929 Automated Tasks and 152 Homemade Tools: How Non-Programmers Reclaimed 83% of Wasted Time

• Scale of the internal contest: 99 participants and 929 digitized work cases
• Time saved: 22,578 claimed hours a year, 7,663 hours counted after strict caps (about 15 million rubles)
• Solutions built by employees: 152 working bots, dashboards, scripts, and pipelines (valued at about 19 million rubles)
• Share of non-developers: 68 of 82 public authors and 83% of all freed-up hours

Every week, our internet marketer used to spend half a day on the same mandatory routine: pull statistics from ad accounts, take a dozen and a half screenshots, consolidate the numbers into a 250-row spreadsheet, and write an analytical memo for management. Today a script she built herself does that job in a few minutes. She never learned any programming language syntax — she simply explained to an AI agent, in plain words, the structure of the source files and the format of the final report.

When we launched mass vibe-coding training at Alpina in early 2026, we wanted to understand how many of these quiet wins had piled up across departments. In August, we announced an internal check-up contest with a transparent leaderboard: show what routine task you automated, how many hours it used to take, and how long it takes now. In the end, 99 people submitted 929 concrete cases.

Why did non-technical staff, not IT people, deliver 83% of all the

time saved across the company?

It's commonly assumed that the main beneficiaries of agentic coding are staff programmers. Our numbers disproved that stereotype. Of the 82 participants who made their cases public, 68 have nothing to do with the IT department — they're marketers, sales managers, lawyers, accountants, and book editors. This group accounted for 83% of all the hours the company saved.

Even by the median figure, marketing beat development: 74 freed-up hours a year per person versus 60 hours for engineers. The reason is clear: for developers, many routine operations have long since been handled by scripts and CI/CD pipelines, while business units had accumulated years of manual Excel reconciliations, copying data between CRM systems, and shuffling documents between folders. A third of participants turned their work into transferable assets — altogether, employees with no tech support built 152 bots, dashboards, parsers, and web services.

How do you honestly calculate the impact of automation without fooling

yourself with pretty numbers?

If you simply added up every participant's self-reported survey, you'd get 22,578 hours saved a year. We deliberately refused to take raw self-assessment at face value and introduced a strict capping system:

• We fixed the annual working-hours base at 1,760 hours and capped the maximum credit per employee at a quarter of that.
• We set an upper limit of 220 hours a year for any single case.
• After automatic cutoffs, exactly a third remained — 7,663 confirmed hours, equivalent to roughly 15 million rubles a year at internal labor rates.

Combined with the valuation of the 152 in-house tools built, priced at external-contractor rates (about 19 million rubles), the total estimated effect came to roughly 33 million rubles. We always stress: this isn't live cash sitting in a bank account — it's time returned to the company and budget saved on outsourced development.

What 3 management decisions actually get a team to use AI every day?

As long as we were just handing out free platform logins, only one in four people showed up to practical workshops addressing their actual work pain points. The turning point came when we introduced three rules: 1. Every session is built strictly around a real weekly task of a specific employee, not abstract examples. 2. A single mandatory AI-workflow standard is introduced for the department's key process. 3. Work quotas get revised alongside the tool rollout: if a specialist with an agent produces a report three times faster but the output target stays the same, they have no incentive to show real productivity gains.

I publish breakdowns of the best of these 929 cases on my channel [@jemal_hamidun](https://t.me/jemal_hamidun). The environment for quickly building your own no-code business tools is at the core of [Nomad AI](https://nomadnet.ai). You can learn to automate your own weekly routine at [Hamidun Academy](https://academy.hamidun.com) with a free 72-hour trial, and grab 370+ ready-made corporate skills while working through your team's processes in masterminds at [Vibe Club](https://hamidun.com/club).