Two operating models at once
Through 2026 Scaled Agile has been releasing AI-Native SAFe alongside Core SAFe — described as a reinvented operating model for the age of AI, published iteratively rather than as a single new edition, with its own big picture, roles and event guidance.
Why it is like this
Releasing a second model in parts lets the framework move quickly without invalidating every existing certification at once. It also means two coherent answers are live simultaneously, and which one "SAFe" means depends on who is speaking.
Not to be confused with
A replacement for Core SAFe, or a finished thing. It arrived in instalments across 2026, and Core SAFe continued to receive its own article updates throughout — so an organisation can be entirely current while using neither the newest roles nor the newest events.
In practice
Before arguing about what SAFe says on anything AI-adjacent, establish which of the two models the other person is describing. They may not know there are two.
Source: What's new in SAFe? — the framework's own change log — AI-Native SAFe release series, 2026. Read on .
SAFe® and Scaled Agile Framework® are registered trade marks of Scaled Agile, Inc. ytyk.org is not affiliated with, endorsed by or licensed by Scaled Agile, Inc. These cards are independent commentary written in our own words; no SAFe text, graphic or Big Picture is reproduced here.
Predict Check yourself
Two colleagues disagree about whether a role exists in SAFe. What is the first thing to establish?
Show the answer
Whether they are describing Core SAFe or AI-Native SAFe — both were live through 2026 and the newer one introduced roles the other does not have.
Configuration and certification are both reasonable follow-ups. The two-models question comes first because it can make both of them right.