People remain accountable
A person remains responsible for decisions and approvals, including when software or AI assists the work.
How We Work
Understand the work. Agree the change. Check the result.
Start with the work people need to do. Use technology and AI where they make that work meaningfully better. We can help choose a change and carry out the agreed work.
We use these steps in proportion to the task. A specific support need or known project can be discussed directly, without an assessment or trial first.
Observe → Understand → Identify Friction → Design → Implement → Prove
What each step means in practice.
- 01
Observe
Talk to the people doing the work. Look at their tools and records, what gets in the way and what they would like to improve.
Typical outputNotes on the work, tools and opportunities to improve it. - 02
Understand
Follow the work from start to finish: who handles it, what information they need and what a useful change would make possible.
Typical outputA process map and the question the work needs to answer. - 03
Identify Friction
Find where work stalls, gets repeated or goes wrong. Check the causes and opportunities to simplify it, and record what still needs an answer.
Typical outputThe issues and opportunities to address first, and open questions. - 04
Design
Compare process changes, existing tools and new options, including AI where it fits. Agree what to change, who will do it and how to check it.
Typical outputA plan for the change, who will do it and how to check it. - 05
Implement
Carry out the agreed work. A project may begin with a small trial before wider use. Test the change with the people who will use it and show them how to run it.
Typical outputThe change in place, with tests and operating instructions. - 06
Prove
Compare the result with the starting point. Record the checks, open issues and who takes over. Use what the checks show to improve, expand or stop a trial.
Typical outputThe check results, open items and who takes over.
A few rules guide the work.
Keep control with the client
Accounts, credentials and work products should remain client-owned or readily transferable where practical.
Explain the recommendation
Important recommendations should trace back to what people told us, what records showed or what we tested.
Ask for only the access needed
Access should match the scope of the work and be removed, reduced, or handed back when it is no longer required.
Define success before claiming it
Where possible, agree what to measure and record the starting point before making a change.
Document the result
Write down the changes and instructions so others can use and support the result.
Implementation ownership
Agree who does what before the work begins.
We carry out the work covered by our agreement and coordinate with your staff, existing providers or specialists where they need to be involved.
KGS handles
The agreed investigation or project, coordination, documentation and checks.
The client provides
The business context, approvals and access, plus a person to take over the result.
Other providers handle
The parts covered by their products, contracts or specialist expertise.
Automation & AI in production
Use a trial to answer a practical question.
For an automation or AI idea, agree what the task should produce, who is responsible and how to check the output. The checks can show whether to improve, expand or stop the trial.
We first check whether a software setting or simpler process can do the job. The amount of testing should match the consequences of an error.
What task is being automated, and who owns the process?
What examples or rules will be used to test whether the output is good enough?
Which outputs need human review or approval?
What happens when the automation or AI is wrong or unavailable?
What measure will show whether it is helping after launch?
After implementation
Leave instructions the next person can use.
We document what changed, how to operate it and who looks after it. The handoff also records open issues and where to go for help.
The materials depend on the project. They may include code, settings, client-owned service accounts, operating instructions and contacts for support or a later transition.
Ongoing support can continue when it is useful and included in the agreement.
What needs to change?
Tell us about the support you need, the project you have in mind or a decision you need help making about technology or AI.