I’ve spent more than 16 years fixing how work actually gets done inside SaaS, financial services and healthcare companies: rollouts, process redesign, big change programs, across Dublin, London, Wrocław and Toronto. One of those was rolling out new technology across a company of 18,000 people, which meant sitting through roughly 17,999 more opinions about “change management” than any one person should have to survive.
Along the way I ended up doing three jobs at once, whether the title said so or not: strategist, builder, coach. Strategist, because someone has to find the actual bottleneck before anyone touches a tool. Builder, because I’d rather ship the workaround than write a memo recommending it. Coach, because the hardest part was never the technology. It was sitting with someone who’d done a task the same way for eight years and helping them do it differently.
More recently I’ve been doing this same work with business owners, specifically around AI. Most start in the same place: a faster Google, a writing assistant, something they poke at for ideas when they’re stuck.
The shift that matters is when AI stops answering questions and starts doing the work. Not “wrote a draft I still have to fix,” but: it prepped the client meeting, checked the proposal against your own standards, pulled the background for a decision, drafted the follow-up, or let your team ask something that used to require finding you.
Once that’s happening, the real question shows up: not “what’s the next AI tool,” but “why does this still have to come back to me at all.”
That’s the job: figuring out what genuinely needs your judgment, and what can be written down, handed off, or handled by AI instead, then building it. If we work together, it’s the two of us for six weeks, not a template and a Slack channel.
I live in Victoria now with my husband and Riker, our Irish setter, who has a very reliable system for making sure nobody skips his afternoon walk.
P.S. I’m not a technical AI guru, and I’d rather you knew that upfront. I can talk about the tech when it’s relevant, but I don’t think the gap between you and a working AI system is “vector database vs. knowledge graph,” and I don’t think nudging a model’s temperature from 0.3 to 0.7 is what gets you your Friday afternoon back.
What I actually care about is why the work keeps landing on your desk in the first place.
See the investment ↓