Most agile teams find out they're broken the same way: a senior engineer quits, a roadmap slips two quarters in a row, or a retrospective devolves into the same five complaints for the eighth sprint running. By then, the damage is already compounding. An agility health radar is the tool that surfaces those problems while you can still fix them—often in 30 minutes, with the team itself doing the diagnosis. Done well, it turns vague frustration into a ranked list of fixable issues. Done poorly, it becomes another ceremony nobody trusts. This guide shows you how to run one that actually changes how your team works.
What is an agility health radar?
An agility health radar is a structured self-assessment that scores an agile team across multiple dimensions—typically clarity, performance, leadership, culture, and foundational practices—then visualizes the results on a radar (spider) chart. Teams use the radar to spot weak areas, prioritize improvements, and track progress over time. The original AgilityHealth Radar, developed by Sally Elatta and now maintained by Agility Insights, is the most widely adopted version, but Spotify's Squad Health Check, Atlassian's Team Health Monitor, and dozens of custom variants follow the same pattern.
The radar format matters. A bar chart hides relationships between dimensions; a radar chart instantly shows whether your team is strong in delivery but weak in leadership, or healthy on culture but starved of clarity. That visual asymmetry is where the conversation starts.
Why agile team health checks matter more in 2026
AI is rewriting what "healthy" looks like for an agile team. The 2025 DORA report found that AI adoption increases throughput but also increases delivery instability—meaning teams ship faster and break things more often. State of Agile data shows the same paradox: teams using AI tools report higher velocity and lower confidence in their own predictability. Lagging indicators like sprint burndown and velocity miss this entirely. A health check catches it.
Three shifts make team-level diagnostics non-negotiable this year:
Smaller, AI-augmented teams. When a team of four ships what eight used to, traditional capacity and team-design assumptions break. A health check exposes the gaps before they become attrition.
Ceremony fatigue. Practitioner communities are openly debating whether sprint planning, standups, and retros still earn their time. A radar tells you which ones your team finds genuinely useful and which feel like theater.
Distributed and hybrid teams. Trust, psychological safety, and shared understanding don't show up on a Jira dashboard. Health checks make them visible.
FixAgile, an Agile training and implementation framework designed for the age of AI, treats team-level health checks as the entry point to every transformation engagement. Org-level maturity models tell you the system is broken. A radar tells you which team, on which dimension, to what degree—and that's where real change starts.
The five dimensions of the agility health radar
The AgilityHealth TeamHealth Radar (v4) measures five core dimensions. Most credible alternatives map to roughly the same areas, even if they relabel them.
1. Clarity
Does the team know why it exists, who it serves, and what success looks like? Sub-dimensions include vision, mission, customer focus, roles and responsibilities, and dependencies. Teams that score low on clarity ship features nobody asked for and burn time renegotiating scope mid-sprint.
2. Performance
The outcomes the team actually delivers: time-to-market, quality, predictable delivery, responsiveness (cycle time), and value delivered. AgilityHealth's research across more than 4,600 teams found that confidence and happiness are leading indicators for these performance metrics—if confidence drops, predictable delivery follows within one to two quarters.
3. Leadership
How the Scrum Master, Product Owner, and engineering lead show up. Are they removing impediments or generating them? Do they coach or command? This dimension is where most "fake agile" problems live—management that treats agile as a tool to make teams work harder, faster, cheaper.
4. Culture
Psychological safety, trust, collaboration, conflict resolution, and continuous learning. Culture scores are the single best early-warning indicator for attrition in agile teams. When culture drops below a 3.5 average on a 5-point scale, plan for resignations.
5. Foundation
The technical and process basics: engineering practices, automated testing, environments, planning and estimating, effective meetings, and tooling. Foundation is also where AI-readiness lives in 2026—teams without solid CI/CD, test automation, and clean backlogs cannot safely absorb AI-assisted development.
A fully rounded radar across all five is the goal. A spiky radar—5/5 on Performance, 2/5 on Culture—is a signal the team is sprinting toward burnout.
How to run an agility health radar assessment
The assessment itself is simple. Running it well is not. Here's the facilitation flow that consistently produces honest scores and actionable outcomes.
Before the session (1 week out)
Pick the right cadence. Quarterly is the sweet spot for most teams. Monthly creates fatigue; annual misses the chance to course-correct. Tie it to the end of a Program Increment or release if you can.
Choose your tool. AgilityHealth, TeamRetro, Echometer, Parabol, and Atlassian's Team Health Monitor all offer ready-to-use radars. For first-time runs, a free template in TeamRetro or a simple Miro board works fine.
Send the survey 3–5 days early. Use a 5-point scale (1 = strongly disagree, 5 = strongly agree) with 4–6 questions per dimension. Anonymous responses produce more honest data, especially the first time.
Pre-read the data. Before the session, the facilitator reviews scores and identifies the two or three lowest-scoring sub-dimensions and any items with the widest score variance. Wide variance often matters more than a low average—it means the team doesn't even agree on reality.
Brief the leadership. Tell sponsors what the session is and—critically—what it isn't. It is not a performance review of the team, the Scrum Master, or the Product Owner. If leadership treats it that way, every future radar will be gamed.
During the session (90 minutes)
A 90-minute facilitated TeamHealth Retrospective format works well: 60 minutes for assessment review and 30 minutes for growth planning. Larger groups can split into sub-groups of 4–6 and converge.
Set the contract (5 min). Restate the rules: psychological safety, no blame, no leadership intervention during scoring, all conversations stay in the room.
Walk the radar (15 min). Share the visual. Point out highs, lows, and variance. Ask: Does this match how the work actually feels?
Deep-dive the lowest two dimensions (35 min). For each, run a structured conversation: What's behind this score? What's one specific example from the last sprint? What's one thing within our control?
Identify three improvement bets (20 min). Constrain the team to three. More than three guarantees nothing improves.
Define ownership and the next checkpoint (15 min). Each bet gets a single owner (not the whole team), one measurable signal, and a date 30–60 days out.
After the session
Publish the radar, the three bets, and the owners somewhere visible (a dashboard, a wiki page, a Slack channel topic).
Add the bets to the team's working agreement so they show up in standups.
Re-run the radar in 90 days. Compare. Repeat.
Scoring interpretation: what your radar actually means
Raw scores mislead beginners. Here's how experienced agile coaches read them.
Average above 4.0 across all five dimensions. Healthy team. Focus on the next problem—usually scaling or a stretch performance goal.
One dimension below 3.0, others above 3.5. Targeted intervention. Pick that dimension, ship one bet at a time, retest in 90 days.
Two or more dimensions below 3.0. Systemic issue. Stop adding new work. The problem is bigger than the team's discretion—escalate to leadership and bring in external coaching.
High variance (standard deviation > 1.0) on any single question. The team disagrees about what's true. This is almost always more important than the average. Run a focused conversation on that question before doing anything else.
Scores climbing across the board, but Culture flat or declining. Classic burnout signature. Performance gains are being paid for in trust. Slow down.
All scores improving consistently. Verify with a leading indicator (cycle time, escaped defects, NPS). If only the radar moves, you may be watching the team learn how to game the survey.
Alternative team health check tools
The AgilityHealth Radar is comprehensive, but it's not always the right starting point. Three credible alternatives cover most use cases.
Spotify squad health check
The most-copied agile team health check in the industry. Eleven attributes (Easy to release, Suitable process, Tech quality, Value, Speed, Mission, Fun, Learning, Support, Pawns or players, Health of codebase) rated red, amber, or green. Strengths: simple, fast, visual. Weaknesses: no AI-readiness dimension, light on leadership, easy to run mechanically.
Use it for: teams new to health checks, or those who want a low-overhead quarterly pulse.
Atlassian team health monitor
Eight attributes scored together as a team in a single 90-minute play. Heavily focused on shared understanding, balanced team, value and metrics, and proof of concept. Strengths: built for live facilitation, strong defaults for hybrid teams. Weaknesses: less granular than AgilityHealth.
Use it for: cross-functional teams that aren't strictly Scrum, or for teams already in the Atlassian ecosystem.
TeamRetro and Echometer
Both tools ship multiple radar templates—Spotify, AgilityHealth-style, DevOps Health, Agile Development Practices—plus customization. Echometer leans heavily into psychological metrics; TeamRetro into ceremony integration. Either is a fast path to a quarterly cadence without building anything yourself.
When to build a custom radar
Build your own when you need to measure something specific that off-the-shelf radars miss—AI-readiness, regulatory compliance, cross-team dependency health. FixAgile's assessment service includes a custom AI-readiness radar with dimensions like prompt-engineering literacy, AI tool adoption, AI-augmented review practices, and human-in-the-loop governance. Custom radars are also the only credible option for non-software teams (marketing, HR, ops) running agile.
How AI is changing agile team health checks in 2026
This is the dimension most articles ignore. AI doesn't replace the conversation—it changes what's worth talking about.
Real-time signal layered on top of self-reports. Modern tooling pulls cycle time, PR review latency, deployment frequency, escaped defect rate, and meeting load directly from Jira, GitHub, and the calendar. When the team's self-reported performance score is 4.2 but cycle time has doubled, the gap is the conversation. AI flow analytics from tools like LinearB, Swarmia, and Jellyfish surface this automatically.
Sentiment analysis on retrospective notes. AI can scan six months of retro notes and surface recurring themes the team itself stopped noticing—the same blocker mentioned 14 times, the trust language that disappeared three sprints ago. Used carefully (with the team's consent), this beats human pattern-matching.
AI-augmented facilitation. AI assistants can pre-summarize the radar, draft the three improvement bets the data implies, and write the follow-up commitments. The Scrum Master's job shifts from running the mechanics to coaching the conversation.
New dimensions to measure. AI-readiness, prompt-engineering maturity, AI-assisted code review quality, and human oversight of AI-generated work all belong on a 2026 radar. Teams that ignore them are flying blind on the fastest-growing source of both throughput and risk.
The core principle holds: AI is the best tool for surfacing signal, but the team is still the only authority on what the signal means. FixAgile's training programs are built around exactly this division of labor—AI for diagnostics, humans for judgment.
Turning health check results into action: a 30-60-90 plan
Most teams produce a beautiful radar and then do nothing with it. The fix is a forced 30-60-90 structure.
Day 1–30: Stabilize. Pick the single lowest-scoring sub-dimension. Assign one owner. Define one weekly experiment. Track in standup.
Day 31–60: Spread. Add the second bet. Begin tracking a leading indicator (cycle time, PR review latency, retrospective participation rate). Schedule a mid-cycle pulse on the two improving dimensions.
Day 61–90: Re-run. Run the full radar again. Compare. Celebrate moves of 0.5+ points. Investigate dimensions that didn't move—usually the bet was too vague or the owner was unclear.
Skip the 30-60-90 and the radar becomes decorative. Use it, and you compound improvements quarter on quarter.
Common mistakes that wreck agile team health checks
Five failure patterns show up across every engagement.
Treating the radar as a report card. The moment leadership uses scores to evaluate the Scrum Master or rank teams, every future score becomes propaganda. Protect the data ruthlessly.
Running it once. A single radar is a snapshot; the value is in the trajectory. Three data points minimum before drawing any conclusion.
No ownership on the bets. "The team" owning an action means nobody owns it. Single accountable owner per bet, every time.
Confusing low scores with low capability. A 2.0 on Foundation often means the team knows exactly what's wrong and lacks the authority to fix it. The intervention is usually upstream of the team.
Ignoring variance. A 3.5 average that hides a 1/5 and a 5/5 is a louder signal than a flat 3.0 across the team. Always look at the spread before the average.
How FixAgile uses health checks to fix broken agile
FixAgile, an Agile training and implementation framework designed for the age of AI, runs a team health check as the first 90 minutes of every transformation engagement. The radar tells us where to focus; the conversation tells us why nothing has worked so far.
From there, FixAgile's hands-on coaching, embedded workshops, and customized training tracks—for Scrum Masters, Product Owners, engineering managers, and executives—target the specific dimensions that scored lowest. Teams stuck on Foundation get technical practice training. Teams stuck on Leadership get coaching for their PO and Scrum Master. Teams stuck on Culture get psychological safety work and, often, a hard conversation with the sponsor. AI-readiness assessments layer on top to evaluate how prepared the team's processes, culture, and tooling are for AI-augmented delivery.
The pattern that works in 2026 is the same one that worked in 2014: measure honestly, focus narrowly, ship improvements, re-measure. The radar is just the cheapest, fastest way to start.
Start with one team, this quarter
If your agile transformation has stalled, your retrospectives have gone stale, or your teams are struggling to integrate AI into their workflows without breaking quality, an agility health radar is the lowest-cost diagnostic you can run. Pick one team. Run the survey this week. Hold the 90-minute session next week. Commit to three bets and a 90-day re-test. That's the entire intervention—and it's exactly what FixAgile's training programs and assessment services are built to scale across an organization once you've seen it work on a single team.
