AI delegation for non-technical roles
Don't just prompt.
Start directing.
A few weeks from now, you send out work that once took most of your day. It took a fraction of that time, the important figures are checked, and you can explain exactly how it was produced. You did not learn to code. You learned how to direct AI.
The uncomfortable truth, said out loud
AI keeps getting better at your tasks.
It produces words, slides, summaries, analyses, and first drafts. Most professionals can feel the shift, even if they do not say the question out loud in meetings:
"Where does that leave me?"
Your value was never just in producing the work. It was in understanding what the work was for, deciding what a good result looked like, recognizing when something could not be trusted, and taking responsibility for what happened next.
As AI takes on more of the production, those responsibilities become more important.
Yet most people have never been taught how to direct AI well. They give it a thin instruction, accept whatever comes back, and then either send out mediocre work or spend the afternoon repairing it.
One repeatable system
Move from doer to director
This is not a collection of prompt tricks. It is a practical directing system you can use with today's tools and with the ones your organization adopts next.
Brief the work
Give the AI the task, the context, the materials, and the required format, so the first version comes back close to right.
Define what good looks like
Set explicit criteria before the work begins, so the result is judged against a standard instead of a vague hope.
Steer as it develops
Catch wrong turns early and correct course in seconds, instead of starting over at the end.
Check the result
Verify with scrutiny that matches the importance and risk of the task, before it goes out under your name.
The four moves sit inside one essential boundary: knowing what you are permitted to hand over, which information must remain protected, when independent verification is required, and when a task needs human or organizational oversight.
By the end of the course
What you will be able to do
Turn a vague request into a clear AI brief that produces a more useful first result.
Define success criteria and get honest critique and pressure-testing instead of automatic agreement.
Plan and steer longer assignments without losing the thread or repeatedly starting again.
Decide when current research is required and verify important claims, figures, and sources.
Work confidently with documents, languages, images, charts, screenshots, spreadsheets, and small workplace tools.
Recognize what is safe to delegate, what should remain private, and when to involve a manager or IT.
You will also develop a practical way to assess new AI models and features without depending on product hype or memorized model names.
Your work, not classroom examples
No matter the daily task
The course will help you do it faster and smarter.
Every skill is taught on the kind of work you already do: reports, updates, comparisons, research, drafts, and messy spreadsheets. You apply each move to your own examples and see the difference the same day, not a collection of disposable classroom exercises.
A free practice environment gives you a safe place to build the skill without connecting company systems or uploading sensitive workplace data. When your real work involves protected information, you learn to work with sanitized or representative material instead.
The work this course makes lighter
Whatever your version looks like, the method is the same.
Eight practical modules
What the course covers
Meet the instructor, see the whole map, and decide if this course is for you.
Understand why directing AI is becoming a core professional skill, see how two people with the same tool get very different results, and learn the essential vocabulary without unnecessary jargon.
- 1.1Why directing AI is now a core professional skill5 min
- 1.2The director vs. the doer6 min
- 1.3The core skill: directing the work, not doing it5 min
- 1.4Plain-language vocabulary you'll actually use6 min
- 1.5What AI already knows vs. what it looks up4 min
Give AI the task, context, source material, constraints, and format it needs. Learn to describe the problem clearly and match the depth of your brief to the stakes.
- 2.1Anatomy of a good brief: task, context, format7 min
- 2.2Context is everything: the brief decides the result8 min
- 2.3Handing over the right material7 min
- 2.4Describe the problem, not the fix, and brief to the stakes8 min
- 2.5Iterate, don't restart: directing in conversation7 min
- +When it goes wrong: the generic, off-target output
Use AI for critique, brainstorming, evaluation, and pressure-testing without leading it toward the answer you already want. Define "good" in advance so the work is judged against meaningful criteria.
- 3.1The yes-man problem: getting honest answers7 min
- 3.2Defining "good" up front: success criteria and checklists6 min
- 3.3Let it work: reasoning models and longer tasks6 min
- 3.4Brainstorming and pressure-testing ideas5 min
- 3.5AI as a reviewer of your own work6 min
- +When it goes wrong: it just agreed with you
Run longer assignments by asking for clarifying questions, reviewing a plan before execution, steering the work in stages, and adding a second review when the stakes justify it.
- 4.1Ask for the questions first5 min
- 4.2Make it plan before it executes6 min
- 4.3Steering mid-run: catching wrong turns early6 min
- 4.4A second set of eyes: one AI reviewing another5 min
- +When it goes wrong: the steer-vs-restart decision
Know when AI is answering from existing knowledge and when it needs current sources. Commission sourced briefings, check claims and citations, and direct work across languages without mistaking fluency for accuracy.
- 5.1When AI looks things up: web search vs. its own knowledge6 min
- 5.2Catching confident-but-wrong answers7 min
- 5.3Deep research: directing a multi-source briefing6 min
- 5.4Directing work across languages6 min
- +When it goes wrong: the citation that doesn't check out
Work with screenshots, charts, whiteboards, and spreadsheets. Create quick visual drafts, perform cautious first-pass analysis, and direct a small tool or workflow that removes a recurring bottleneck.
- 6.1Showing AI what you see: screenshots, charts, whiteboards6 min
- 6.2Directing quick visuals and drafts5 min
- 6.3First-pass data analysis on your own spreadsheets7 min
- 6.4Building a small tool to remove a recurring bottleneck7 min
- +When it goes wrong: the tool that half-works
Understand what can safely be handed over, how oversight should increase as AI moves from drafting toward deciding or acting, and what to do when sensitive information or an incorrect result slips through.
- 7.1Understanding the risks and benefits of what AI can do6 min
- 7.2What's safe to hand over and what isn't7 min
- 7.3Shadow AI: the hidden risk in getting ahead6 min
- 7.4Directing well when there's no policy yet5 min
- 7.5Owning the result: trusting outputs when your name is on it7 min
- +When it goes wrong: after a mistake gets out
Use the time you save for higher-value work, communicate your contribution honestly, and build a durable method for evaluating new tools as they appear.
- 8.1Move up the value chain: redeploy the time you save6 min
- 8.2Showing your work: AI etiquette and visible contribution5 min
- 8.3Staying current: sizing up a new model when it lands4 min
Built for real work
AI will sometimes misunderstand the brief, invent a fact, flatter your idea, lose the thread, or produce something that only half works. This course does not hide those moments behind perfect demonstrations. From Module 2 onward, every module teaches you to diagnose the failure, decide whether to steer or restart, keep what is still useful, and recover.
Every recovery method from the course, gathered into one practical reference you keep using long after the last lesson.
Yours to keep
Tools you'll reuse every week
The briefing checklist
A thinking aid for task, context, materials, format, and success criteria. Not a magic block of text to paste into every conversation.
The Failure-Mode Playbook
Every "When it goes wrong" fix on one page: what failed, why, and the recovery move that saves the work.
The new model test kit
A ten-minute routine for sizing up any new AI model with your own work, so the skill keeps paying after today's tools are gone.
Designed to outlive the tools
Why this course will hold up
It teaches judgment, not prompt collecting.
You learn how to think about the work before, during, and after AI touches it.
It applies across real workplace tasks.
The same method works for writing, research, multilingual work, images, spreadsheets, longer assignments, and small automations.
It builds confidence without blind trust.
You learn when a quick review is sufficient, when independent verification is necessary, and when AI should not be used at all.
It integrates responsible use into the work itself.
Data safety, sanctioned tools, oversight, and accountability are taught alongside the capabilities that make them necessary, not as a compliance lecture at the end.
It is built for what comes next.
Models and interfaces will change. The ability to brief, judge, steer, and verify will remain valuable, and the final module shows you how to assess new models yourself.
Your instructor
Learn from someone who has taught thousands
Andrei Gheorghiu has spent more than two decades as an IT consultant, auditor, and trainer, teaching thousands of professionals in information security, IT governance, and audit.
He wrote a book on building applications with large language models, and what drives him now is simple: helping people upskill and stay relevant in the age of AI. This course brings that experience to the people who need it most, professionals who direct work rather than write code.
Who it is for
Capable. Busy. Not a developer.
For you
You are responsible for real outcomes. You are not a developer, and you are not trying to become one. You do not want to study AI for its own sake.
You want to use AI confidently, protect the quality of your work, and become the person others trust to get good results from it.
Less risk, better output
A team that completes this course gets useful output from AI without creating unnecessary risk. Employees protect data, use sanctioned tools, verify important outputs, and ask before improvising outside established boundaries.
For organizations developing formal AI governance (including those using ISO/IEC 42001 as a reference), the course strengthens the everyday awareness and practical competence those programs expect. Policy tells people the rules. This course helps them apply the rules in daily work.
Enroll
Start directing today
Full course · one payment
Everything you need to move from doing to directing.
- All 8 modules: 3.5 hours of focused teaching
- Free practice environment, safe from company data
- Hands-on lab and quiz for every module
- The complete keepsake toolkit: checklist, playbook, test kit, and more
- Watch in order or jump to what you need
For teams and organizations
Per-seat pricing that scales with your team.
- Everything in the individual course, for every seat
- Volume discounts and invoice billing
- Supports workforce awareness for AI governance programs, including ISO/IEC 42001
- Employees who protect data and ask before improvising
Watch the introduction free before you decide. If you don't recognize yourself in the first four minutes, keep your money.
No course can promise that AI skills protect a job, and you should be suspicious of any that does. This one promises something more credible: a professional capability you can demonstrate, through work that is faster to produce, easier to check, and stronger when it reaches another person.
From doing to directing
The tools will keep changing. The ability to direct them well will keep mattering.
Become the person others trust to brief, steer, and verify AI-assisted work.