Testing work with a defined scope and a report you can check.
Each engagement starts with one concrete problem and a written scope. Nothing starts, and no access is granted, until both sides have agreed what is in and out.
One failure, investigated until it can be reproduced.
Start here when
A specific bug keeps coming back, only happens sometimes, or your team has run out of time to chase it.
What you send
- The version where it happens
- What you expected and what you see
- Access to a safe environment, agreed in writing
What you receive
- Reproduction steps someone else can follow
- Findings on cause, with the evidence behind them
- A focused fix if it is feasible and agreed, with regression evidence
- Remaining gaps, stated plainly
Not included
- Open-ended debugging
- Unrelated repairs
- Any guarantee that every defect is found
A bounded release check before rollout.
Start here when
A release is due and you want agreed user journeys and API cases checked by someone outside the team that built them.
What you send
- The journeys and endpoints that matter
- A staging environment and test data
- What would block the release for you
What you receive
- Findings with expected and observed behaviour, and evidence
- Coverage notes: what was tested and what wasn’t
- A retest of agreed fixes
- A list of blocked cases and untested areas
Not included
- Exhaustive coverage
- Production monitoring
- A guarantee of a defect-free release
Staging AI-agent evaluation
For teams about to hand over an agent, or changing its model, prompts or tools. Agreed scenarios, observed against expected behaviour, with traces and a rerun.
Ledger evidence assessment
For teams whose AI-generated reports will need to be reviewed later. We look at one workflow and report what evidence is kept, what is missing, and whether Elvane Ledger would help.
How we handle access, data and tools.
- Written authorization first. We only test systems you have authorized in writing, and only in the environments named in the scope.
- Staging, not production. Unless the scope explicitly says otherwise.
- No secrets by email. Credentials and confidential data move only through a route we agree.
- AI tools by agreement. Whether and how AI tools may be used on your code or data is agreed before work starts.
Start with one problem.
Email what you’re trying to validate and roughly how big it is. We’ll tell you whether a narrow engagement fits.