Between 1 January and 16 October 2026, our team of three recorded 12,288 commits, compared with 1,421 over the same period in 2024, when we were four. We merged 22 times more pull requests and committed to 41 repositories, up from 15.
This rise coincides with our move to working with AI coding agents, guided by a senior team that keeps control of every decision. The figures show this change; they cannot isolate the effect of AI, because the number and type of projects also changed.
Three years of data at a glance
Same period, three years
1 January to 16 October
- 2024(team of 4)
- 2025(team of 4)
- 2026(team of 3)
- Commits
- 20241,42120252,017202612,288
- Merged pull requests
- 202474202514220261,620
- Repositories with commits
- 202415202517202641
- Source code lines added
- 2024173,5032025388,54020262,667,698
- Test share of added code
- 20241.2%202511.1%202631.8%
Source: Git history of the 62 repositories in the Appik-Studio GitHub organisation, extracted on 16 October 2026. Only team commits are counted: six forks, upstream history, external contributors and bots are excluded.
Pull requests: GitHub search across the Appik-Studio organisation, pull requests merged between 1 January and 16 October of each year. Lines added: source files only, excluding generated files and commits over 10,000 lines. Test share: lines added in test files as a share of all source lines added.
With one fewer person, the team produces 8.6 times more commits and adds 15 times as many source code lines. Almost one added line in three is now a test, compared with 1% in 2024.
The shift in late 2025
Team commits per month, January 2024 – October 2026
Around 150 commits a month in 2024, then a rise from late 2025 up to 1,977 commits between 1 and 16 October 2026.
- June 2025: first instruction file for coding agents in a project.
- Late 2025: the team goes from four to three people.
October 2026: 1–16 only.
Source: Git history of the 62 repositories in the Appik-Studio GitHub organisation, extracted on 16 October 2026. Only team commits are counted: six forks, upstream history, external contributors and bots are excluded.
Show the monthly data
| Month | Commits |
|---|---|
| Jan | 100 |
| Feb | 123 |
| Mar | 149 |
| Apr | 50 |
| May | 121 |
| Jun | 102 |
| Jul | 96 |
| Aug | 390 |
| Sep | 178 |
| Oct | 182 |
| Nov | 182 |
| Dec | 150 |
| Month | Commits |
|---|---|
| Jan | 402 |
| Feb | 164 |
| Mar | 164 |
| Apr | 104 |
| May | 85 |
| Jun | 130 |
| Jul | 215 |
| Aug | 373 |
| Sep | 313 |
| Oct | 257 |
| Nov | 194 |
| Dec | 394 |
| Month | Commits |
|---|---|
| Jan | 750 |
| Feb | 453 |
| Mar | 987 |
| Apr | 1,362 |
| May | 1,075 |
| Jun | 1,018 |
| Jul | 1,786 |
| Aug | 1,288 |
| Sep | 1,592 |
| Oct (1–16) | 1,977 |
The first instruction file for coding agents appeared in our projects in June 2025. From late 2025, activity takes off and then stays well above earlier levels. Three habits were adopted over the same period.
Every project has its instructions. A file versioned with the code describes the conventions, tests and security rules. An agent starts from the right context, and every corrected mistake becomes a rule.
Every change is checked. Tests and automated checks run before anything is merged, in a pull request.
Tests are part of delivery. An agent prepares the first version of a test and the team reviews it. Regressions are caught before they reach your users.
According to the 2025 Stack Overflow survey, 84% of developers use or plan to use AI. The tool alone does not make the difference. Google Cloud's 2025 DORA report describes AI as an amplifier of a team's strengths and weaknesses, and a randomised controlled trial by METR found that experienced developers were 19% slower with early-2025 tools. These studies suggest that the method matters as much as the tool.
What it changes for your project
- A first version sooner, so choices can be validated with real users.
- Tests from the start, so the product can evolve without breaking what works.
- A senior team over time, with the same people as contacts.
AI can speed up writing code. Scoping, architecture, review and release to production remain with our team. These steps determine the quality of the result.
Frequently asked questions
Does AI write all your code?
No. Agents help write the code, tests and documentation. The team decides what to build, reviews the changes and remains responsible for every release to production.
Is a project cheaper with AI?
Writing code can be faster, but the budget depends mainly on scope, integrations and the level of quality required. Our article on the cost of a mobile app in Switzerland details these factors.
Your project with a small, well-equipped team
Do you have a mobile app, web platform or business tool to build? Let's talk about your project.
SaaS
santé
Swiss Made Software