Open practical research

First how you work now.
Then a check against VOLY.

Share one real AI coding-agent run: the task, the result, and how your team verified it.

Open practical research

First your way.
Then a check against VOLY.

First we capture how you work today: one latest task, your acceptance criteria, and the actual outcome. Then we run the same task through VOLY on identical criteria and compare the metrics.Participants receive the final benchmark first and a short verification-process diagnostic. A public profile is optional.

01Your working context

These signals let us compare similar workflows instead of mixing fundamentally different ways of working.

Which AI coding tools did you use in the last 30 days?
02One concrete run

Describe the latest instance. A past action is more useful than a forecast of what you might do.

What exactly did you ask the agent to do?
What result did you expect?
What triggered you to run the agent at that moment?
03Replay with VOLY — optional

If you want, we can run the same task through VOLY on identical criteria. Share a public repository and the task statement to enable it. This step is voluntary.

Which metrics interest you in the comparison?

The repository stays private to the research team; only aggregated metrics, not code, reach the report.

04How you defined done

‘Good’ and ‘fast’ are hard to measure. Name the observable evidence behind the acceptance decision.

Which checks actually ran?
05What actually happened

Successful and failed runs are equally useful—we study the gap between an agent report and an accepted change.

06Public profile — optional

Leave contact details for an interview or let us feature you as a participant. We never publish your name or image without separate permission.

No boxes are preselected. Publication permission can be withdrawn using the contact details in the privacy policy.

1 / 5

Want to go deeper?

The survey is the start,
not the whole project.

After the survey, you can share an anonymized case, propose a testable hypothesis, or share an expert opinion.

01

Propose a case

You do not need to disclose the repository: context, task, result, and acceptance method are enough.

02

Propose a hypothesis

Test a hypothesis with a fixed task, equal budget, and acceptance criteria set in advance.

03

Share expert opinion

We invite experts to publish their own results or an expert opinion in their field.

You may participate anonymously. Personal data and photos are published only with the participant’s separate consent.

Deeper participation

Send a proposal