TestCon Europe 2026

AI-Driven Software Quality Measurement in Practice

About four years ago, Roman Rodomansky's team built a tool to track software quality across many engineering teams. Now AI agents review and summarize its data across more than 50 projects at once. At TestCon Europe in Vilnius, Roman shows how it works, and how measured quality ties to engineering output and revenue.

Upcoming

Friday, October 23, 2026

10:10–10:55 Vilnius time (EEST)

IMAX hall, Forum Cinemas Vingis
Savanorių pr. 7, Vilnius, and online

Starts in 5 days

Add to Google Calendar Apple or Outlook (.ics)

Conference: TestCon Europe 2026
Format: 45-minute talk
Where: Vilnius and online

The session on testcon.lt

50+projects whose quality the framework measures at once
~4 yearssince the team built its first quality tool

What the talk covers

The framework started as an in-house tool for tracking quality across engineering teams. Recently the team added AI agents to its key parts, such as reviewing code automatically and summarizing quality data. The talk walks through:

How the framework is built, and which of its parts matter most.
Where AI agents take over manual work, and make quality easier to see and to read.
How measured quality connects to engineering output and to business results, revenue included.

Its topics: quality metrics, AI agents, observability, business impact and productivity.

Session details

When: Friday, October 23, 2026, 10:10–10:55 Vilnius time (EEST, UTC+3).
Where: the IMAX hall at Forum Cinemas Vingis, Savanorių pr. 7, Vilnius. TestCon also streams online.
Conference: TestCon Europe 2026 (opens in a new tab), a conference for software testing and QA professionals.

The full abstract is on the session page (opens in a new tab). We'll add the recording and the slides here after the talk.

The story so far

The talk picks up from Roman's WAWTech 2025 talk in Warsaw, Our Playbook to Measure Software Quality Using AI (opens in a new tab): how his team moved from 500-question review playbooks to agents that suggest answers from each project's documents, route questions by role and turn findings into tickets. The video is there, if you want the background before Vilnius.

For the people side of AI, how to measure adoption across people, teams and the whole organization, see his DevWorld 2026 talk, Measuring AI Adoption Across People & Orgs (opens in a new tab).

In AI Adoption Score

The framework itself runs in Roman's team's own delivery. AI Adoption Score brings its checklist side to any team: the Software Quality Management (opens in a new tab) survey reviews one project against 248 practices in 14 areas, asks each person only their role's questions, and scores every area out of 10 with what's missing listed most important first.