What this is
This wiki is a curated, continually updated summary of the lessons Shepherd Thomas is learning as we help Australian mid-tier organisations adopt AI. It is not a marketing blog and not a comprehensive encyclopaedia of AI adoption. It is a working set of notes, refined over time as we see the same dynamics repeat across engagements.
Notes fall into three kinds:
Patterns. Recurring dynamics we have seen across multiple client engagements. Patterns describe what tends to happen, why, and what it implies.
Heuristics. Short, practical rules of thumb that have earned their place through experience. They are not laws and they are not always right, but they are reliable enough to share.
Case studies. Longer narrative pieces that illustrate patterns in action. Case studies are drawn from real engagements and abstracted enough that a reader at the client firm cannot recognise their own engagement from the note body. Some are composites drawn from multiple engagements; some are single-engagement accounts with strict abstraction applied. Either way, the confidentiality test is the binding rule — see below.
Why we publish this
The wiki reflects a specific view about the current economics of knowledge. Consulting firms have historically guarded their methodologies as intellectual property because knowledge was scarce and scarcity was commercial. AI is collapsing that scarcity. The value of knowledge at rest — written up, documented, cataloguable — is on a rapid path toward zero. Protecting it is beside the point.
What retains value is something else. Call it knowledge in action: the translation of a pattern into a specific client's real situation, under their constraints, with their people, in the face of their history and politics. Knowledge in action is not a document, and it is not something an AI can reliably produce alone. It is the work of coaching and supporting clients to act appropriately in their specific context, and that is the work Shepherd Thomas sells.
Publishing the wiki is consistent with that view. If what you find here is useful to you on its own terms, take it. Where we add value is in the implementation layer, not in the documents. There is no commercial reason for us to hide any of this, so we do not.
How it is produced
The wiki is fed by a disciplined extraction process run over our working notes, meeting transcripts and engagement materials. That process produces draft notes. Every draft is reviewed and refined by a human before it is promoted to the public wiki. There is no automated path from client engagement to publication.
When a note is updated with fresh evidence, the note itself evolves rather than being replaced. The Last updated date on each page reflects the most recent substantive change. Older thinking that has been superseded is either updated in place or, where the change is material, preserved as a deprecated note linked to its replacement.
Confidentiality and what you will not find here
Our client engagements are private. That is not a marketing line; it is the foundation on which the work rests. Anything that could identify a specific client is removed before a note enters the draft pipeline.
Case studies are written with enough abstraction that a reader at the client firm cannot recognise their own engagement from the note body. Sometimes that is achieved by compositing multiple engagements; sometimes by a single-engagement account with strict abstraction on sector, size, timing and outcomes. On the rare occasion a client is named openly in a case study, it is because identification no longer poses a confidentiality risk and both the firm and the client are comfortable with the disclosure.
If you recognise your own organisation in something you read here, that is a reflection of the pattern being common — not of us writing about you specifically.
How to read the wiki
There is no required reading order. The wiki index groups notes by type; the themes index groups them by topic. Internal links between notes use standard hyperlinks. Notes should stand on their own, but many are richer in context when read alongside related ones.
If you would rather be notified as new notes are published or updated, the RSS feed is the easiest way to follow along.
Engagement
If the work described here is the kind of work you would like help with, write to us.