Following our initial announcement on June 16th, where we launched AI-Native SAFe as the operating model designed for the Age of AI, we promised to guide our community through this evolution step by step. On June 30th, we took the next step with Session 2 in our virtual event series: From Outputs to Outcomes with AI-Native SAFe.
While the focus of our first session was establishing the strategic need for an AI-Native operating model, Session 2 shifted gears into execution. In an AI-assisted world, delivery is no longer the primary bottleneck. Because AI tools and autonomous agents allow us to build and ship code faster than ever before, the critical bottleneck has moved. The real challenge is ensuring that what we are building is valuable, safe, and aligned with strategic intent. In short, organizations need to shift from managing outputs to managing outcomes.
To support this goal, we released the next iteration of AI-Native SAFe guidance, circled in the image below, which can be accessed directly from the AI-Native SAFe Big Picture.
An overview of the key themes covered in session 2 is summarized below.
Moving to an Outcome-Driven Product Development Approach
Working on this new version of SAFe is a deeply collaborative effort. We were very fortunate to collaborate on this session with Mik Kersten, whom many of you know from his groundbreaking book From Project to Product. Mik has a new book coming out this July: Output to Outcome.
It is based on a very simple premise: In a world where AI can endlessly amplify outputs, organizations risk optimizing for the wrong thing. In his book, he includes the case of Cognition, the creators of Devin, the autonomous AI software engineer. Their teams develop multiple new agents and features every single week. In their weekly leadership alignment meetings, they review this work. If an output aligns to the outcomes they are trying to achieve, they pursue it; everything else is discarded. This is what “good” looks like in the Age of AI: an unprecedented pace of innovation, all while keeping humans and teams firmly in the loop.
In the early days of Agile, working software was the primary measure of progress. But when AI drastically reduces the cost of generation, simply shipping a feature is no longer a differentiator; it is merely an output. The answer isn’t to manage the output directly. We manage outcomes, and that is how we extract value from increased output. To support this paradigm shift, we have released a brand-new, standalone guidance article: Outcome-Driven Product Development in AI-Native SAFe. Central to this guidance is a cycle, shown below, that applies at every level of the Framework: Team, ART, and Portfolio. Its core mission is to accelerate the flow from desired outcome to realized value while generating rapid feedback loops.
- Outcomes: Expressed as OKRs, they range from a Portfolio’s contribution to enterprise strategy down to a Team’s intention for the upcoming PI.
- Priorities: Portfolios prioritize by directing funds to outcomes; ARTs and Teams prioritize by committing capacity to the work that achieves them.
- Outputs: For Portfolios, these are products and solutions; for ARTs and Teams, they are Features, Prototypes, Enablers, and Experiments.
- Measurements: Continuous monitoring allows us to identify winning outputs and ruthlessly pivot away from or cease everything else.
- Value: In AI-Native SAFe, value means customer outcomes achieved and the business results they drive.
AI Necessitates Changes to Investment Strategy
Traditional portfolio strategy formulation is performed annually, often taking 6–9 months. In the Age of AI, that planning cadence creates a strategy that is outdated before it is executed. AI-Native SAFe shifts portfolios to a new approach to investment strategy, placing multiple small bets, failing fast, and dynamically reallocating resources quarterly.
Furthermore, AI-Native solutions break the assumption of fixed IT overhead. Budgets are now highly variable, driven by the economics of token consumption. Context-heavy prompts, system instructions, and automated loops can scale costs far faster than a traditional budget anticipates. Portfolio leaders must now set operating-cost guardrails for token spend, separate fixed and variable costs, and continuously monitor consumption.
Steering to Success with an Inspirational Product Vision and Outcome-Driven Roadmaps
At the ART level, outcomes are no longer just a retroactive summary of what we planned; they are active inputs to the planning conversation. They provide cross-ART alignment, inform team context, and serve as the baseline for our new Sense and Respond events.
New Outcome-Driven Roadmaps help move organizations from mapping outputs to visualizing intent, milestones, releases, and incremental key results.
Crucially, this roadmap is powered by an effective Product Vision that is:
- Informed by Data: Utilizing AI-driven sentiment analysis and machine learning to synthesize massive market and customer insights in real time.
- Aligned to Strategy: Bound by portfolio outcomes and strict financial guardrails.
- Anchored to Technology: Viable and technically feasible through strong architectural input.
Translating Outcomes into Reliable Execution
How do we translate high-level outcomes into reliable execution without AI hallucination? AI needs three pillars to act as a reliable partner: Intent (Why), Specifications (What/How), and Context (Where). When these elements are connected into a unified information stack, no human has to brief an AI agent from scratch. The agent safely draws the context, policies, and goals it needs directly from the surrounding ecosystem.
This environment requires Curated Data. Data that is complete, accurate, maintained, and securely accessed. Built deliberately ahead of need on the Architectural Runway, curated data fuels every step of the Outcome-Driven Cycle.
AI-Native SAFe Virtual Event Session 2
If you missed the live broadcast of Session 2, watch the full recording below to see how to transition your organization from output volume to outcomes with true business impact.
Watch the Session 2 Recording Here
And join us for our next session, AI-Native SAFe Teams, Roles, and ARTs. To secure your spot and participate in the live Q&A, visit our registration page. This virtual event series runs over 12 weeks and culminates in a full reveal at the SAFe Summit San Diego on September 14-18.
We are excited to be on this journey with you and build the future together.
Stay SAFe,
Andrew Sales, SAFe Chief Methodologist