About
Good AI and automation start with understanding how work flows.
Good AI and automation start with understanding how work flows between people, processes and software.
The science of workflow optimisation
The thinking
The handovers between people, processes and software are often where work breaks down.
Most workflow problems aren’t really technology problems. They’re the result of processes that have grown over time, workarounds that have become normal, and handovers where information gets lost, duplicated or stuck.
Adding another tool, AI feature or automation doesn’t necessarily fix that. Sometimes it just makes a complicated process harder to understand.
Most tools are capable enough. The reason work still drags is usually structural: an unclear handover between people, processes and software, a step nobody owns, duplicated data entry, or an automation nobody has checked since the day it was built. AI and automation are powerful tools when applied where they make sense — but only after you understand the workflow underneath.
The method
Investigate, identify the bottleneck, design, build, test, refine and document.
The method is deliberately unglamorous. It is the same seven stages on every project.
- 01InvestigateMap the workflow as it exist right now. Not the ideal version of it.
- 02Identify the bottleneckFind the exact point where work stalls or requires human intervention.
- 03Design the approachCreate a hypothesis for a single change. Simple first, automation later.
- 04Build the solutionBuild the minimal viable change inside your live workflow, written down.
- 05Test in the live workflowRun with live work. Monitor for unexpected friction or dropped context.
- 06Tweak and iterateAdjust until the friction disappears and you are confident it works.
- 07Document the outcomePublish the findings and hand over documentation to the process owners.
Why Test & Tweak exists
Understand first. Improve deliberately. Keep what works.
Test & Tweak takes a different approach: understand how the work actually happens, identify where it breaks down, and make deliberate improvements. Test the change in the real workflow, learn from what happens, and keep what works.
Because better operations aren’t built by automating everything. They’re built by understanding what matters, improving what doesn’t work, and making those improvements stick.
Working principles
- Start with how work flows between people, processes and software — then apply AI and automation where they make sense.
- Never automate a process until you understand how it actually runs.
- Test changes in the real workflow and measure before and after.
- Document every failure as openly as every win.
- Scoped, bounded engagements with documented handovers
Who runs this
Test and Tweak
Practical workflow optimisation for consultants, agencies and service businesses. Every project is scoped, tested with real work, and documented for complete ownership.
Client names, qualifications and company history are never added unless they are real and the client has agreed.