For teams drowning in manual steps
AI and workflow automation
I find where your people are acting as the integration layer, then remove that. Practical automation with a named owner and a runbook — not a pilot that dies in six weeks.
What you get
Automation work fails for predictable reasons: no owner, no documentation, and a build nobody can modify. Each of these exists to prevent one of those.
Workflow mapping and opportunity assessment
Where the manual steps actually are, how much time each one costs, and which are worth automating. Some are not, and I will tell you which.
Automation builds
Working automations in n8n, Zapier, Power Automate, Make or Composio — built in whichever platform you already pay for, not the one I prefer.
AI use-case scoping with an ROI case
Which tasks genuinely benefit from a model and which are better served by a rule. Scoped with a number attached before anything gets built.
SOP development and standardisation
The written process behind the automation, so the work is still doable when the tool changes or the integration breaks.
Vendor-agnostic tooling evaluation
An honest comparison when you are choosing a platform, with the switching costs and the lock-in named rather than glossed over.
Handoff and runbooks
A named owner on your side, trained, with documentation covering what to do when it fails at 2am. This is the part most automation projects skip.
What this looks like in practice
Designed and operate a semi-automated outbound system for a technology advisory firm, integrating Apollo.io, AI lead scoring and message generation, GoHighLevel CRM sequencing and reporting — targeting mid-market CIOs, CTOs and CFOs across a large solution portfolio.
How the engagement runs
Every engagement runs the same three phases. You know the scope, the deliverables and the exit before I start billing.
Discovery
A working session with the people doing the work. I map the current state, find where it breaks, and tell you plainly whether you need me — and for how long.
Build
Defined scope, fixed deliverables, a milestone plan you can hold me to — documented as I go rather than written up at the end.
Handoff
SOPs, runbooks and a trained owner on your side. The goal is that it keeps working after the engagement ends.
Tools I work in
Vendor-agnostic by default. These are the platforms I can be productive in from day one rather than day thirty.
Common questions
Do I need AI, or just automation?
Usually just automation. Most of the time lost in a business goes to moving data between systems that do not talk to each other, and that is a deterministic problem with a deterministic fix. AI earns its place where the input is genuinely unstructured — summarising, classifying, drafting. I scope both and say which applies.
Which automation platform do you build in?
Whichever you already own. I work in n8n, Zapier, Power Automate, Make and Composio, and I would rather build in your existing stack than add a subscription. If none of them fit the job I will say that too.
What happens when the automation breaks after you leave?
There is a named owner on your side and a runbook covering the failure modes. That is a condition of the engagement, not an optional extra — an automation nobody can maintain is a liability I have handed you.
How do you decide what is worth automating?
Frequency multiplied by time multiplied by error cost, weighed against build and maintenance effort. Some manual steps are cheaper left manual, particularly low-volume ones with a lot of judgement in them.
Can you work with our IT and security constraints?
Yes. Most of my work has been inside organisations with real change control — semiconductor manufacturing, enterprise IT operations, managed services. Building something that cannot pass review is wasted effort.
Contact
Tell me what’s stuck.
A short call is free and usually enough for both of us to know whether there’s a fit. Bring the messy version — I don’t need it cleaned up first.