See what engineering leaders finally get to ask about R&D productivity.

These are real questions Aidealy answers - about your engineers and your AI agents, measured as one workforce, from the work they actually ship. You ask in plain English; Aidealy investigates and answers. No dashboards to configure, no queries to write.

Worried this is just another dashboard, or a stack-ranking tool?
It's neither - here's how we think about that.

Aidealyacmecorp
Rank my R&D workforce by real productivity this quarter.

Analyzed your engineering data · 7/7 steps

Productivity Rank · last 90 days

Productivity score · by engineer

90-day window · signed, uncapped

  1. 1Sarah Wong156
  2. 2Marcus Johnson128
  3. 3Alice Chen119
  4. 4David Park89
  5. 5Jamie Chen-18

A negative score means cleanup outweighed shipped work.

  • Human
  • Net cleanup
Ask a question about your data
Illustrative example - not customer data

Why these questions

Notice what these questions have in common: every one is an aggregate question - a ranking, a trend, a correlation across your whole workforce over time. Not "who touched this file last week" (that's digging anyone can do by hand), but "rank the engineers who caused the most production breakage this quarter" - the kind of question you could never answer by hand. This is what it looks like to actually measure engineering productivity in the AI era, instead of guessing at it.

And for the first time, the list includes your AI agents. They open pull requests, review them, even merge them now - a real part of how your software gets built - so Aidealy measures them as part of one workforce, right alongside your engineers.

Most of these had no good answer until now - not because the data didn't exist, but because no one could connect it and just ask. Ask about one engineer, one team, or your whole R&D organization. Browse them by what you want to understand.

A rank is where the investigation starts - not where it ends. The why is always one more question away.

The library

Browse by what you want to understand.

22 prompts shown

Productivity & ranking

One signed score per engineer - the Aidealy Productivity Rank - built from what reaches production and its quality. See it engineer by engineer, or averaged by team. The ranking is where the investigation starts, not where it ends.

Show me the productivity rank of my R&D workforce over the last three months - and who moved up or down.
The same rank across rolling 7-, 30-, and 90-day windows, so you see direction, not just a snapshot.
Which engineers moved the product forward the most - and whose work created the most rework?
The signed rank in plain terms: real contribution is rewarded, and the cleanup someone's work caused is subtracted. A rank can even drop below zero - not as a punishment, but because that quarter the cleanup outweighed the output. The number is where you start asking why, not a verdict.
Show each team's average productivity rank over 90 days with its top engineer.
The same score, seen from higher up: each team's number is the average of its engineers' scores, with the team's top engineer named. One more question takes you from a team to the engineers in it.

Bugs & stability

Bugs are the hard case - they surface days or weeks after code reaches production. Aidealy measures how long code lasts, and traces a late-surfacing bug back to the pull request that introduced it - and to both its author and its approver.

What counts as a "critical bug"? See how we define it on the FAQ.
Rank the engineers who caused the most critical bugs in production over the last three months.
Production breakage attributed to whoever introduced it, visible as a pattern at a glance, then traceable to the exact pull request behind it.
Which engineers caused the most critical bugs last quarter - and who approved them?
The reviewer-accountability view: a bug ties to the person who wrote it and the one who reviewed and approved it - not just author blame.
Whose code lasts longest in production - and whose needs fixing soonest after it merges?
Code longevity as a quality signal: work that survives is rewarded; code corrected hours after merge is the opposite of done.
Which pull requests introduced bugs that only surfaced weeks after they merged?
Each later correction is traced back through your Git history to the pull request that introduced the line - even when it surfaces long after anyone remembers the change.

Code quality

A single 0–100 score for your actual source code - the Aidealy Code Quality Score, scored against the full Clean Code discipline - tracked as it moves over time and across every pull request.

Give me a Clean Code quality overview of my repos - and whether it improved or declined this quarter.
One quality score across your codebase, with the trend: where it's climbing, where it's slipping.
Rank our repos by how much code quality improved or declined over the last 90 days - best and worst movers.
Quality movement ranked by repo, so you know where to look first.
For code written in Cursor or Claude Code, how does AI-written code quality compare to human-written?
Pull requests are grouped by how much of their code AI wrote and compared on the same quality score - so you can answer "is the AI's code actually good?" with the score, not a gut call.
Where is test quality lagging behind production code across our teams?
Production code and test code are scored separately, as the discipline prescribes - so weak tests don't hide behind strong production code.

Delivery & review

Where work is slowing down - drawn from how long pull requests sit in review, get approved, and merge across your Git history. Review is the new AI-era bottleneck: code is written faster than it gets read.

Where is delivery slowing down - which engineers or teams have the slowest reviews and approvals?
Review-cycle bottlenecks ranked by team and person. The hold-up stops being a feeling and becomes a name and a number.
Show review-cycle times - time to first review, time to approval, time to merge - by team this quarter.
Turns a vague "reviews are slow" into the exact stage that's eating your cycle time.
Across this quarter, which reviewers most often let pull requests sit before a first response - and is it getting better or worse?
Stalled reviews surfaced as a pattern and a trend by reviewer, not a one-off callout - so you can see whether review is keeping up as code volume climbs.
How many hands-on hours went into PR #512 - and how did that split across the people who worked on it?
Hands-on hours are person-hours credited to the people who wrote the code - a bot accrues none. It's a floor, recorded from the editor, and it never feeds the rank.

AI usage & cost

What your team is actually doing with AI - the purpose behind each interaction, how much of the code is AI-assisted, and the new working patterns the AI era created - plus what your AI usage burns in tokens, by model and by engineer, all from the work itself. A token count is a fuel bill, not an achievement.

What is each team actually using AI for - debugging, planning, documentation, tests, security?
Every AI-assisted session is classified by purpose. "Are we getting value from AI?" stops being a debate and becomes a per-team, evidence-based answer.
For code written in Cursor or Claude Code, how much is AI-assisted - and is it linked to more rework after merge?
AI-vs-human authorship recorded at the source, set against longevity - so you can see whether AI-written code is holding up or coming back.
Which engineers work most effectively across several AI sessions in parallel?
A genuinely new pattern made visible: the strongest engineers run several AI sessions in parallel rather than one at a time. This is a minor signal - what ships and its quality always lead - and a window into how work is changing, not a reward for being busy.
How much are we spending on AI by model and by engineer this quarter?
Token usage and cost, tracked by model and by engineer - useful context for the bill, never a measure of worth. A token count is a fuel bill, not an achievement.
How many tokens did each PR cost us, by model?
Token cost tied to the unit of work that's actually tracked today - the pull request - so you can weigh AI spend against what each change delivered.
What did each PR cost us this week in hands-on hours and AI spend - and which bot-opened PRs took the most human hours to finish?
Effort and AI spend side by side, per PR. Bots never accrue hands-on hours - a bot has no editor - so the human who finished a bot-opened PR gets the credit; and these are person-hours, not elapsed time.
How many tokens did we spend on a given ticket, by model?Coming soonJira integration
Tying AI cost to specific tickets needs your issue tracker connected - that's on the roadmap.

Now ask your own questions.

You've seen the questions. Ask yours in the free demo, on a fictional company's data. Or talk to us and we'll get you set up to ask them of your own R&D: your engineers, your AI agents, and the code they ship together.