Say what is actually true
Estimates are ranges with stated assumptions. Risks are raised while they are still cheap to address. If something is not going well, the client hears it from us first.
Values
Six commitments, written to be checkable
Values are only useful if someone can tell when they have been broken. These are written so a client could hold us to them.
What we commit to
Six values, and what breaking each one would look like
Design for the people who stay
The measure of a good solution is whether the organisation can operate and change it after we leave. That shapes our bias towards configuration, documentation and enablement.
Start with the business, not the product
Product capability is a means. We begin with the decision, process or obligation that needs to change, and select technology against that.
Govern early, not after an incident
Security, privacy and platform governance are design inputs. Retrofitting them costs more and produces controls that are stricter than they needed to be.
Prove it before scaling it
Pilots exist to produce evidence, not enthusiasm. We define what would count as success, and what would count as a reason to stop, before we build.
Leave the estate better than we found it
Content cleaned up, permissions reviewed, decisions recorded, teams enabled. The value of an engagement should outlast its scope.
What this means in practice
If an estimate turns out to be wrong, you hear it in the week we know rather than at the end of the phase. If a design decision was ours and it did not work, we say so and pay for the correction where that is fair. If a request is outside what we can do well, we say that instead of learning at your expense.