Which course, and is it available today
This is about the "AI for Business 2.0" programme from Google Ukraine, built together with the Ministry of Economy of Ukraine and the educational platform EdEra. The first intake started on 10 November 2025; the organisers extended access to its materials and certificate until 10 May 2026. The second intake opened on 12 October 2026, and on the day of checking registration is open.
What the course page says as of 27 October 2026.
- Format. Five modules, each opening on a Monday: 12, 19 and 26 October, 2 and 9 November. A module contains videos of up to 15 minutes, a workbook with ready-made prompt structures, and a test. Expert webinars every Thursday; recordings are kept.
- Time. The organisers estimate 2–3 hours a week, which is 10–15 hours for the whole course.
- Certificate. After watching more than half of the module videos and passing the final test.
- Access after the end. According to the answers on the course page, registered participants keep permanent access to the videos, workbooks and prompts. So it can be taken after 9 November too, at your own pace.
- Price. Participation is free. You will pay only for tools, if after the course you choose a paid plan, and that is covered below.
If you start today, three modules are already open, the fourth appears on 2 November and the fifth on 9 November. At one module a week you will finish by 30 November.
The five-module programme and small-business tasks
The module descriptions on the course page are written in the language of a "corporate system" and "KPIs for every department". A company of two to ten people has no departments, so we translated each module into tasks of that size and pointed to the article on this site to read when the module ends.
| Module and access date | What the course description promises | What applies to a company of up to 10 people | Continue on this site |
|---|---|---|---|
| 1. Personal productivity · from 12 Oct | An audit of your working time, "resource drains", a personal ecosystem of assistants that frees up at least 30 % of your time | One task you do by hand every day, and a count of minutes before and after. A prompt for a specific task, not a pack of "100 prompts" | A prompt for a business task, three assistants tested on five tasks |
| 2. AI for business growth · from 19 Oct | Media campaigns without agencies, your site in AI Overviews answers, CRM automation, round-the-clock customer support | Texts and descriptions, answers to typical customer questions through a bot with handover to a person. "A site in AI answers" for a small business means pages with verifiable facts, not a separate tool | A chatbot for business, AI product descriptions, AI content and Google |
| 3. Data, projects, operations · from 26 Oct | A central knowledge base, financial forecasting, hiring, market risk intelligence "without the risk of information leaks" | A knowledge base built from your delivery rules, prices and return terms. A sales forecast for a shop with 200 orders a month is a spreadsheet job; AI writes the explanation next to it. The module's main question for you: which data may go into a chat at all | What data you can give an AI assistant, when you need a CRM |
| 4. No-code automation, agents · from 2 Nov | Your own apps built through dialogue with AI (vibe coding), services linked into an "autopilot", networks of autonomous assistants | One scenario: an order from the site creates a CRM card and a message to the manager. Your own app is rarely needed; more often you need a feature that already exists in your service | Automation without a developer, what not to automate |
| 5. From pilots to scale · from 9 Nov | An implementation roadmap, control over API spending, getting the team on board, measurable KPIs | For ten people the "roadmap" is a decision on how many seats and which plan to buy, and one number a month: hours saved against the cost of subscriptions | What AI costs for a team of ten, the industry series hub |
Note the third module. Two of its speakers listed on the course page are lawyers, and its description contains the words "without the risk of information leaks", so data questions are in the programme. But the rules that constrain a business in Ukraine depend on the industry, and this is exactly where a one-size course cannot be precise by definition.
What in the modules applies in your industry
In the industry series hub we already laid out what can be handed to AI in each type of business and which legal provision stands behind each prohibition. Here the same logic is mapped onto the course modules. Three columns are three groups of businesses with different constraints.
✓ — take it from the module as is; ≈ — with the condition in the cell; ✗ — this part of the module is not applied in your industry. The provisions behind the conditions, with links to the texts of the laws, are in the series hub and the industry breakdowns.
For the second column the order of actions is the reverse of the module order: first the data-handling rules from the article about data for an AI assistant and the breakdown for your industry, and only then module one. The industry breakdowns in the series: online shop, café and restaurant, beauty salon, clinic, law or accounting firm, construction and renovation, online school, real estate agency, car service, hotel or guesthouse, family farm, small manufacturer.
What the course promises and how to measure it yourself
The course page carries several numbers and claims that should be read as hypotheses, not results. We did not verify them and cannot: they are the organisers' wording. But each can be turned into a measurement in your own company.
- "Frees up at least 30 % of your time". Measurement: one task, a week without the assistant and a week with it, minutes in a spreadsheet. If the task took 90 minutes a day, 30 % is 27 minutes. Less than an hour a week saved means the task is wrong, not that the course is bad.
- "Your site at the top of AI Overviews answers". Measurement: in Search Console, look at impressions and clicks for the pages you rewrote after the course, 4–6 weeks later. What actually changed in Google's requirements for text in October 2026 we covered separately.
- "Round-the-clock customer support". Measurement: 50 real customer questions through the bot before launch, and the share of answers handed to a person. The method is in the article about chatbots.
- "Without the risk of information leaks". This is not a measurement but a settings check: whether model training on your data is switched off and whether you are on a business plan. Where to find those switches is in the article about data for an assistant.
The testimonials from first-intake graduates on the course page ("discharge summaries three times faster", "visibility three times better") are also participants' numbers, not the organisers' and not ours. Treat them the same way: as what someone somewhere got, not as what you will get.
A plan for the first month after the course
The course ends with a test and a certificate. For a business it ends when one task is done differently every day and that shows in minutes. Four weeks is enough for that.
- Is there a task someone does by hand every day for more than 30 minutes?Product descriptions, answers to the same questions, letters to clients, discharge summaries. If there is none, the course stays general education, and that is fine. Yes → week 1: measure without the assistant
- Does that task need client, patient or contract data?If so, data rules and a business plan come first, and pick a different task for the test, one without that data. No → week 2: the same task with the assistant
- Does the task pull data from two systems and repeat without changes?An order into the CRM and a message; a form entry into a spreadsheet and an email. Then it is not an assistant but automation, and it is counted differently. Yes → week 3: one automation scenario
The weekly plan, in that order.
- Week 1, measurement. The same person does the task as usual and writes down the minutes every day. Without this week you will not be able to tell an impression from a result later.
- Week 2, the task with the assistant. A personal or free plan, a prompt for this task from the course workbook or built our way. Record the minutes and the number of results you had to redo.
- Week 3, the count and a second task. If the saving is under an hour a week, or you had to redo more than one result in five, the task is wrong: take the next one from the module table, not the next service. If it works, take an automation scenario from module four, but only one.
- Week 4, the money decision. Count the hours saved per month against the cost of plans for the number of people who actually need them. A pricing table with a check date is in the article about AI for a team of ten.
An illustrative example for scale. A shop owner writes 25 product descriptions a week at 20 minutes each: 500 minutes, 8 hours 20 minutes. With the assistant a draft takes 8 minutes and 5 minutes go to checking: 325 minutes, 5 hours 25 minutes. The saving is 175 minutes a week, about 2 hours 55 minutes, and 700 minutes over four weeks, 11 hours 40 minutes. That clears the one-hour-a-week threshold. But if six of the 25 descriptions had to be rewritten, that is 24 %, more than one in five, and the prompt and the product fact base need fixing first, not a team-wide plan.
What the programme description does not cover
Three things that are absent from the module descriptions on the course page and that matter more for a small business in Ukraine than the choice of model.
- Tool prices per person. Module five promises control over API spending, but a small business more often pays for seats in a subscription than for tokens. A comparison of the plans of three assistants and of automation services, with a check date, is in the series hub and in the calculation for a team of ten.
- A comparison of assistants from different providers on Ukrainian tasks. A course from one provider is naturally built on its own tools; the descriptions of modules two and four also name automation services from other companies. How three assistants differ on five business tasks in Ukrainian we tested ourselves.
- Industry prohibitions. Medical data, attorney-client privilege, fiscal documents. A course for everyone cannot break them down by industry, and without them modules one and three are dangerous in a clinic or a law firm. That is what the industry series is for.
Our position
Take the course: it is free, the materials stay, and a workbook with prompt structures is worth two or three hours a week. But take it with a spreadsheet of minutes beside you, not with lecture notes. One task, a measurement before and after, a money decision in week four. The numbers in the course description, 30 % of your time or the top of AI answers, stay hypotheses until you see them in your own spreadsheet.
The position does not apply if you have been using an assistant and automation for a year and came to the course for the certificate or the community. Then the measurement will show you nothing new, and it makes sense to go straight to modules four and five.
Checked on 27 October 2026 on the "AI for Business 2.0" programme page on the EdEra platform (modules, access dates, format, certificate, access after the end) and on the page of the programme's first intake (access extended to 10 May 2026). The shop example is illustrative; substitute your own minutes.