The AI-native SAFe PI Planning toolkit, in depth.

The assistant lives in a dedicated chat view with persistent threads. It uses tool calls against your actual workspace — ARTs, PIs, teams, capacity, backlog, risks and retros — so answers and suggestions reference your real plan instead of generic SAFe templates. Every tool call is visible, auditable and bounded by your SAFe role permissions.

Anywhere you create work, an AI-assist button is one click away: draft an epic from a prompt, draft a feature with WSJF inputs, generate a slate of features under an epic, or break a feature into 4–15 user stories with acceptance criteria and Fibonacci points. Every suggestion is previewed in a selection dialog — you pick which items to keep and the rest are discarded.

Approved stories don't just land in the backlog. They automatically populate the originating feature's swimlane inside the Backlog column on the breakout team's board, so the team's planning workspace is pre-staged the moment Product Management approves the AI draft.

Create as many wiki spaces as you need per organisation — engineering, product, ops, onboarding. Each wiki has nested pages, a markdown editor with live preview, full version history with diff and restore, and threaded comments on every page.

Permissions follow your SAFe role model: viewers read, members edit, admins manage. And because the AI copilot can read your wikis, the knowledge your train writes down becomes context the assistant uses when answering questions or drafting work.

Each feature carries the four WSJF inputs — Business Value, Time Criticality, Risk Reduction & Opportunity Enablement, and Job Size — editable directly in the table. The WSJF score (Cost of Delay ÷ Job Size) is recomputed on the fly and the backlog re-ranks instantly so Product Management can negotiate priority live with stakeholders.

Click any feature for a Jira-style detail view: rich-text description, acceptance criteria, linked epic, owning team, status, tags, custom fields, comments, attachments and a full activity history. Everything is keyboard-navigable and works on mobile during planning events.

Custom fields, statuses, tags and epics are inline-editable per ART, so PlanTrain adapts to your taxonomy without configuration projects, custom forks or expensive plug-ins.

Rows are teams, columns are iterations, and dedicated lanes sit at the bottom for cross-team dependencies and milestones. Drag a feature card onto an iteration cell to schedule it; add a dependency from any feature, team or directly on the dependency lane.

Dependencies are first-class tracked entities with a producing team, a consuming team, a target iteration, a status and tags. They render as cards in the dependency lane and are mirrored as a small inbound chip on the consumer team's row at the requested iteration, so both sides see the same commitment.

Filter the entire board by team, status, epic or tag, and zoom in or out for room-size displays. PI objectives sit in their own row at the top of the board so the whole train can see what success looks like.

Every participant sees a five-card vote on their device. Votes are anonymous by default and appear on the facilitator screen the instant they are cast. The aggregated histogram updates live, with per-team breakdowns so the RTE can see exactly where the doubt is.

Below 4? PlanTrain ships a guided re-vote workflow: capture the reasons for low votes, run a 15-minute adjustment, and re-vote. The full vote history is preserved across PIs so you can spot trends in confidence, predictability and team morale.

Pick from built-in templates — Start / Stop / Continue, Mad / Sad / Glad, 4Ls (Liked, Learned, Lacked, Longed for), Sailboat, Lean Coffee — or build your own by adding, removing and renaming columns. Every column is colour-coded with its own label.

Cards capture short or rich text from the team. Group similar cards by drag-and-drop, then run a vote (configurable votes per person) to surface the top topics.

Action items are first-class: assign to a person, set a due iteration and pull them automatically into the next PI's I&A workshop. Nothing falls through the cracks.

Each epic carries a hypothesis statement, lean business case fields, MVP definition, and a Continue / Pivot decision tracker. Linked features inherit the epic's strategic intent and roll status up automatically.

The portfolio view shows epic health, total invested capacity and projected vs actual outcomes — enough to support quarterly portfolio reviews without exporting to a separate BI tool.

Out-of-the-box dashboards show predictability per team, velocity and throughput, risk and dependency counts, and confidence-vote history. Every chart is filterable by PI, team, epic and tag.

Sign in with email and password or with Google. Each user belongs to one or more organisations, and roles map directly to SAFe — Owner, Admin, RTE, Product Manager, Product Owner, Scrum Master, Member and Viewer. Row-level security on every database table guarantees that data from one organisation is never visible to another, regardless of the entry point.

Enterprise controls such as SAML/OIDC SSO, SCIM provisioning, audit log, data residency and formal ISO 27001 / SOC 2 attestations are not yet part of the product — they are on the roadmap and we're transparent about it.