⚡ Velocity: Now vs. 3 Months Ago
Weekly Commit Velocity (Nov 2025 — May 2026)
Monthly PRs
Monthly Commits (AI vs Non-AI)
3-Month Velocity Comparison: Feb 2026 vs. Mar 2026 (Peak)
| Repo | Feb Commits | Mar Commits | Change | Feb PRs | Mar PRs | PR Change |
|---|---|---|---|---|---|---|
| digital-qa-mobile | 45 | 38 | -16% | 3 | 12 | +300% |
| essity-presentations | 37 | 48 | +30% | 3 | 23 | +667% |
| help-ct-com | 1 | 7 | +600% | 45 | 48 | +7% |
| ct_abap_analyser | 0 | 4 | New | 0 | 0 | — |
| code-evaluation-skills | 0 | 2 | New | 0 | 1 | New |
| Total | 83 | 99+ | +19% | 58 | 92 | +59% |
⏳ PR Cycle Time Analysis
Overall Merge Speed
of PRs merged in under 1 hour
89% merged within 24 hours. Median cycle: <1 minute. The team moves fast with minimal review bottlenecks.
Median Cycle by Repo
digital-qa-mobile has the longest cycle (1.9h median) — more complex review process for AI test automation code.
Monthly Cycle Time Trend
Nov started at 70h median, dropped to near-zero by Dec as team streamlined flow.
📊 Commit Patterns & Work Rhythms
Commits by Day of Week
Commits by Hour (UTC)
Work Pattern
97.4% of commits are weekday. Peak: Tuesday (172). Weekend: 19 commits total. Core hours: 11:00—18:00 UTC (13:00—20:00 EET), matching Ukraine/Eastern Europe timezones.
Busiest Day
commits on Mar 10 & Mar 18, 2026. Week of Mar 10 was the peak week at 89 commits. This aligns with L'Oreal and PUMB demo stand launches in digital-qa-mobile.
Velocity Acceleration
Week-over-week growth from Nov W47 (20 commits) to Mar W10 (89 commits). The team went from steady trickle to sustained high output.
🤖 AI Impact Deep Dive
AI-Related vs Non-AI Commits Over Time
Commit Message Quality
| Category | Median Length | Multi-line | Has Rationale |
|---|---|---|---|
| AI repos | 270 chars | 41% | 36% |
| Non-AI repos | 60 chars | 58% | 21% |
| Claude co-authored | 351 chars | 63% | 27% |
AI-assisted commits have 4.5x longer commit messages and are 71% more likely to include rationale (why, not just what).
AI Adoption Curve
% of commits in AI-related repos. Nov was 81% AI because digital-qa-mobile was the primary active repo.
Kosyakov Dmitry: Claude Sonnet Day
All on Feb 27, 2026. Built a complete iOS medical data extraction app with Claude Sonnet 4.5 in a single session. Average commit message: 524 chars (2.5x longer than org average). Scaffold, refactor, test, optimize — full lifecycle in one day.
👑 Top Contributors
Contributor Velocity & Acceleration (Nov 2025 — May 2026)
| Contributor | Repos | Total | Nov | Dec | Jan | Feb | Mar | Apr | May | Commits/Mo | AI % | Avg Msg Len |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| dorohin | essity-presentations | 181 | 0 | 24 | 24 | 48 | 85 | 0 | 0 | 45.2 | 0% | 455 |
| Oleksandr Eskin | digital-qa-mobile | 140 | 15 | 21 | 8 | 49 | 47 | 0 | 0 | 28.0 | 100% | 312 |
| vbelyakov | help-ct-com | 92 | 0 | 18 | 0 | 11 | 31 | 20 | 12 | 18.4 | 0% | 44 |
| vbelyakov-ct | help-ct-com | 106 | 223 PRs, median cycle <1 min — documentation powerhouse | 13.2 | 0% | 50 | ||||||
| Alexey Basov | digital-dev-demo ct_abap_analyser | 38 | 0 | 7 | 5 | 2 | 21 | 1 | 2 | 6.3 | 36% | 21 |
| Artem Korenev | Specs-fork | 42 | 2 | 1 | 3 | 2 | 0 | 0 | 0 | 2.3 | 0% | 25 |
| denysvarchenko | CTFoundationModels | 22 | 0 | 0 | 0 | 0 | 1 | 14 | 0 | 3.7 | 68% | 326 |
| Kosyakov Dmitry | DemoApp | 8 | 0 | 0 | 0 | 8 | 0 | 0 | 0 | 8.0 | 100% | 524 |
| illia-dorohin-ct | essity-presentations | 8 | 30 PRs for UAT releases — deployment orchestrator | 2.0 | 0% | 76 | ||||||
| Artem Orynko | code-evaluation-skills | 5 | 0 | 0 | 0 | 0 | 5 | 0 | 0 | 5.0 | 100% | 483 |
| Volodymyr Hulchenko | ios-code-review-rules | 2 | 0 | 0 | 0 | 0 | 0 | 2 | 0 | 2.0 | 100% | 79 |
Top Contributor Monthly Acceleration
📦 Repository Health Dashboard
All 21 Repositories — activity, health, and AI adoption
| Repository | Language | Commits | PRs | Authors | Active Months | Velocity (c/mo) | Open PRs | AI |
|---|---|---|---|---|---|---|---|---|
| help-ct-com | CSS/HTML | 209 | 230 | 5 | 10 | 20.9 | 2 | |
| essity-presentations | TypeScript | 189 | 30 | 2 | 4 | 47.2 | 0 | |
| digital-qa-mobile | Python | 172 | 23 | 4 | 5 | 34.4 | 4 | Claude CI |
| Specs-fork | Ruby | 69 | 0 | 9 | 24 | 2.9 | 0 | |
| ct_abap_analyser | TypeScript | 24 | 0 | 1 | 3 | 8.0 | 0 | AI Workflow |
| digital-dev-demo | Java | 19 | 26 | 2 | 3 | 6.3 | 7 | GenAI Auto |
| CTFoundationModels | Swift | 15 | 1 | 1 | 2 | 7.5 | 1 | Foundation |
| DemoApp_FoundationModels | Swift | 8 | 0 | 1 | 1 | 8.0 | 0 | Claude 4.5 |
| z_assist_chat | JavaScript | 8 | 0 | 1 | 1 | 8.0 | 0 | RAG / SAP |
| code-evaluation-skills | Markdown | 6 | 1 | 2 | 1 | 6.0 | 1 | Claude Skills |
| Landmark-app | TypeScript | 5 | 3 | 2 | 2 | 2.5 | 0 | |
| AFNetworking | Obj-C Fork | 3 | 0 | 1 | 1 | 3.0 | 0 | |
| ios-code-review-rules | Markdown | 2 | 0 | 1 | 1 | 2.0 | 0 | AI Review |
| digital-qa-reports | Python | 1 | 0 | 1 | 1 | 1.0 | 0 | |
| CTBankDemo | Swift | — | 0 | — | — | — | 0 | |
| ct-quiz-game | SCSS | — | 0 | — | 1 | — | 0 | |
| playwright | TS Fork | 1169* | 4 | 100+ | — | — | 1 | |
| Charts | Swift Fork | — | 0 | — | — | — | 0 | |
| efcore | C# Fork | — | 0 | — | — | — | 0 | |
| RMStepsController | Obj-C Fork | — | 0 | — | — | — | 0 | |
| essity-presentations † | * Playwright count is upstream fork history • † Highest velocity repo at 47.2 commits/month | |||||||
🚀 AI Adoption Timeline
🏆 Key Findings
PR Throughput: 8.4x Growth
From 11 PRs (Nov 2025) to 92 PRs (Mar 2026). This wasn't gradual — it was a step function triggered by team scaling + AI tooling adoption.
AI Commits = Higher Quality Messages
AI-repo commits average 270 chars vs 60 chars in non-AI repos (4.5x). Claude co-authored commits average 351 chars. AI isn't just faster — it produces better-documented code.
Near-Zero PR Cycle Time
84% of PRs merge in under 1 hour. The team has minimal review friction — most changes flow directly to main branches. This maximizes velocity but may indicate a review coverage gap.
Force Multiplication
Oleksandr Eskin (1 person) maintains AI test automation across 4+ client projects. Kosyakov Dmitry shipped a full iOS app in 1 day. AI enables individual contributors to produce team-level output.
9 New Repos in 2026
57% of all org repos were created in 2026 (5 months). The AI tooling investment is spawning new projects at an accelerating rate: ABAP analyzer, Foundation Models, code review rules, evaluation skills.
Note: LOC Data Unavailable
The current GitHub token has search-only access to private repos — cannot clone or fetch per-commit additions/deletions. To add lines-of-code analysis, grant the token repo scope or clone repos locally with proper access.