What a Private AI Tutor Actually Does — and Why One-to-One Beats a Course

There are more AI courses in London than ever — and most people who take one still aren’t using AI in their actual work a month later. The videos were fine. The problem is that a generic course teaches AI in the abstract, on someone else’s example, at a pace set for a room full of strangers, and then leaves you to translate all of it into your own business on your own.

A private AI tutor closes exactly that gap: applying the tools to your real work, with someone who has already done it sitting beside you.

I’m Vitalij Kolotikov. I run private AI training in Canary Wharf and online, and I build real businesses and software with AI before I teach any of it. This is what one-to-one AI tutoring actually looks like, who it helps, and why — for most people — it beats a course.

What a private AI tutor actually is

“Tutor” tends to bring school to mind. This is closer to working with an experienced colleague for an hour.

A private AI tutor:

• works with you, one to one (or a small group of your choosing) — not a class;
• uses your own work as the material — your marketing, your website, your admin, your product idea;
• adapts to your level, whether you’ve never opened ChatGPT or you already automate things;
• and stays fixed on one outcome: you leave able to do something you couldn’t do before.

Nothing is theoretical. If you run a hotel, we work on your bookings. If you’re a marketer, we work on your campaigns. If you’re changing careers, we start a portfolio piece you can show.

What a one-to-one session looks like

A first 60-minute session usually runs like this:

1. We start with a real task you brought — not a demo. Something you actually need done.
2. We pick the two or three tools that fit it. Out of the hundreds out there, most are noise for your situation.
3. We build it together, live. You drive; I guide. You see exactly which prompt, which setting and which step made the difference.
4. You leave with the thing — a working draft, a set-up tool, a repeatable process — and notes you can follow again alone.

The point of the first hour isn’t to cover everything. It’s to get one useful result and remove the “where do I even start” feeling that stops most people before they begin.

Why one-to-one beats a group course

A good course has its place. But for actually putting AI to work, private one-to-one training wins on the things that matter:

• Relevance. Every minute is spent on your work, not a generic example you have to reinterpret later.
• Pace. We move faster where you’re comfortable and slow down where you’re not. No waiting for the room; no being left behind.
• Implementation. You don’t leave with intentions — you leave with something built and running.
• Honesty. When a tool isn’t worth your time, I’ll say so. A course selling itself rarely will.

Group and company sessions are still the right call sometimes — team training brings everyone to a shared standard. But that’s a different goal from getting you productive quickly.

Who private AI tutoring is for

Most people I work with fall into one of three groups:

• Business owners and managers who want AI applied to real operations — marketing, sales, service, reporting — with a plan, not a set of slides. → AI for business
• Beginners and career changers with no technical background, who want practical, employable skills and a real project to show. → AI for beginners
• Professionals and small teams who already use AI a little and want to go from occasional to fluent.

You don’t need to code, and you don’t need to have used AI before. The pathway adapts to wherever you’re starting from.

What you can actually learn

People arrive with very different goals, so sessions cover very different ground. The most common:

• AI for marketing — campaigns, content, SEO support, email, competitor research and a more consistent system.
• AI automation — taking the repetitive admin, follow-up and reporting off your plate.
• ChatGPT, properly — prompts, assistants and workflows you can rely on, not one-off tricks.
• Websites and apps — planning and building real software with AI, even without a developer background.

Not sure which you need? That’s fine — choosing the right focus is part of the first session. You can also see every AI course in one place.

Being honest about results

I won’t tell you that one hour turns you into an AI expert. It doesn’t, and anyone who promises that is selling something. Here’s what’s realistic:

• After one session — you’re oriented, the right tools are set up, and you’ve made one useful thing.
• After three — you have momentum in one area: a marketing system started, a working site, or an automation running.
• After five to ten — the skills combine into a real project you can keep building on, and you’re confident continuing alone.

Progress depends on the work you do between sessions. I give you the direction, the tools and the method; you bring the reps.

I don’t only teach AI tools. I build businesses and software with them — then I teach you exactly how.

What good AI training should cover in 2026

Knowing which button to press is no longer enough. In January 2026, Skills England published an AI foundation skills benchmark for work built around three connected domains: technical ability, non-technical judgement, and responsible or ethical use. That is a useful test for any tutor or course.

A practical session should therefore improve more than your prompt:

• Task judgement: recognising when AI is useful, when an ordinary tool is better, and when the task should stay human.
• Briefing and evaluation: giving clear context, then testing the output against facts, audience and purpose.
• Data judgement: knowing what not to paste into a tool and how to work with anonymised examples.
• Human oversight: deciding who checks an output before it affects a customer, colleague or important decision.
• Repeatability: turning one good result into a documented process you can run again.

The strongest evidence that training worked is not a certificate or a list of tools. It is a small piece of work you can reproduce a week later without the tutor beside you.

Use the first session as a small experiment

Before you book a long programme, define a one-hour test. Bring one task with a real output and a clear finish line. For example:

Weak goal	Testable first-session goal
“Learn AI for marketing”	Turn one approved case study into an email and two channel-specific drafts, with a fact-check list
“Automate my business”	Map one repetitive enquiry workflow and build its human-reviewed first-draft step
“Get better at ChatGPT”	Create and test a reusable prompt for a task you perform every week
“Build an app”	Define one user journey and produce a working, testable first screen

At the end, ask four questions: Did we finish something useful? Can I explain why it worked? Do I know the risks and limits? Can I repeat the process? Four honest “yes” answers are a better reason to continue than enthusiasm alone.

In person in Canary Wharf, or online

Sessions run one-to-one or in small groups — either on Level 14 of The Cabot in Canary Wharf, with floor-to-ceiling windows over the Thames, or online, wherever you are. Companies can also have the training delivered at their own office.

The price is simple: £145 + VAT for a 60-minute lecture, whether it’s one person or a small group. Nothing upfront — you pay in person or after the session.

How to start

The best first step is a single lecture. Bring one real goal — a task, a campaign, a business idea, or simply “I want to finally use this properly” — and we’ll make progress on it in the first hour.

If you’d like to see who you’d be working with first, read more about how I work. When you’re ready, book a private lecture and tell me what you’d like to achieve.