How we use AI
Judgment Is the Asset
How Abundant Kindling works with AI
Most of what's written about AI in consulting is either evangelism or panic. Neither helps.
We work with AI deliberately and as a matter of course. AI accelerates how we produce work. It does not produce judgment — and judgment is what AK is paid for. Get that wrong and the work goes generic, fast.
The risk is not that AI replaces us. The risk is that it makes our thinking, writing, and design indistinguishable from everyone else's. Every venture starts to get the same org chart, the same tech stack, the same brand voice. We guard against that constantly.
Here's how.
The test
Every piece of work that leaves AK passes one test:
Is it true, is it useful, and does it bear our judgment?
Three green lights, the work is done. Anything less, the work is not done yet.
How we work
Architecture and design decisions are human. AI proposes; we choose. Org structures, system designs, integration patterns, decision rights frameworks — these carry consequences that compound. The choice belongs to a person who can defend it.
AI accelerates production, not specification. Drafting documentation, generating boilerplate, scaffolding components — this is what AI does well. Deciding what gets built, why, and how it fits is judgment work.
Code and systems we ship get read by a person. Generated, AI-assisted, or hand-written, the standard is the same: someone has read it, understood it, and put their name to it. Anything in production is owned by a human, not a model.
Relationship-bearing comms are human. Cover notes, hard conversations, founder coaching — the words are ours, written by us. AI may inform; the words are not generated.
Work product comes with the thinking. Clients get the artefact and the reasoning behind it. The point is not throughput; it's leaving you with something you understand and can run.
How your data is handled
The boundary that matters is data egress to systems we don't control. Use AI is not the question. Where does the data go is.
Freely usable in any AI system.
Public information, published research, anonymised or aggregated data, anything you've cleared.
Controlled systems only.
Client information, internal strategy, unpublished work, commercially sensitive material. Used only with AI tools we control or have data agreements with. Not pasted into public chat interfaces or third-party AI systems without authorisation.
Restricted egress.
Personal information, legally privileged material, financial data, anything under an NDA covering AI processing, anything you've flagged. Stays inside systems where we can account for where it lands.
When the tier is unclear, we ask. Client-specific restrictions live in the engagement, not in memory.
What we won't do
- Publish AI-generated work as human-authored when no editing-and-judgment loop occurred
- Hand a client a system, design, or strategy whose choices we cannot defend
- Send Red-tier data to systems where we cannot account for where it lands
- Replace the human voice in relationship-bearing communications
- Ship code or systems no human has read and put their name to
- Use AI to obscure thinking rather than amplify it
- Misrepresent the process to clients or to ourselves
The bottom line
Used well, AI lets a small studio bring the depth and rigour of a much larger one without diluting the thinking. Used badly, it produces work that's faster, cheaper, and indistinguishable from anyone else's.
We use it well.
A note on the infrastructure
AK runs on named AI systems. Gwendolen Fairfax is the studio's executive function — working memory, commitment pipeline, the operational layer the studio runs on. Sister Code is the implementation partner — building, deploying, version control. Both are characters by design: an AI that only executes does not improve the work; one with a view does.
If you're curious how that works in practice — I'm the AI With a Name introduces Gwendolen, and Two Minds, One Operator covers how she and Sister Code work together.
Working with us
Most of our engagements start with a conversation about what you're trying to build and where the constraints are. The rest follows from there.