The audit

How the audit works

A clear job with a clear end. Here's what happens, what we need from you, and where we stop and ask.

Three people going through a printed report in a bright glass-walled meeting room
  1. Step 1
    We talk

    Tell us what you’ve built

    What it does, who uses it, what it connects to and what worries you. We agree the scope, the price and how we’ll get access, in writing, before anything starts. You also confirm we have your permission to test.

  2. Step 2
    We map it

    We work out how it really fits together

    We draw every part: the app, its database, spreadsheets, connectors, email and outside services. We also list every login your system depends on and who controls it. This is often the first time anyone has seen it all on one page.

  3. Step 3
    We check it

    Security and ownership

    We look for open data, exposed secret codes, weak sign-ins, too much access, and places where one person holds everything. See everything we check.

  4. Step 4
    We test it

    Safe testing

    We try things as a stranger would, and as each kind of user would, within agreed limits. We use test accounts, and we stop at the point of proof rather than downloading your data. If a test would create a record on your live system, we ask first.

  5. Optional
    Real use

    Real-world user testing

    If you’d like, we also do the real jobs your staff and customers do, on the devices they use, and note what trips them up. More about user testing.

  6. Step 5
    We think

    What would make it better

    Having lived in your system for a while, we suggest improvements. These stay separate from the risks, so you can see what’s urgent and what’s just worth doing.

  7. Step 6
    You read it

    The report, and a conversation

    You get the report, and we go through it with you. There’s nothing to sign up to afterwards. If you want us to stay involved, we can, but the report stands on its own.

If we find something urgent at any point, such as customer data open to anyone, we tell you the same day. You won’t read about it for the first time in the report.

What a home-grown system usually looks like

Few of these systems are just an app. By the time a business relies on one, it has grown a good deal around it, and the risks tend to sit in the joins.

INSIDE YOUR BUSINESSONLINE SERVICESPEOPLEYour main systemStock, orders or accountsStaffOffice and on the roadCopying serviceRuns on a scheduleSpreadsheetFills a gap in the appYour appBuilt with an AI toolDatabaseCustomers and ordersEmail serviceSends notificationsPaymentsCard providerCustomersSign in from anywhere12345
Where the risks usually sit: 1 an all-powerful login to your main system, 2 copying that can fail silently, 3 secret keys in public code, 4 data with no access rules, 5 a spreadsheet anyone can edit.

What we need from you

  • An hour at the start, to tell us about the system.
  • Access, set up in the way we agree. You keep control of every login, and passwords never appear in our report.
  • Someone who can answer the odd question, and give the go-ahead for a test.
  • Your permission in writing. Without consent, testing can be a crime, and writing it down protects both of us. Why a tester needs permission.

How findings are rated

RatingWhat it means
RedA real risk now. Fix it first.
AmberA weakness to fix in the coming months.
GreenDone well. Keep it that way.
CheckNot confirmed yet, or one for your adviser to confirm.

Each finding is also marked tested, meaning we tried it within safe limits, or seen, meaning we looked at it. That way you always know how firm our evidence is.

Afterwards

Remove the access you gave us, and we’ll remind you. We list every change we made during the work, such as a test account, and remove what we added. Then the report is yours to use however helps you most.

Last updated 6 November 2026.

Find out where you stand.

Tell us what you’ve built. We’ll agree the scope and the price with you before anything starts.

Talk to usSee an example report