the included material, without reaching for company data.

  3. 03

    Build a brief

    Context, task, format and check. You fill in four parts.

  4. 04

    Run it in your tool

    You hand AI the file and the brief. You get a first version of the result.

  5. 05

    Check and log the time

    You run the 60-second check and log the time. You add corrections and a second check.

Which tool

Use the AI you already have.

  • ChatGPT
  • Claude
  • Gemini

You work with ChatGPT, Claude and Gemini, including free plans. The model map in module 1 compares more tools and tells you when it is worth paying. Available features and limits depend on the plan you pick. The course does not include AI tool subscriptions.

Who it is for / who it is not for

This is for you if you want to improve work that keeps coming back.

Yes, if:

  • You keep coming back to similar reports, documents or summaries and want to do them faster.
  • You already have AI, but you still start similar tasks from scratch.
  • You want to start on practice files and set the rules for using data from your job.
  • You need to check a result before you show it to a client, your manager or your team.

This scope will not match your need if:

  • You are looking to learn programming or build AI applications.
  • You want to pass AI answers along without comparing them to the source.
  • You need a community or daily emails to keep coming back to the material.
  • You want to learn only from finished video recordings. Recordings are in progress.

You pick a task by your own role

  • Analytics
  • HR
  • Marketing
  • Sales
  • Project management
  • Consulting
  • Team management

What you have after every module

You finish each stage with something to return to at work.

Start with the Start module. Then pick material for the task in front of you. You do not have to go through the whole library in order. You get 52 lessons across 9 modules. Connectors and MCP are optional.

9modules
52lessons
Module 0 · StartYou pick a task and learn to check the result.

You choose recurring work, go from source material to result, and log the time. Start has 7 lessons and 77 minutes of material. Set aside up to 90 minutes for onboarding.

Module 1 · Model mapYou match a tool to the task.

You use the model map to judge a choice against the requirements of your own work. The map compares more tools than ChatGPT, Claude and Gemini, and flags when it is worth paying.

Module 2 · Prices and limitsYou know what to check before you pick an AI plan.

You compare your task requirements against a tool price and limits table. The course price does not include AI subscriptions.

Module 3 · Task libraryYou pick a task from your own role.

You get 32 tasks with CSV, XLSX, DOCX and PDF files. Each one links source material, a brief, execution and a check. Tasks take about 15 minutes. The quick check takes 60 seconds; corrections and a second check take extra time.

Module 4 · Brief repositoryYou come back to a saved instruction.

Briefs are collected in a repository on GitHub. Public notes show the way of working. The private GitHub repository for members collects briefs with change history and updates in one place.

Module 5 · Data and conversationsYou prepare the conversation about using data with AI.

You start on practice files. Before you reach for data from work, you check company rules and agree the use with your employer or client. The course helps you prepare that conversation; it does not grant permission on their behalf.

Module 6 · Changes in AIYou judge what is new by your own work.

You learn to read changes in AI tools and check whether they help with the task you actually need to do.

Module 7 · Measurement and your resumeYou have a concrete result to describe.

You measure your work and prepare a resume entry based on a task you actually did. You can describe what you do with AI, how you check the result, and what time you log.

Module 8 · Connectors and MCP, optionalYou decide whether you need connections to other systems.

You learn about connectors and MCP, the way of linking AI to other tools and data. You can skip this module and keep working from files and briefs.

Less to keep up with

You come back for an instruction and get on with your task.

There is no community and no daily emails. You pick the material you need for your own task. You use briefs with context, task, format and a check.

A year of access with updates

Get moving faster on a task that will come back.

Your annual access includes:

  • You start faster

    A Start module that walks you through picking a task, writing a brief, and checking the result. 52 lessons across 9 modules in total.

  • You have something to practise on

    32 tasks by role with CSV, XLSX, DOCX and PDF files. You get material to practise on and a criterion for checking the result.

  • You are not rewriting instructions from scratch

    A brief repository on GitHub. You can come back to an instruction for similar work.

  • You come back when the work changes

    A year of access to the materials and updates.

  • A monthly Task clinic and the Short List

    60 recorded minutes every month: a walkthrough of the Short List of tool changes, one task done live, and answers to questions sent by email.

  • A certificate for work, not for watching

    AI Power User with a tools stamp. You earn it after the baseline test and the three artifacts from module 7, renewed with every year of membership.

Do the math with your own rate

The time you get back has your own value.

Do the math with your own rate: at $50 an hour, the course pays for itself after you recover about 2 hours, roughly one weekly report done faster for a month.

An AI tool subscription alone runs about $20 a month, whether you use 5% or 50% of what it can do.

This is an example for plugging in your own rate and your own AI subscription cost. It assumes 2 hours recovered on reports across a month in total; it is not derived from the 70 to 40 minute illustration above. Usage levels are an illustration, not a measurement of users or a plan limit.

Divide $99 by the value of your own working hour. Include time for learning, preparation, checking and corrections. The course helps you organize work with AI; it does not replace a subscription. If your employer pays for the tool, do not count their bill as your own expense; on a free plan, skip this cost.

Presale · annual access

$99 / year

Paid upfront for the year.

Your first result in 90 minutes. Then tasks of about 15 minutes each. 52 lessons, 32 ready-made tasks with files, a year of access with updates.

Full version launches 10 November 2026. Time to launch: --.

Buy in presale · $99

Presale until the launch on 10 November 2026. Buy today and open the text lessons with exercises right away; you get the full version with recordings on launch day. Your year of access starts at launch.

The price covers the course. It does not cover AI tool subscriptions.

Before launch: text lessons and exercises. Video recordings arrive 10 November 2026.

Access after payment

  • Email after paymentUsually within a few minutes you get an email with a link to set your password on the platform.
  • 12 months from purchaseUpdates to the materials are included for the whole access period.
  • Renewal by new purchaseNothing is charged automatically.
  • Receipt for expensesStripe checkout gives you an itemized receipt you can file as an expense.
  • Downloaded files stayYou can keep using them even after your platform access ends.
One-task guarantee

If within 30 days of purchase you do not have a recurring task you now do faster, request a full refund within that window.

See the guarantee terms

Founding package

Start with help, and keep your price.

Your brief, reviewed

Once a year, send one brief in reply to your welcome email. Within 7 days you get an answer: what to fix and why.

Closes the objection: I do not know if my instruction will get me the result I need.

Your price, frozen for good

Buy before 10 November 2026 for $99 and pay the same price at every renewal. Regular price from 2027: $129 a year.

Closes the objection: next year access might cost more.

One-task guarantee

One task, faster. Or the full payment comes back to you.

The full payment comes back

If within 30 days of purchase you have not sped up one recurring task, you get a full refund.

The files stay

The files you downloaded stay with you even after a refund.

One email

No questions asked. Write to info@tonilacki.com.

Request a refund within 30 days of purchase. I refund the full payment within 14 days of your request, the same way you paid.

info@tonilacki.com
Toni Lacki, author of The AI Operating Manual

Toni Lacki · author

You learn to translate technology into everyday work.

My name is Toni Lacki. For over 10 years I have taught companies how to adopt new technology. Today I roll out AI across departments of global companies. I also build my own course platform, Ludus.

3

client academies built from scratch

5,000+

people certified in 6 months in one of my earlier programmes

PLN 1,000,000+

in corporate course sales

In that programme: +18% product adoption and −22% support tickets. These are results from earlier work, not a promise about this course.

The AI Operating Manual is a first edition. I do not have participant reviews yet. The numbers above are from my earlier work.

See the way of working on a sample task

Before you decide

Check the details and choose the right access for you.

What do I get during presale?

After payment you get an email with access to the platform right away, and you open text lessons with practice files, briefs and check criteria. The full version with video recordings launches 10 November 2026. Your year of access starts at launch, so presale does not shorten your access. The price is the same in presale and after launch: $99 a year.

Do I need to know how to code?

No. You write briefs in plain language. You work with files. GitHub is used as a materials repository. The connectors and MCP module is optional.

How much time should I set aside?

Up to 90 minutes for onboarding. Start covers 7 lessons and 77 minutes of material. Tasks from the library take about 15 minutes. That is a practice-time estimate, not a promise about how long every piece of professional work will take. 30 days is your guarantee window from purchase.

Will I get a finished video course?

Video recordings are in progress. The offer includes the lessons, tasks and files for independent study. No date has been set for when the recordings will be added.

Do I have to use my company data?

No. Start with the included practice files. Before you use your own data, check your company rules and agree with your employer or client what you may feed into AI. The data module helps you prepare that conversation. Buying the course does not replace that permission.

What if I do not speed up a single task?

If within 30 days of purchase you do not have one recurring task you now do faster, email info@tonilacki.com within that window. I refund the full payment within 14 days, the same way you paid. The files you downloaded stay with you. One email, no questions asked.

Do I have to go through every module in order?

Start with the Start module. Then pick a task from your role and reach for the material you need for that work. You do not have to go through the whole library. You can skip connectors and MCP.

What does $99 a year cover?

A year of access with updates: 52 lessons across 9 modules, 32 tasks with files, and the brief repository. You pay upfront for the year. AI tool subscriptions are not included.

Is the 70-minutes-to-40-minutes example a participant result?

No. That is an illustrative example from the Start module showing how to log time. It is not a measurement of a participant, and it is not a promise about your own result. Count your own time together with preparation, execution, checking and corrections.

Are there participant reviews yet?

This is the first edition. I do not have participant reviews yet. The numbers in the author section are from my earlier work. On this page you can see a sample brief and the checks used on a result.

Does this make sense if I already use AI?

Yes, if you still start similar tasks from scratch or find it hard to check the answer. You use briefs, check criteria and time tracking. Having the tool alone does not tell you how to do your own task.

How is this different from just asking a chat tool a question?

You get connected materials to practise with: a source file, a brief, execution and a check. You can come back to briefs for similar work. You do not have to build the whole exercise yourself from a blank page.

Will checking every result take 60 seconds?

No. 60 seconds refers to the quick check in the exercise. At work, how much checking you need depends on the task and the source material. If you find an error, you need time to fix it and check again.

Which tool can I work with?

You work with ChatGPT, Claude and Gemini, including free plans. The model map in module 1 compares more tools and tells you when it is worth paying. Available features and limits depend on the plan you pick. The course does not include AI tool subscriptions. You check your company data rules in module 5. We do not assume every task fits every plan.

How do I get access, and what happens after a year?

After payment you get an email with a link to set your password on the platform, usually within a few minutes. Access lasts 12 months from purchase, updates are included. Renewal is a new purchase, nothing is charged automatically. Downloaded files stay with you after access ends.

Can I buy this for my team?

Email info@tonilacki.com. A team version is in preparation.

Is the training available in English?

Yes. This is the English edition: all 52 lessons, the 32-task library, the practice files and the members repository are in English. The Polish edition is a separate product on antonilacki.com.

Is there a certificate?

Yes, but earned by work, not by watching lessons. The AI Power User certificate with a tools stamp (month and year) is issued on the platform once you mark the baseline test from module 0 and the three module 7 artifacts as done: the resume entry, the one page for your manager and the manager's quarter plan. You can add it to your LinkedIn profile; when you renew access it gets a fresh stamp.

What is the monthly Task clinic?

Once a month, 60 minutes, recorded and available in the course: a walkthrough of the Short List of tool changes, one task from the library done live, and answers to questions sent by email. There is no group or chat to follow; the recording stays in the course.

What do founding members get?

Buy before 10 November 2026 and you keep $99 a year for good; the regular price from 2027 is $129. You also get one brief reviewed by email within 7 days, once a year of membership.

Your next task

Have a task that keeps coming back? Get the time back on your next attempt.

Check the scope of a year of access for $99 and the terms of the one-task guarantee.

See scope and price