Organisational AI Transformation and Support

Most engineering organisations have bought AI tools. Far fewer can say whether delivery has improved as a result. That gap is where I do most of my work.

I help CTOs, CEOs and heads of engineering decide how AI should change their organisation, put governance and measurement in place, and avoid the expensive mistakes that come from treating adoption as a tooling exercise. My view, argued in Are We Rearranging AI Deck Chairs?, is that the return on AI comes from the same systems and discipline that were always the real work, applied to one more kind of contributor.

Who this is for

Technical leaders whose board or CEO is asking what AI spend is returning.
Organisations that have rolled out coding assistants and want evidence, not anecdotes.
Leaders under pressure to reduce engineering headcount because of AI, who want a considered view first.
Teams where one experienced outside opinion can change the direction.

Problems I am usually asked about

Adoption without a view of return. Licences are bought and usage is reported, but delivery is not measured. Are We Rearranging AI Deck Chairs? covers what to measure instead.
Headcount decisions. AI Is Not an Engineering Strategy sets out why writing code and engineering are not the same job, and what a board should ask before cutting.
More output, less understanding. Is Anybody Actually Reading This? looks at generated tickets, replies and pull requests that nobody has really read.
Cost and control. The Box is Open and The Real Cost of Intelligence cover pricing risk and where AI should run.
Governance. Building AI on Ungoverned Data explains why data foundations decide what AI can safely do.

What I do

Strategy, governance and economics. A usage policy people will follow, clear ownership, and an honest view of cost that does not assume today’s pricing lasts.
Team and role design. How roles, review and career paths change when AI contributes code. See The AI-Era Engineering Org in Practice.
Technical practice. Documentation, context and testing habits that make AI output trustworthy. See Helping AI Agents Remember.
Measurement. Delivery measures such as DORA, read with care, so you can see whether AI is helping. See Decoding the 2025 DORA Report.

How it starts

A free introductory call of 15 to 30 minutes, then, if it makes sense, a fixed-price audit of one to two weeks. Some clients continue on a monthly retainer. The details are on the engagements page.

Questions

How do you tell whether AI is working?

I start from delivery outcomes, not usage. DORA measures and review quality are a sensible place to begin, with an honest note about what the numbers cannot show. Vendor-reported productivity figures deserve caution, and I label evidence as strong or weak.

Is this only for engineering teams?

The work centres on engineering, but the questions usually come from the board or CEO. I help technical leaders answer them in plain terms.

Where are you based?

Australia. I work remotely with clients globally. See the contact page for how that works across time zones.

Browse all writing on AI adoption, or see how engineering leadership and operational measurement support this work.