Patterns & Coaching

Pattern Insights analyzes your sprint or iteration data to surface trends, risks, and opportunities — first as rule-based cards, then through four AI coaching lenses.

Proactive signals

At the top of the view, AgileNotes raises signals — its proactive “look at this now” alerts — without being asked. They’re derived from the team’s AI-native memory: a risk that’s gone stale (open too long with no movement), a dependency that’s overdue, or a pattern that’s getting worse sprint over sprint. Each signal is colored by urgency (info / warn / critical) and is dismissible once you’ve acted on it. This is the difference between a dashboard you have to interrogate and one that taps you on the shoulder.

Severity levels

Scopes

Toggle between the current sprint/iteration and the whole PI. The rule engine uses the scope’s sprints plus their notes — a wider scope yields steadier averages, a narrower one reflects today. Different cards can fire under each scope; a “WIP within target” card might appear in PI View but not in the current iteration if today’s WIP is running high.

AI Analysis — the four coaching lenses

Below the rule cards, a four-card grid re-frames the same data through canonical coaching lenses:

• Systems Thinking — recurrence, feedback loops, dependency networks.
• Lean — TIMWOODS waste, with overprocessing / rework as the headline signal.
• Theory of Constraints — the workflow state holding the most accumulated ticket-days is the constraint.
• Flow — cycle-time spread, batch behavior, aging cost, WIP discipline, push vs pull.

Click Generate Coaching Lenses the first time; afterward the section persists until you click Regenerate. The four cards stream in one at a time as the AI produces them, rather than all appearing at once. Each card has a pencil — your edits to the analysis or coaching paragraph are saved alongside the AI output and survive until you choose to regenerate.

Why? — see the evidence, then 👍 / 👎 to teach it

Both the AI insight cards and the four coaching lenses are AI-native: grounded in your real data, and they learn from your reactions.

• Why? — where an insight is built from specific tickets or notes, a Why? button appears. Open it to expand a “Grounded in:” strip showing the tickets (by key) and the notes behind that card. It’s how you confirm the analysis reflects what actually happened rather than a generic guess. (Purely metric-driven cards — WIP, cycle time — have nothing to cite, so they show no Why?.)
• 👍 / 👎 — every AI insight and coaching lens carries thumbs. 👍 (“Helpful”) keeps that framing. 👎 (“Not helpful”) opens an optional box where you can say what was off, or what you’d prefer next time. That guidance is remembered two ways: it steers the very next generation, and it becomes a standing team directive the AI honors on every AI surface from then on — so a correction you make once sticks. (Even an empty 👎 still tells it to drop that framing.)
• Learned Directives — review and remove what the AI has learned from your 👎 guidance under a team’s AI settings → Learned Directives (admins and Scrum Masters).
• Pencil edits — rewrite any AI paragraph in your own words; your version is saved alongside the output and counts as feedback too.

Retro Prep — a one-page brief for the morning of your retro

Below the AI Analysis sits Retro Prep: an occasional-use brief the Scrum Master generates the morning of the retrospective. It replaces the archaeology you’d otherwise do by hand — re-reading old notes, checking whether last retro’s actions went anywhere, deciding what to bring up. Click Prepare Retro Brief and it assembles four sections:

• What happened — a grounded narrative of the sprint or iteration: the goal and whether it held, the shape of the work, notable events from your notes.
• What the data says — the metric picture in plain language, against your team’s own baseline, with values shown exactly as the app displays them.
• Open follow-through — the accountability section, stated neutrally: last retro’s actions and their fate, actions that have stayed open across iterations (with their ages), and how this window’s risks played out.
• Suggested threads — up to three discussion prompts, each tied to a specific pattern or follow-through fact and phrased as an open question directed at the system, never at a person.

Every section carries the same Why? provenance strip and 👍 / 👎 feedback as the coaching lenses, so you can check what it’s grounded in and teach it your preferences. The brief is cached per sprint until you Regenerate.

Connecting insights back to evidence

• View Related Notes on any insight jumps to the Notes view filtered to the notes that drove that pattern. (Metric-only insights — WIP, cycle time — have no related notes, so they don’t show the button.)
• In PI or ART view, a “Continue Capturing Notes” card appears until the scope has five sprints’ worth of history; each dot fills as another sprint’s data lands. At 5/5 it hides — you’ve unlocked the full pattern depth.

Ask AgileNotes — the grounded assistant

The Patterns view also has an assistant you can ask a direct question — “What should I raise in standup?”, “What’s most at risk right now?”, “What patterns are getting worse?” (it suggests these to start). It’s deliberately one question, one answer rather than a chat thread: ask, get a grounded reply, ask the next thing.

Every answer is grounded in your team’s memory and shows its citations — the exact risks (⚠️), facts (📌), patterns (🔁), signals (⚡), notes (📝), and metrics (📊) it drew on — plus a confidence label (high / medium / low) so you can see how much the data supports it. As everywhere else in AgileNotes, if it can’t ground a claim, it won’t make it.