“Are our trains committing to outcomes — and is the AI on every team governed and paying off?”
Scaled Agile’s AI-Native SAFe moves trains from features to outcomes, puts AI on every team and adds new ways of working — the AI Value Architect, PI Outcome Planning and ART Sense and Respond. Each one assumes something most stacks don’t have: outcomes, teams, AI, spend and decisions on one live record. Dotwork is that record. Commit to outcomes from it, sense and respond on live data every iteration, see which AI each team uses, what it costs and whether it is governed — and move from Core SAFe one train at a time, with each train’s past increments kept as they were run.
Runs in
Covers
Scaled Agile’s operating model for AI
The pressure
The new model asks trains to commit to measurable outcomes in a day, check them every iteration and treat AI as a capability on every team. That only works if the outcomes, the data that measures them, and the AI doing the work are already on a record. In most companies they sit in a planning tool, a BI dashboard and a dozen personal AI accounts nobody can see.
What it looks like today
Full coverage
Outcomes, AI-native teams, the AI-native ART, the new events and the new role, each mapped to the room that runs it — beside the money, the architecture and the AI itself.
Built for AI-Native SAFe (2026) and Core SAFe 6.0
From outputs to outcomes.
Product, Builder, Domain Expert and AI.
The six pillars of the train.
Shorter events, prepared before anyone sits down.
Where the AI Value Architect works.
Core SAFe’s portfolio, carried forward.
What changes
Features planned over two days
PI and Team Outcomes committed in one, starting from last PI’s results and the current plan
Inspect & Adapt once a PI
Sense and Respond every iteration, on live outcome data
AI as a personal tool nobody can see
AI as a team capability — inventoried, governed and costed
AI value claimed in a slide
AI spend traced to the outcome it was meant to move, and counted only with evidence
The AI-native train, measured
Follow one PI Outcome down to the team outcomes behind it, the people and AI doing the work, and what that work cost — tokens, models and labor. Then follow it back up to the metric that says whether it moved. Value counts only when the link to the outcome and the evidence are both in place.
Where other tools stop
Planning tools hold the board. AI tools hold the usage. Neither knows which outcome the AI was working toward or whether it paid off. On Dotwork the outcome, the team, the AI and the spend sit on one trace.
Flexed to how you run
Scaled Agile describes AI-Native SAFe as a journey, not a restart, and Core SAFe remains supported. Dotwork lets every train be where it is.
Core and AI-native trains side by side
Run AI-Native ARTs beside Core SAFe trains on the same portfolio, budgets and strategy — nothing counted twice.
Your AI, not ours
Register whatever AI your teams use — copilots, coding agents, your own models — and see it on the same model as the work.
Your vocabulary
Keep the words your trains already use for outcomes, events and roles. Reports still line up with AI-Native SAFe.
Review depth set per train
Each team or train sets how much review its AI needs — light for low-risk tools, deeper for high-risk ones — so governance keeps pace with the work.
Moving trains across
Dotwork maps what changed and proposes how each train’s records move. You approve every change, and past increments keep the structure they were run under.
Two-day PI Planning becomes one-day PI Outcome Planning, ending in a commitment vote
Planning drafts outcomes, capacity and dependencies from the record; past PI objectives stay as they were
Inspect & Adapt and the System Demo become ART Sense and Respond, every iteration
Live outcome data and decisions each iteration, with the Inspect & Adapt history kept
PI objectives become PI Outcomes and Team Outcomes
Objectives remapped to outcomes and key results, each move proposed for approval
Agile teams become AI-Native Teams — Product, Builder, Domain Expert and AI
Team records gain the AI they use, what it costs and how it is governed
A new role: the AI Value Architect on every ART
A ready-made view of AI adoption, spend and value for each train
AI governance and ethics become a pillar of the ART
Guardrails registered on the train, with intake and review on the same model
Every framework, one model
An AI-native train runs into AI regulation, AI spend and outcome measures on day one. On Dotwork those frameworks read the same records as the train, instead of living in three more tools.
EU AI Act · NIST AI RMF · ISO 42001
The ART’s AI governance pillar, mapped to the EU AI Act, NIST AI RMF and ISO 42001 requirements it has to meet.
Explore AI RegulationFinOps for AI · token and model spend
Token and model spend metered per team, so each train’s AI has a cost and an owner.
Explore TokenomicsGoals tied to funded work
PI Outcomes and Team Outcomes as key results, measured from live data.
Explore OKRsTeams funded around products
Trains organized around products, funded around outcomes.
Explore Product Operating ModelFunding value streams, not projects
The portfolio, lean budgets and guardrails carried over from Core SAFe.
Explore Lean Portfolio ManagementDelivery flow and performance metrics
Delivery flow read from the same work, to see whether AI is speeding it up.
Explore Flow & DORAAI that runs the framework
AI-Native SAFe asks for shorter events and faster learning. The assembly that used to fill the week before each event is drafted from the model, and every recommendation stops at the person who owns the call.
Drafts candidate PI Outcomes from the strategy, last PI’s results, capacity, dependencies and the budget behind each value stream.
Shows which outcomes moved, which stalled and why, from the metrics bound to each one.
Links new PI and Team Outcomes to the strategic outcomes and metrics they belong to, and flags outcomes with no measure.
Registers the tool or agent against the team and train, with its risk level, access and owner.
Checks AI spend against the outcomes it was meant to move, and counts value only where the evidence is in place.
Tracks dependencies between teams, trains and shared AI platforms and flags those putting outcomes at risk.
Proposes how the train’s objectives, events and team records map to the new model.
Answers questions about outcomes, trains, AI use and spend, with a link to each record it used.
Where it runs in Dotwork
Dotwork does not ask you to adopt a new method. The framework becomes a view over the same model these solutions already read and write.
Govern & Deliver
“Is the work we are doing still the work we meant to fund?”
ExploreGovern & Deliver
“What AI is running, what can it access — and is it worth what it costs?”
ExploreFund & Optimize
“What does our AI actually cost — and what did it return?”
ExploreFund & Optimize
“Do we have the people we funded?”
ExploreGovern & Deliver
“Is every team working on what matters?”
ExploreGovern & Deliver
“What does this capability cost to run — and what depends on it?”
ExploreGovern & Deliver
“Can we prove our controls were actually working on what we shipped?”
ExploreGovern & Deliver
“Are we building what customers need — and is it paying off?”
ExploreHow it compares
Agile portfolio tools coordinate the events. AI and copilot tools report their own usage. AI-Native SAFe on Dotwork holds the outcomes, the teams, the AI and the money on one model — the thing the new framework assumes.
| Capability | SAFe & agile portfolio tools | AI & copilot dashboards | AI-Native SAFe on Dotwork |
|---|---|---|---|
| PI and Team Outcomes linked to strategy | Partly | ||
| Outcomes measured from live data | Partly | ||
| Sense and Respond every iteration, prepared in advance | Partly | ||
| Every AI on every team, inventoried | Partly | ||
| AI spend by team and train | Partly | ||
| AI spend traced to the outcome it moved | |||
| AI governance and guardrails per train | Partly | ||
| Core SAFe and AI-native trains on one portfolio | Partly | ||
| Move trains to AI-Native SAFe one at a time | Partly | ||
| Wired to strategy, budgets and architecture |
What an AI-native train needs underneath
AI-Native SAFe is a way of working. These are the records it runs on — and the reason a planning tool plus a chatbot can’t carry it.
Outcomes on the record
Strategic outcomes, PI Outcomes and Team Outcomes linked to each other and to the metrics that measure them — so planning starts from evidence, not a blank board.
Context the AI can read
Strategy, budgets, architecture, decisions and delivery on one model, so the AI on every team works from the same facts as the people.
AI as an inventoried capability
Every AI tool and agent registered against the team and train that uses it, with its access, cost and owner.
Guardrails and value on the same model
Governance, ethics guardrails and AI spend held beside the outcomes, so the AI Value Architect can show what works and what doesn’t.
Who it is for
“Can PI Outcome Planning start from a record, and Sense and Respond from live data?
“Which AI is each team using, what does it cost, and is it moving our outcomes?
“Which outcomes did we commit to, and which are actually moving?
“Are the PI Outcomes we voted on tied to the strategy we fund?
“How do we move trains to AI-Native SAFe without losing what Core SAFe gave us?
“Is every AI on every train registered, within its guardrails and reviewed?
Before you buy
How to get a number instead of a pitch.
1
Good answer · Spend, the team, the outcome and the metric on one trace.
Bad answer · A token report from one tool and an outcome slide from another.
2
Good answer · Yes — both on the same portfolio, with each train’s history kept as it was run.
Bad answer · Only after the whole portfolio is migrated.
3
Good answer · Nothing. It drafts outcomes, risks and options; people commit.
Bad answer · It reprioritizes the backlog automatically.
Core SAFe, too
AI-Native SAFe doesn’t replace Core SAFe, and neither does Dotwork. Run SAFe 6.0 in any configuration, keep trains on the version they run today, and move each one to AI-Native SAFe when it is ready — on the same portfolio, budgets and history.
Core SAFe 6.0, every configuration
Essential, Large Solution, Portfolio and Full SAFe — agile teams, Agile Release Trains, Solution Trains and Lean Portfolio Management.
Every event and artifact
PI Planning, ART Sync, the System Demo and Inspect & Adapt; the portfolio kanban, epics and lean business cases; lean budgets, guardrails and the architectural runway.
Earlier versions, kept as they ran
Trains still on SAFe 5.x keep their Program Increments and structure. Moving to 6.0 renames them to Planning Intervals without losing a past increment.
Mixed portfolios
Core SAFe trains, AI-Native ARTs and teams outside SAFe on one portfolio and one set of budgets, with nothing counted twice.
Try it on your train
A guided working session with Dotwork AI on your real context — how teams and trains are organized, how work moves, where AI is already in use and how decisions get made.
1
Map the way work actually flows
Use Dotwork AI to capture the teams, rituals, dependencies, funding paths, decisions and work structures that make up your operating model.
2
Leave with something you can keep
Download a readable operating model document and an OWL or RDF model your team can inspect, share and build from.
3
Go deeper when it is useful
If there is a strong fit, we shape a working Dotwork instance around your model and populate it with demo data.
More in Strategy & Delivery
Strategy & Delivery
The methods that turn strategy into funded, delivered work.
The model underneath
Steer, Fund & Optimize and Govern & Deliver all read and write the same Enterprise Operating Model. The first module you license builds the model. Every module after it adds context instead of another silo — and Dotwork AI reads all of it, stopping where the decision is yours.
See the Enterprise Operating Model© 2025-2026 Dotwork, Inc. All rights reserved.