Independent consultant / Newbury, UK
AI & workflow automation. With a job to do.
I build custom software and automation around business operations. Where AI has a useful role, I combine it with the systems, rules and human decisions the process depends on.
Independent consulting and hands-on delivery · Newbury, UK
Start with the task, the inputs and the decision.
Before choosing a model or tool, identify what a person is trying to finish. What arrives? What information do they need? What can be handled by a clear rule, and where does judgement remain necessary?
That makes it possible to decide whether the right solution is a process improvement, an ordinary integration, an AI-assisted step or a combination.
Information has to be prepared
Explore whether AI can help organise, extract or draft information before a person checks it and uses it in the next step.
Work is repeatedly copied between tools
An integration or deterministic workflow may be the useful part. Connect the systems and make ownership and failure handling clear.
A prototype needs an operational path
Define the access, review, exception handling, testing and support arrangements needed before a tool is used in live work.
A small, testable piece of the process.
Define a useful outcome
Agree the task, the people involved and how they will judge whether the output is usable.
Check inputs and access
Identify what data is needed, which systems own it, who may access it and what may be sent to an external service.
Test representative cases
Include incomplete inputs and exceptions as well as straightforward examples. Decide when the workflow should ask for help or stop.
Build in review and handover
Make responsibility for checking outputs explicit. Agree documentation, training, operating costs and support before launch.
Business operations behind the technical work.
My automation background includes reporting and sales operations processes at Vodafone Business, including IoT reporting, heads-of-sales reporting and Salesforce workflow improvements.
Today I combine that operational experience with custom software development and AI tooling. As external CTO at Kona Libre, I am also directly involved in application development, testing, release coordination and operational documentation.
A defined first engagement.
Focused discovery is £2,500 for one agreed process and up to four of my working days. We define the question and outputs before starting; a technical proof or demo is included when it helps answer that question.
A larger systems assessment starts at £6,000 for one workstream. Implementation, external specialists, model or platform charges and ongoing operations are agreed separately.
Well-defined implementation uses a £575 day rate as its planning basis. Complex workstream leadership and delivery uses £750 per own working day. The estimate includes the agreed coordination, testing, documentation and handover.
Questions to answer before automating.
Does every workflow need AI?
No. If a rule or an existing system can complete the task reliably, that may be sufficient. The choice should follow the work and its constraints.
Can the workflow act without a person checking it?
That depends on the task, permissions, possible consequences and agreed acceptance criteria. We make the review and escalation points explicit before deciding what can run automatically.
How will we know whether it helps?
Agree a baseline and a practical measure for the process, such as time spent, exceptions, rework or usable outputs. Any estimate should be tested against the actual workflow.
Can you work with what we already use?
The existing systems, contracts and access constraints are part of discovery. A custom tool, an integration or a change to the current process can each be considered.
A useful next step
What would you like the workflow to do?
Bring the process and the problem. We can use a free 30-minute first conversation to identify a sensible starting point.
No VAT added — I’m not VAT registered.
View the full engagement terms.