AI implementation for partner-led professional services firms.
Independent consultant, based in London. I work with fractional CFOs, mid-sized law firms and professional-services teams that need working systems in production.
Who I am
I am Richard O'Rourke, an independent consultant based in London. My background is twenty-five years across engineering, energy and finance: three companies founded and led as chief executive, a £30m community-energy impact fund, business-development roles at EnerNOC and MITIE, and an early career as an R&D engineer. I now work with a small number of clients at a time on AI projects that need to reach production.
What I have learned
Most AI projects fail for a small number of reasons, and the model is rarely one of them. They fail because the problem was framed around the technology rather than the business; because the data was not ready; because the evaluation criteria were never written down; or because the team that builds the prototype is not the team that has to operate it.
I first learned this at Kinematik, the R&D-software company I co-founded in 2000. The product was patented, venture-funded and sold to US pharmaceutical companies, and it was delivered as a consultant-led product for a reason: the software was the straightforward part, and each deployment depended on whether the people who would use it had been involved in building it.
What I bring is experience of taking systems from idea to production, judgement about which problems are worth automating, and the discipline to decline the ones that are not.
How I work
Principles
Start with the problem, not the model. A large part of the work is establishing whether AI is the right answer at all.
Deliver something small and working early. A working system in two weeks is worth more than a strategy document in three months.
Define the evaluation before building. If a result cannot be measured, it cannot be released with confidence.
Plan for operation. Production is largely logging, monitoring and the human in the loop, and that is where most of the effort goes.
Engagement formats
I work in three formats:
Audits — a focused diagnostic of one workflow, team or stalled proof-of-concept. One to two weeks.
Sprints — building a working prototype or production rollout alongside your team. Four to eight weeks.
Retainers — embedded with a leadership team through a longer rollout. Monthly commitment.
I take on a few clients at a time. Pricing is by engagement and agreed in advance.
What I do not take on
Strategy work with no execution path. Where a long report is the deliverable, other firms are better placed to provide it.
Process automation presented as AI. If conventional automation is the right answer, I will say so.
Fully autonomous customer-facing agents in regulated workflows. Not at this stage.
Projects where data must remain in the UK but the budget does not allow for it. UK-resident inference is achievable, but it carries a cost.
These constraints are worth establishing at the outset of any AI engagement.
Who you would be working with
Richard O'Rourke
Founder
Twenty-five years across engineering, energy and finance: three companies founded and led as chief executive, a £30m community-energy impact fund, and an early career as an R&D engineer.
View profile
Discuss a specific project
Send a short note about the work you are considering: what you have tried, where it has stalled, and what a successful outcome would look like. I will reply personally.