Do You Need To Master Core SAFe to Go AI-Native SAFe?

This week, Scaled Agile closed out the AI-Native SAFe webinar series, with Andrew Sales spending the whole session on the brand-new AI-Native SAFe Implementation Roadmap. As with every post in this series, this isn't a comprehensive summary, just my take on the key messages. The story so far: what AI-Native SAFe means for your organisation, from outputs to outcomes, AI-Native ARTs, teams and roles, PI Outcome Planning and Sense and Respond, and scaling innovation.

The new roadmap organises the change journey into five outcome-driven phases: Align to Outcomes, Set Priorities, Launching AI-Native ARTs, Measure Performance, and Amplify ROI. Anyone who has worked with the Core SAFe Implementation Roadmap will recognise plenty of it. The focus of this post is, of course, the differences. 

A Roadmap That Assumes a Foundation

The first four steps of the roadmap, spanning Align to Outcomes to Set Priorities, are coloured grey. Andrew explained that this recognises that organisations that have already implemented Core SAFe will have done some of this work. This is further reinforced by the Achieving AI-Empowered Agility article, now titled The Four Critical Lean-Agile Shifts in the Age of AI, which explicitly states that AI-Native SAFe is "built on the foundation of Core SAFe". While I anticipated this and it makes complete sense for an organisation using SAFe, I am curious about how a non-SAFe organisation would navigate this roadmap: building the lean-agile foundation and the AI-native operating model in a single journey. Nothing in the guidance says they can't. As far as I know, nobody has watched it happen yet, but I’m excited to give it a go!

Andrew noted that organisations already using SAFe will likely revisit some of the grey steps as they move to AI-Native SAFe. Specifically, he called out the need to address AI governance and ethics, and the opportunity to validate that your ARTs are genuinely organised around products. I expect many organisations will benefit from revisiting the shape of their existing ARTs. While Core SAFe advocates the periodic refresh of value streams and ARTs, not all organisations heed this advice. Making this opportunity explicit when implementing AI-Native SAFe should help organisations shake off the inertia that sets in after many years of successful framework application.

An Outcome-Driven Approach to Change

Andrew shared that Scaled Agile had applied the same outcome-over-output thinking that sits at the heart of AI-Native SAFe to the change journey itself. He warned that organisations can fall into the trap of "experimenting with everything, everywhere, all at once" and counting ART Launches (an output metric) rather than outcomes. He prescribed transformation OKRs as the antidote. This is not new thinking; it is Core SAFe's transformation OKR approach carried forward. In roadmap terms, the OKRs are set in the Align to Outcomes phase and come back in Measure and Grow, near the end of the journey. What was new was the expectation that key results be measured daily on executive dashboards or, if not practical, at least quarterly.

Launching AI-Native ARTs

The twelve critical success factors of AI-Native SAFe are also new: three elements inside each of the four Lean-Agile shifts for the Age of AI. This is a familiar SAFe pattern: a numbered list of success factors with a self-assessment to score yourself against. (Some of you might recall the Essential SAFe self-assessments used in SAFe 4 and 5 and the Ten Critical ART Success Factors). The 12 critical success factors of AI-Native SAFe map cleanly to the journey we have been on through the AI-Native SAFe webinar series. 

These 12 success factors provide the backbone for a new two-day course, Launching AI-Native ARTs, scheduled for release on 29 September. Andrew described the objective of the training as bringing the right people into one room, working through the AI-Native ART Launch Readiness Assessment as a gap analysis, and converging on a launch plan. The article is specific about who the right people are: attendees "should have both the necessary breadth of knowledge and the leadership authority to identify all critical activities that need to be planned." 

The article states that the course combines training and workshop elements: attendees learn the four Lean-Agile shifts and the 12 success factors, then apply them to their own ART. The resulting launch plan covers what must be in place before the first PI Outcome Planning event and before the first Sense and Respond event, and what can wait until after both.

The AI-Native launch readiness assessment resembles the SAFe DevOps Health Radar, which underpins the SAFe DevOps course, leading me to anticipate that a similar format will be used for the Launching AI-Native ARTs course. I’m super curious to see the assessment, so I will be attending the inaugural run of the course at the SAFe Summit in San Diego. Perhaps I will see some of you there!

Measure Performance and Amplify ROI

The Measure and Grow step of the roadmap recommends the same RTE and SAFe Advanced Scrum Master (SASM) training we see in Core SAFe. This makes sense: as Andrew mentioned, the focus of these courses is building high-performing teams and ARTs, which is a requirement of both Core SAFe and AI-Native SAFe. What puzzled me was the inclusion of SASM, given there was no Scrum Master role mentioned in the AI-Native team content shared back in session 3. I will be interested to see how this evolves.

The Scale AI Capabilities step of the roadmap suggests establishing an AI Council and evolving the LACE to include AI. The AI Council is a cross-functional executive body spanning IT, legal, compliance, and finance that sets enterprise AI policy and assumes accountability for ethical, legal, and financial risks. My field experience runs the other way, with organisations standing up this kind of function early, sometimes before anything else, as AI usage policies are often the first hurdle for organisations leaning into the Age of AI. The article doesn't say why the council sits where it does, and I am genuinely curious whether this is a sequence or just an observation about where councils tend to appear. The evolution of the LACE to take on AI responsibilities and act as an AI Centre of Excellence is logical and reflects what we are seeing in the field. 

What Does AI-Native SAFe Mean for Current Scaled Agile Courses?

I had been wondering what AI-Native SAFe would mean for both the existing AI-Native training and SAFe training suites (and I don't think I'm alone in that regard!). Webinar 6 provided the answers we have been waiting for. There are two separate streams of training that support AI-Native SAFe: the AI-Native courses (hands-on AI content taught by certified AI-Native trainers) and the AI-Native SAFe courses (the operating model taught by AI-Native SAFe SPCs). 

The three AI-Native courses listed on the roadmap are available today: Leading the AI-Native Organization, AI-Native Value Architect, and AI-Native Foundations. The AI-Native SAFe courses are yet to be released. 

In addition to the Launching AI-Native ARTs course mentioned above, Andrew announced three new courses: Implementing AI-Native SAFe, AI-Native SAFe Fundamentals, and AI-Native SAFe for Teams, all coming in the next few months. What wasn't clear to me is whether these are brand-new courses or an evolution of the existing ones. My read is that they are new and separate, as the AI-Native SAFe Implementation Roadmap uses the current LPM, RTE, SASM, and Advanced SPC (ASPC) badges. But that is my guess. I haven't seen anything from Scaled Agile confirming it either way. What is clear is that the current LPM, RTE, SASM and ASPC courses are all part of AI-Native SAFe. 

One more thing on training. Andrew was explicit that the roadmap deliberately doesn't show every course Scaled Agile offers. I understand and appreciate the intent, but I also worry about it. In my experience, the classes and workshops that appear on the roadmap have always had more demand than the ones that don't (remembering, of course, that correlation is not causation). If it isn't on the map, organisations tend to skip it, whether they need it or not, and this is how we end up with train wrecks! ;-)

How Do You Get Up to Speed on AI-Native SAFe?

If you are trying to work out what all of this means for your own upskilling, the upgrade path for active SAFe Studio members went live on 25 August: free, on-demand, with an expert path for SPCs, ASPCs and SPCTs and a certified member path for every other SAFe certification, any type, any version. Even without a certification, members of the SAFe community can take the upgrade and earn an AI-Native SAFe Ready badge.

As for me, I need to stop blogging and go do the upgrade path!

Stay Tuned

That wraps my blog series on Scaled Agile’s AI-Native SAFe webinars. The message continues to get clearer: AI-Native SAFe is an evolution of a lean-agile operating model, and the base assumption is that you have one. The full reveal comes at the SAFe and AI Summit in San Diego, 14 to 18 September. I'll be there, and I’ll be sure to share what I learn.