The method has a formal name. We made it smaller.
Everything we do starts from an outcome and works back. That is not something we invented, and it is worth knowing where it comes from before you decide whether to trust it.
The order of the questions is the whole idea.
Business-outcome-driven enterprise architecture, usually shortened to BODEA, is Gartner's name for an approach that inverts the usual order of an architecture programme. Traditional enterprise architecture starts with how technology is built and documented. BODEA starts with why a change is wanted and what business outcome it has to produce, and only then asks how it gets implemented.
That inversion is the entire reason it works at forty people as well as at forty thousand. Almost none of the machinery of a large architecture programme survives the trip down to a small organization. The order of the questions does.
- Choose the system
- Implement it
- Discover what the work actually needed
- Negotiate with the system you already bought
- Name the outcome you need
- Work back to what the organization must be able to do
- Ask which of those abilities is actually weak
- Only then talk about systems
The approach is Gartner’s, described in its enterprise architecture research. Gartner, enterprise architecture research. What we have stripped out is the machinery that only makes sense at ten thousand employees. What we kept is the sequence above, which costs nothing to run at forty.
AI is arriving in the architecture before it arrives on an agenda.
Three findings from Gartner, all pointing the same direction. Read together they say something specific about a small organization: the technology is not the bottleneck, and it is not going to be. The bottleneck is whether anyone has decided what the thing is for and who answers for it.
Over 40% of agentic AI projects will be canceled by the end of 2027
Gartner attributes the cancellations to escalating costs, unclear business value, or inadequate risk controls. Its analyst notes that most current propositions lack significant value or return on investment.
40% of enterprise applications will feature task-specific AI agents by 2026
Up from fewer than 5% in 2025. The practical consequence for a smaller organization is that AI arrives inside software it already pays for, switched on by a vendor rather than chosen in a meeting.
By 2028, AI agents will help deliver half of enterprise architecture outcomes
Gartner's Predicts 2026 research expects agents to absorb the routine end of architecture work, leaving the human side of the discipline concentrated on strategic advice and on governing the AI itself.
Gartner, Predicts 2026: AI Agents Will Redefine How EA Works
The first of those is the one worth sitting with. Gartner does not attribute the cancellations to models that were not clever enough. It attributes them to cost, unclear business value, and inadequate risk controls, which are governance failures with a technology costume on. An organization that already starts from the outcome and works back is running the exact discipline those projects were missing.
The second matters more than it looks. If AI arrives switched on inside software you already pay for, then the decision to adopt it was never yours to make in a meeting. It was made by a vendor, and the only remaining question is whether anyone at your organization is going to notice, decide what it may touch, and write that down.
Bring your AI question to one conversation.
We will tell you honestly which parts of this apply to an organization your size, and which parts are enterprise machinery you can safely ignore.