davidfacer.com / aimaturitymodels.com / AI-Native Maturity Models / AI-Native SDLC / D2

Buyer & user persona development

Almost every product team has a persona deck somewhere. Far fewer can say when it was last true. Personas rot quietly — usage patterns shift, buyers change, and the static document keeps being cited long after it stopped matching anyone real.

This dimension asks a sharper question than "do you have personas": does your organization have a living representation of its users and buyers, or a snapshot from whenever someone last had time to update the slide deck?

Where most organizations start (Nascent)

Personas begin as static documents, made once — at launch, at a redesign — and rarely revisited. Buyer personas often live separately from user personas, owned by different teams that never reconcile them. The first real step is establishing one shared structure with clear ownership and a minimum update cadence, using a consistent AI-assisted research method instead of whatever an individual happens to prefer.

Where the real gains happen (Modeled → Integral)

The meaningful shift isn't refreshing personas more often on a schedule — it's replacing scheduled refreshes and document distribution with a live, shared intelligence system: personas connected to ongoing usage, research, and feedback signals, with common underlying data and AI intelligence that Product, Marketing, Sales, and Engineering can all draw from directly, rather than everyone interpreting a document their own way. From there, the next real gain is making that system the thing functions directly consume — not a document Product hands off, but a shared asset Marketing and Sales draw from themselves, with Engineering connected through the requirements system.

What the top of the curve actually looks like (Telemetric)

At full maturity, persona intelligence converges with market discovery (D1) and positioning (D3) into one continuously updated layer. A meaningful shift in who the buyer or user actually is propagates automatically into roadmap input, messaging, battlecards, and strategy narrative — each rendered for its audience, all traceable to the same underlying representation, none of it manually reformulated.

Why this dimension matters

Personas are the organization's working model of who this is actually for — get that stale, and every downstream decision (what to build, how to position it, who to sell it to) inherits the staleness silently, without anyone necessarily noticing which decision went wrong or why.


Drafted from the ratified Shared Intelligence Layer source, including the reconciled Level B revision (2026-07-23): AI-assisted persona research and synthesis is now a consistent method at this level, not an ad hoc one — reasoned from coherence with D1-B and D3-B, which already used identical language at that letter. D2 carries no open review flag.

Drafted from the SDLC model’s real locked content.

Feed This to Your AI© 2026 David Facer