From a week of work to a useful next step
You set a goal, spend time on it, and move on. By Friday, it can be difficult to remember which goals you finished or whether a change to your routine helped.
AI Review connects the intention you recorded before each sprint with the outcome you recorded afterward. That gives you something concrete to reflect on when planning your next sessions.
An example review
Finding
28-day window
You completed your goal less often in debug sessions than in your other sessions.
One change to try
“Keep debug blocks under 45 minutes.”
NEDIO checks the relevant sessions automatically. There is no habit checklist to maintain.
No clear pattern
28-day window
No clear pattern emerged this week.
Six comparisons were checked. None was strong enough to report, so none is shown.
Observation instead
You logged 6 sprints across 4 separate days — your steadiest week this month.
A count, not a comparison — a factual recap of what your week held.
Illustrative example using 28 days of sprint history.
If you accept the suggested change, NEDIO checks whether you follow it in the relevant sessions and tracks the outcomes you record. Once enough evidence is available, it reports whether goal completion appears to have improved.
The comparison gives you a reason to investigate your debugging routine. Task difficulty and other changes in your work still matter when interpreting the result.
How AI Review works
Record your goal and outcome
Before a sprint, write what you want to finish, choose a work mode, and set a duration. At close-out, record whether you met the goal, what you finished, and anything you want to return to. These short entries build the history your review uses.
Review patterns in your own sessions
Each week, NEDIO examines a rolling 28-day window. It compares groups of your sessions and checks whether a difference is strong enough to report. A finding needs enough sessions across separate days, and it must pass statistical checks that account for testing multiple possible patterns. At most one comparative finding is shown each week.
Choose one change to test
When there is a finding worth acting on, AI Review can suggest a specific change that NEDIO can check during your sessions: keeping a block under 45 minutes, keeping the same goal throughout a sprint, or taking a parked item as your next goal. You decide whether to accept, adjust, or leave the suggestion.
Keep a record of the result
An experiment needs enough relevant sessions before NEDIO can assess it. Depending on the evidence, the review may report that the outcome appears to have improved, that no clear improvement was detected, or that there is still too little information. Resolved experiments stay in your history, including inconclusive results.
The numbers are calculated before a language model helps phrase the result, and the wording is checked against those numbers.
NEDIO checks the relevant session records automatically, so there is no separate habit checklist to maintain.
What you get when there is no clear pattern
Most weeks will not produce a comparative finding.
Your review may instead show a factual observation, such as how many sprints you completed across how many days, or an item you have carried forward. It may simply report that no clear pattern emerged.
These quieter weeks still contribute to your session history. You can review what happened without treating an uncertain comparison as advice.
Your data and the role of AI
AI Review uses the goals, work modes, durations, and outcomes recorded in NEDIO.
What NEDIO does not read
Your code, file names, and repository names. This also applies when you use its VS Code or Cursor extension.
What the model receives
The model is handed counters and labels the product defined: counts of sprints, time windows, a work mode, a length range, and the direction and size of a difference. It writes one sentence from them. Sprint goals, “Done when” lines, reflection notes, support messages, your email address or account identifiers, and source code, file names, or repository names are never sent.
NEDIO calculates the comparison separately and checks the generated wording; a fixed template is used if validation fails.
The one exception. There is one exception. An observation about a next step you keep writing down has to quote it to be worth reading, so after the same the parking line has appeared in 3 sprints, it is sent to Anthropic to be phrased, shortened to 120 characters. You can turn this off in Settings → Email and privacy. The same observation about the same line is written from NEDIO’s own template, and the text stays with us.
Your review is personal to your account. NEDIO has no manager dashboard or shared team reports.
What the results can tell you
AI Review can show whether recorded outcomes got better during an experiment. It cannot establish that the experiment caused the improvement.
Your tasks, workload, and working conditions may also have changed.
Use each result alongside your knowledge of the work. Completing a difficult investigation and finishing a small edit represent different kinds of progress, even when both count as completed goals.
When the “work” you reviewed was a comfortable refactor instead of the bug you were avoiding, that pattern has a name: AI made procrastination look like work.
Frequently asked questions
How much history do I need?
It depends on the number and spread of your sessions and how large the differences are. NEDIO shows a readiness estimate based on your own history. A few sessions may support a factual recap before they support a comparative finding. A comparison is shown only with at least 14 qualifying sprints on 8 separate days, at least 14 days since your first close-out, and a difference that is both large enough and statistically significant (p at most 0.05, with a correction for the number of comparisons tried). Most weeks nothing clears that, and the review says so.
Which sessions count toward AI Review?
A sprint counts toward AI Review when it has a completed close-out and at least 10 minutes of focus time. Only the first 3 sprints of a calendar day count, and the review looks at the last 28 days.
How long does an experiment run?
An experiment needs at least 14 days and 8 occasions when you followed the change before a verdict is possible. After 12 weeks with too little evidence, it closes as unanswered. It may need longer than that to gather evidence.
Is AI Review included in the free plan?
No. AI Review is Pro in full: the weekly recap, observations, comparisons, one experiment at a time, and the ledger. Free accounts keep their weekly focus statistics. Those are minutes, blocks, goal outcomes, and how your work splits across modes and durations. Your recorded history is already available if you upgrade later.
What does AI Review send to a language model?
The model is handed counters and labels the product defined: counts of sprints, time windows, a work mode, a length range, and the direction and size of a difference. It writes one sentence from them. Sprint goals, “Done when” lines, reflection notes, support messages, your email address or account identifiers, and source code, file names, or repository names are never sent. There is one exception. An observation about a next step you keep writing down has to quote it to be worth reading, so after the same the parking line has appeared in 3 sprints, it is sent to Anthropic to be phrased, shortened to 120 characters.
Can I stop my text going to the model?
You can turn this off in Settings → Email and privacy. The same observation about the same line is written from NEDIO’s own template, and the text stays with us. Turning it off does not remove AI Review or the observation.
Can I use AI Review with sessions from my editor?
Yes. Sessions recorded through NEDIO in the browser, VS Code, and Cursor share the same history when you use the same account.