The AI adoption gap

You budgeted for the agents; whether they pay off is decided at the team layer, where adoption spreads unevenly. Free on the page: the 6-question adoption pulse and the 5-point working agreement, no mandates required.

Read the playbook

Senior engineer attrition

Gallup puts replacement cost at one-half to two times annual salary, plus the knowledge that leaves, and it announces itself months ahead. Free on the page: the 6 early-warning signals and the cost-of-one-exit worksheet.

Read the playbook

Is the platform team working?

You staffed a platform team on the promise that an hour spent there saves many hours here, and nothing in the organization measures whether it did. Teams that route around it never file a ticket saying so. Free on the page: a six-question adoption pulse asked of the served teams, and four signals a platform is solving the wrong problem.

Read the playbook

The verification gap

AI agents multiplied how much code your team ships; the human capacity to confirm any of it is correct stayed flat. Verification, not generation, is now the bottleneck. Anthropic’s own Claude Code team calls it the biggest problem of shipping 8x more. Free on the page: the bad-vs-sad escape tracker and the 5-point verification working agreement.

Read the playbook

Output vs outcome

Your team ships visibly more than it did a year ago, and nobody can say which of it mattered. Output is measured continuously and outcomes are barely measured at all, because building used to be the expensive part and no longer is. Free on the page: a six-question outcome audit for your last ten shipped items, and an outcome-first definition of done.

Read the playbook

Improving DORA metrics

Your dashboard says cycle time is 4 days. It cannot tell you which practice made it 4, or what to change to make it 3: a dashboard is a speedometer, and a speedometer cannot press the pedals. Aurora Coach works on the practices behind the four DORA metrics. Free on the page: 12 example experiments, one line each, three per metric.

Read the playbook

Aligning engineering with strategy

You presented the strategy, everyone agreed, and Monday looks exactly like the Monday before. Alignment cannot be communicated into existence after a decision was taken without what the teams knew. Free on the page: six questions that surface what your teams know before the planning cycle, and the response record that turns asking into negotiating.

Read the playbook

On-call burnout and alert fatigue

The rota records who was available. It records nothing about who was woken, what the following day cost, or which engineer gets called whether or not it is their week. Free on the page: a six-question health pulse asked when a rotation ends, and an alert noise audit that turns a shared feeling into a number.

Read the playbook

Developer loneliness

Debugging together, reviewing together, thinking out loud: work that used to need a teammate now happens between one engineer and their agents. The productivity is real and so is the isolation. Problems stop surfacing early and knowledge stops moving between people. Anthropic’s own Claude Code team hit it too. Free on the page: 5 connection rituals and the 4 check-in questions that catch isolation early.

Read the playbook

AI skill atrophy

Your juniors have to become seniors, and leaning on AI skips the reps that used to get them there: the struggle through a bug, the read of unfamiliar code. That is why AI-assisted juniors score two letter grades lower on comprehension and debugging, and why the mode of use decides everything. Free on the page: the 6-point team learning agreement and 4 learning-pulse questions.

Read the playbook

Code review process for AI-generated code

Review capacity is a fixed number of human hours, and AI-generated volume went past it: queues lengthen, review depth drops, more PRs get approved without discussion. Reviewing more carefully is arithmetic that does not work; the shape of the process has to change. Free on the page: three example review-process commitments a team could adapt.

Read the playbook

Sprints that keep slipping

Your team commits to a sprint and lands about two thirds of it, sprint after sprint, so everyone plans harder. Velocity is a forecast built from history, and agent-assisted work widened the spread that history described. Free on the page: six causes of chronic slip with a diagnostic question for each, and a fifteen-minute planning retro.

Read the playbook

Nobody has talked to a customer

Your engineers build what the ticket says, accurately and fast, without having met the person who will use it, so accuracy against the ticket is the only quality they can judge. Free on the page: six questions that reveal a team building blind, and a discovery cadence agreement small enough to survive a busy quarter.

Read the playbook

Developer experience survey

You survey the team to find what slows engineers down, the results arrive as a deck, nothing anyone can point at changes, and next year fewer people answer. What fixes that is the loop after the survey, not more questions. Free on the page: the 16-question survey grounded in SPACE and the DevEx framework, with one-click copy.

Read the playbook

Squad health check (Spotify model)

Something is off with the team and nobody can quite name it. A health check turns that into a conversation the team runs on itself, so what was a feeling becomes eleven things you can point at, compare over time, and act on. Free on the page: the complete Spotify squad health check, all 11 dimensions with strong and weak descriptions, workshop instructions, and a copy button. Plus the improvement loop the model has always been missing between workshops.

Read the playbook

Measuring psychological safety

Teams can feel when it is not safe to speak up, but a feeling is not something a lead can manage, track, or show anyone. Free on the page: what Edmondson's 7 survey items measure, with the 7-point scale and scoring and one-click copy. Plus the loop that turns a one-off reading into a visible trend.

Read the playbook

Postmortem follow-through

Your postmortems produce good action items that die in the backlog until the incident recurs. Free on the page: the copyable action-item template and the follow-through checklist for the half of the process nobody solved.

Read the playbook

Retrospective action items that stick

Your team agrees on the same fix every retro and it never gets done, so people start treating the retro itself as theatre. Free on the page: 9 before-and-after rewrites and the copyable 6-field template, plus the loop that turns an action item into a commitment tracked through periods.

Read the playbook

Comprehension debt

Your team ships more code than anyone on it fully understands. Nothing breaks on merge day; the cost surfaces later, in debugging, incident response, onboarding, and the next change to code nobody wrote. Free on the page: the proxy measures that make the gap visible and three example commitments for paying it down.

Read the playbook

Many problems, one loop. Aurora Coach senses through Coaching Sessions and check-ins, analyzes, and recommends; the team refines, commits, and executes; the next period re-evaluates, with every team, every period. That loop is what continuous improvement software is for.

How does Aurora Coach help with AI-era problems specifically?

The problems are new; the method is not. Comprehension debt and review bottlenecks respond to the same loop as any ways-of-working problem: sense what is actually happening, analyze it, commit to a small change, and re-evaluate next period. Aurora Coach runs that loop with every team, every period; each use case page shows what it looks like applied to one problem. The constant across them: the work is increasingly done alone with agents, and the problems are still solved as a team.

Do I need Aurora Coach to run these experiments?

No. Each page gives you something you can run with a whiteboard and a spreadsheet. Aurora Coach turns the one-off exercise into a working system: the AI gathers context from the whole team every period, analyzes it across six domains of team effectiveness together with outside research, and recommends concrete next steps. The team decides, commitments stay owned by the team, and the next period’s analysis shows whether the change held, including the periods when delivery pressure is high.

Every playbook above is free to use as it stands. To see what the loop could do for your own team, use Map your ROI below: it asks a few short questions about your setup, then gives you a quick read on what Aurora Coach could change for you. Free, no signup, about two minutes.