Automation & AI
Connect your tools, reduce repeated work and give AI a practical job in your business.
Overview
Useful automation begins with understanding the process. I look at how information moves through the business, where people repeat the same tasks and where a handover gets stuck. Then we can work out what should happen automatically and where a person needs to stay involved.
That might be connecting a form to your CRM, following up on website activity or building a tool to help with documents, notes or emails. AI can be part of the answer when the task benefits from it, with the inputs, outputs and review steps agreed up front.
- Website, form and CRM connections.
- Activity-based email follow-up.
- Document, email and administration tools.
- Practical AI workflows with human review where needed.
Why I’m a good fit
I work across the website and the systems behind it. That gives me a useful view of where a process begins, what happens to the information and what someone needs to do next.
For InnoFund, I connected form submissions directly to HubSpot and built tools to understand where traffic comes from and where it goes. For My Oud Box, I built custom email funnels based on website activity and connected loyalty across online and in-person purchases.
I bring that same practical approach to AI work: start with a defined task, try it against real examples and make the behaviour understandable. The aim is a process you can use and keep an eye on as the business changes.
FAQs
Does every automation need AI?
No. Rules, API connections and existing software can handle plenty of tasks. We’ll choose an approach based on the job, the information involved and how reliably the result needs to behave.
Can this work with my existing tools?
Often, yes. We’ll check the available integrations, APIs and access permissions before committing to a solution. That helps identify what is possible and any ongoing software costs.
How do we keep control of what it does?
We agree what the workflow can do, what needs approval and how to spot a failure. For AI tasks, that includes testing representative examples and deciding where outputs need a person to review them.
Where should we start?
Bring me one task that takes too long or keeps getting repeated. We can map the current process, identify a useful first improvement and agree what success would look like before building it.
