AI EvidenceProof
Here is what we can evidence.
Named clients, third-party certification and how AI work is measured on your own results.
Certified in the Claude Partner Network Microsoft Solutions Partner ISO 27001 ISO 9001 Independent since 1999
AI clients
Working with us on AI
Businesses we are working with, each at a different stage of the route.
Managed IT
27 years of operational record
The figures are from Cardonet's managed IT service.
cardonet.ai is the AI practice of Cardonet, independent since 1999 and certified to ISO 27001 and ISO 9001. The engineers behind Cardonet's 24/7 managed IT service run this work under the same audited controls.
- 98% Customer satisfaction
- 98% SLA achievement
- 27 Years independent since 1999
Who you are hiring
You are hiring one company in three roles, with the same people throughout. More than 60 of us are engineers, and the AI work stands on 27 years of complex projects: cloud migrations, office moves that took every system with them, passwordless sign-in rolled out across a business and hotel openings. Our engineers have slept on site to see that kind of work through.
- Advisory We show you where you stand, agree the rules with your leadership and choose the tool once the rules exist. That advice arrives as Govern, the first phase of our Track rollout, with named outputs and named participants, so you can hold us to it.
- Integrator We build AI into the systems you already run, whether that is Microsoft 365, Google Workspace or the system your people live in, under sign-in you control. Before anyone's access grows we check who can see what, because a connected tool can open anything its user's account can.
- Enabler We bring your first team into daily use on the work they already do, and show them where AI helps and where to stop. From that day the team that ran your Track rollout keeps what is live running, improving and governed, with support that answers 24/7/365.
Certification
Audited by a third party
ISO 27001, ISO 9001 and Cyber Essentials, and screening for every person on the team.
- ISO 27001 and ISO 9001 Audited and independently verified against an established standard.
- Cyber Essentials Certified against the scheme's five technical controls.
- Screening Every person on the team is BS7858:2019 screened and DBS checked, before they touch your systems or your data.
Partnerships
Cardonet is Certified in the Claude Partner Network.
Microsoft Solutions Partner for Microsoft 365 Copilot.
The Claude Partner Network is how Anthropic works with the firms that put Claude into production for customers. Partners advance through Certified, Select, Preferred and Global Premier tiers, and we are at the first of them.
Escalation
What happens when it goes wrong
One escalation route, available 24/7/365, into the same team that built your Track.
- A named route to a named team The people who set up your Track are the same people you reach when something goes wrong.
- Review points are defined Adopt, the phase that brings your first team into daily use, sets out who checks what and when, and where output can stand alone. AI gets things wrong, so we treat that as a design constraint.
- Nothing widens without a check A permissions review sits between your first team and everyone else, and we do not skip it to meet a deadline.
Results
Measured on your own work
The numbers that count come from your own work, measured before you start and again after.
Cardonet has 27 years of operational record behind everything above, and AI work is held to the same standard of evidence.
Adopt ends with a test on your own work, and every build closes with Test against your own work, so the numbers that come out are yours. Before any of that, a permissions review checks who can see what before anyone's access grows. We will not show you a number nobody can verify.
What we will publish
Test results from named engagements, for every client who allows it.
- Test results We will publish the result of each Adopt test and each build's Test for every client who allows it, each from a named engagement rather than an average across a cohort.
Where to go next
If you came for the controls
Guardrails sets out what you end up holding, where the boundary sits, the permissions review in full and the list of things we do not claim.
If you want to see the method
Two phases with named outputs and named participants, a permissions review and a test on your own work to close it.
If you are not ready to talk yet
Ten questions take about two minutes and give you the two or three things worth doing first, and nothing is sent anywhere.
