An architect presents the same architecture deck to executives, engineers and a security review board, and each audience is dissatisfied. What is wrong with the approach?

What is wrong

One artefact at one altitude cannot serve three audiences with different decisions to make.

Each audience needs different content because each is answering a different question:

• Executives ask: what does this cost, what does it enable, what is the risk, when does it deliver, and what decision do you need from me? They do not need components.
• Engineers ask: how does this actually work, what are the interfaces, what happens when it fails, and what am I expected to build? They need mechanism, and a deck without it reads as hand-waving.
• Security reviewers ask: what are the trust boundaries, where does data flow, what is authenticated where, and what is the blast radius of a compromise? They need a data-flow and trust view that neither of the others requires.

The fix

Three artefacts derived from one architecture, each complete for its audience rather than the same material at different lengths:

For executives — one page: the decision being asked for, options with costs and risks, the recommendation and the rationale, and what happens if nothing is decided. Diagrams only if they carry an argument.

For engineers — the working design: component responsibilities, interfaces, data ownership, failure behaviour, and the specific things that will be hard. Written down so it survives the meeting.

For the review board — the views they assess against: trust boundaries, data flows and classifications, authentication and authorisation at each boundary, and the threat model with its accepted risks.

The failure this prevents

Executive audiences shown component diagrams disengage, and correctly — the diagram does not help them make their decision. Engineers shown an executive summary do not know what to build, so they invent the missing detail and the architecture diverges. Reviewers shown either produce findings late, when changing the design is expensive.

The principle

Pick the altitude from the decision, not from the diagram you already have. An architect who presents one artefact to everyone is optimising for their own effort, and paying for it in three separate failures of communication.