Senior QA lead + AI agent · EU-based · Working with teams in Europe and the U.S.

The machine it doesn't work on. QA outsourcing and release testing for SaaS teams.

Release testing & QA outsourcing for SaaS teams — one senior lead instead of a team to manage.

I run your release testing with an AI agent wired into your Jira, Confluence and test management tool. It reads your specs, writes the test cases, builds the test cycles, runs your API and automated tests, and files the bugs. I review everything, test what needs a human, and sign off. The output of a small QA team, with one person accountable for it.

Book a 15-min call See how it works

Fixed-price start from €2,500 · Read-only by default · Everything stays in your tools

release reportACME · 4.2

$ womm release --version 4.2 --project ACME

✓Read 38 Jira tickets and 11 Confluence pages

✓Drafted 46 test cases — plain-English UAT + technical — reviewed and approved

✓Created test cycle in Xray · 46 cases · assigned to 3 testers

✓API smoke: 71 checks · auth, validation, OpenAPI contract, p95 latency

✗List endpoint returns records already removed by retention (404 on detail) — High

✗Invalid filter value silently ignored instead of 400 — Medium

✓Existing Playwright suite run in CI · 212/214 pass · 2 flaky, flagged

!9 cases need human judgment — assigned for manual testing

✓Bugs filed in Jira with evidence · release report sent to management

A sample release run. Yours shows your tickets, your tools and your bugs.

Release testing is mostly glue work

Most QA time before a release doesn't go into finding bugs. It goes into reading specs, writing cases, creating tasks and cycles, chasing assignments, copying results between tools and writing the status report. That work is necessary, slow, and exactly what an agent with API access does well.

Test cases are always a week behind the release

Cases are drafted straight from your tickets and specs, the same day they're ready.

Our business testers can't follow the test cases

UAT cases are written in plain English for people who don't read API docs. Technical cases stay technical.

Nobody knows what's actually been tested

Every case, result, defect and piece of evidence lives in your test management tool, linked to the ticket it came from. The release report writes itself.

How it works

01
Connect
Read-only API access to Jira, Confluence, your test management tool and your test environment. The agent maps your product: areas, risks, existing tests, gaps.
02
Draft
For each release, the agent reads the tickets and specs and drafts the test documentation and cases. I review and approve every one.
03
Build
One run creates the whole structure: Jira tasks, test cases, the test cycle, tester assignments, links back to source tickets.
04
Run
API smoke and regression tests, plus your existing automated suites. Failures are triaged before anyone sees them.
05
Test by hand
What needs human judgment gets tested by a human: me, or your business testers working from cases they can actually follow.
06
Report
Results, evidence and defects recorded in your tools. A release report your management can read in five minutes.

Agent vs me

The agent does:

Reads tickets, specs and history across your tools
Drafts test cases and documentation
Creates and syncs tasks, cases and cycles
Runs API and automated tests, collects evidence
Drafts bug reports and the release report

I do:

Decide what's worth testing and how deep
Review and approve everything before it's written to your tools
Test what needs judgment: flows, UX, edge cases
Decide whether a bug is really a bug
Sign off the release and answer for it

Works with what you already use:

Jira
Confluence
Xray
Zephyr
AIO Tests
TestRail
REST APIs / OpenAPI
Playwright
Cypress
CI pipelines
AWS CloudWatch (read-only)

Different tool? If it has an API, I can connect to it.

An agent in your tools. Here's how it's kept safe.

Read-only by default. Anything that writes is a dry run first, then approved by me.
Your credentials stay in your environment, never in prompts, logs or reports.
Staging and synthetic data by default. No production data unless scoped and agreed.
Separate browser profile and test accounts — nothing runs as your people.
NDA and GDPR data processing agreement signed as standard.
No platform of mine to buy: scripts and tests live in your repository.

Pricing

Fixed price, fixed scope. Start small, keep going only if it works.

QA Health Check

€2,500–4,000

2 weeks

A review of your test process, tools and coverage, an API smoke test, and a prioritized report with a 30/60/90 plan.

If it finds nothing worth fixing, you don't pay.

Recommended start

Release Sprint

€3,500

one release

One release tested end to end: cases, cycle, API and automated runs, manual testing on critical paths, management report.

Test Operations

from €3,000

/ month

One-off onboarding €1,500, then every release handled as above. API regression suite grows in your repo with each release. Weekly written report.

I take on a small number of clients at a time.

Invoicing in EUR or USD.

What you receive

Test cases in your tool, linked to the tickets they cover
A test cycle with assignments and results
API test scripts in your repository, yours to keep
Bugs in Jira with steps, expected vs actual, and evidence
A release report: one page for management, details for engineers

Why this way

vs. a QA outsourcing team

One senior person, no handoffs, no juniors learning on your product. The agent covers the volume.

vs. AI testing tools

They click through your UI. I run your whole test process, inside the tools your team already uses.

vs. hiring

No recruitment, no ramp-up, stop any time.

Who's behind it

Tomasz Kordiak

Senior QA lead · 19 years in software quality · ISTQB Advanced

I've spent 19 years making sure software works on more than one machine: as the first and only QA on startup products, and as a specialist inside large enterprise programmes. More than a decade of that was in regulated environments (banking, payments, compliance, EU trust services) where "it mostly works" isn't an option.

I've built QA functions from zero more than once: test strategy, automation, release gates, and UAT run with business stakeholders who don't read API docs. Today I specialise in the quality of AI and agentic systems: evaluating LLM features, red-teaming them, and testing agents that make real decisions. I use the same kind of agents to run test processes faster than a team could by hand.

19+ years in QA and test automation
Previously at Microsoft, Credit Suisse, Swift and Nokia
Built QA from scratch for a Belgian e-invoicing and digital post startup
Python, Playwright, Cypress, Java, C# · API, performance (k6, Gatling), AWS
ISTQB Advanced Level certified
Founder of the Test:Fest software testing conference
Polish and English · years of work with teams in Belgium and the U.S.

The person on your first call is the person doing the work.

What I won't promise

That AI replaces your QA team
100% coverage or zero bugs in production
An agent that decides on its own that your product works

FAQ

Is AI testing my product unsupervised?

No. The agent drafts and runs; I review and approve anything written to your tools, and I sign off.

What access do you need?

API tokens for Jira, Confluence and your test management tool, plus access to a test environment. Read-only to start; write access only where we agree.

Do you need production data?

Usually not. Staging with synthetic data is the default.

Our test management is a mess. Is that a problem?

That's usually the first thing I fix. The Health Check tells you where it stands.

We don't have automated tests. Can you still help?

Yes. I start with API tests, which pay off fastest, and build from there. The code stays in your repo.

Is it really just you?

Yes, plus the agent. That's the point: one accountable person instead of a team you have to manage.

How do we work across time zones?

Async by default: findings land in Jira and Slack as they come, with a written summary. Calls in a fixed overlap window: 8:00 AM–12:00 PM ET (14:00–18:00 CET), Monday to Friday — whatever time zone you're in, there's a slot that works for both of us.

What happens if we stop?

You keep everything: cases, cycles, scripts, reports. Nothing to migrate.

Fifteen minutes, and you'll know where your testing stands.

Tell me what's coming up — a release, a launch, a process that keeps slipping — and I'll tell you what I'd do first and what it would cost.

Book a 15-min call

Replies within one business day, from me.