Evidence · strength 0 of 5 · searched to 2 Nov 2026
Does AI save time on notes and routine records?
Not enough evidence yet to say whether AI saves time on notes and routine records.
What trials found when officers and office staff wrote routine reports and meeting notes with AI: one strong trial found no clear time saved.
VerdictNot enough evidence yet
No good evidence
ImpactNot shown
We show no impact figure until at least two studies qualify.
Evidence strength0 of 5
There is not enough evidence to rate this yet.
Cost per user a year$$
Office plans list about $20–$30 a user a month (checked 2 Nov 2026). Specialist report tools are priced separately.
Pilot rating under our v0.3 rules. The rules are published on our method page but not yet independently reviewed, so ratings may change as the method is checked and new studies arrive. Our method.
What the studies found
No impact figure yet. Fewer than two studies qualify, so under our rules we do not show how big the effect is. The studies we found are listed below, with what each measured.
How sure are we?
- Starting level2 of 5, from 2 studies graded A, B or C.
- Down one: old models0 of 3 studies show they used a model released since May 2025 (the 18-month window); the rest used older models or do not say which.
- Down one: not independent2 of 3 studies were run, funded or written by the company whose AI was tested.
- Down one: mostly not peer-reviewedMost of the weight is on working papers or preprints, together with another problem above.
- Publication biasnot testable with fewer than 10 studies.
- Result0 of 5. There is not enough evidence to rate this yet.
Behind the average
None of these studies reports results for different groups of people or tasks.
How, where and why it works
Why it works
1 of 4. AI drafts the record from audio or notes, and a person edits it.
For whom and where
0 of 4. No study tests who gains.
What it takes
1 of 4. In the police trial, officers used the tool on about a third of their reports.
The full cost
0 of 4. Studies give no measured cost.
Four separate scores from 0 to 4, never added up: how much the studies tell us about why it works, for whom, what it takes and what it really costs (the EMMIE method).
What to do next
Measure your pilot →If you are trying it yourself
- Read the draft against your own notes before it is filed.
- Record which items you changed.
If you are rolling it out
- Check your records rules before use.
- Run a timed comparison from system logs.
What it takes
- Time real reports before and after, rather than asking people if they feel faster.
- Keep the author responsible for every line of the record.
What can go wrong
- The one pre-registered trial found no clear time saved across 755 police reports.
- In a Microsoft study, AI-written meeting notes covered slightly fewer key details.
The studies
| Study | Grade | Design | People | AI model (released) | Company role |
|---|---|---|---|---|---|
| No man's hand: artificial intelligence does not improve police report writing speed Ian T. Adams, Matt Barter, Kyle McLean, Hunter M. Boehme, Irick A. Geary Jr., 2024. Journal of Experimental Criminology 22:137-154 (journal) | A | field RCT, deployed work | 85 patrol officers randomised (43 control, 42 AI) at Manchester Police Department (NH); 755 reports in the pre-registered analysis; difference-in-differences check on 6,084 reports over one year | Axon Draft One (drafts a report narrative from body-camera audio). Paper says it is built on ChatGPT 4; a footnote says Axon reports the model in use is 'GPT4-mini' (Mar 2023) | none |
| Early LLM-based Tools for Enterprise Information Workers Likely Provide Meaningful Boosts to Productivity: Copilot in Teams Meeting Study (MSR-TR-2023-43, first Microsoft AI and Productivity report) Alexia Cambon, Brent Hecht, Benjamin Edelman, Donald Ngwe, Sonia Jaffe, Amy Heger, Mihaela Vorvoreanu and others (study authors: Amy Heger, Mihaela Vorvoreanu, Alexia Cambon), 2023. Microsoft Research technical report, December 2023 (working paper) | C | other quasi-experiment, realistic task | 57 Microsoft employees: 33 with Copilot access, 24 without. Access was rolled out division by division; assignment was not random | Microsoft 365 Copilot in Teams (early version; underlying model not named) | authors employed by vendor: Authors are Microsoft employees evaluating Microsoft's own product on Microsoft's own staff (vendor-run) |
| Generative AI in Real-World Workplaces: Impact of Copilot on Cognitive Load (Kleiner, Meijer, Rotella, Presson study) Report eds. Sonia Jaffe, Neha Parikh Shah, Jenna Butler, Alex Farach, Alexia Cambon, Brent Hecht, Michael Schwarz, Jaime Teevan; study by Madeline Kleiner, Max Meijer, Katie Rotella, Nora Presson, 2024. Microsoft Research technical report 'The Second Microsoft Report on AI and Productivity Research', July 2024 (working paper) | D | other quasi-experiment, realistic task | 40 Microsoft employee volunteers, half with Copilot and half without (how they were split is not stated) | Microsoft 365 Copilot in Word (version and model not stated) | authors employed by vendor: Microsoft researchers testing Microsoft Copilot on Microsoft staff (vendor-run) |
Grades: A, a well-run randomised trial in real work; B, a strong natural experiment in real work or a well-run trial on realistic tasks; C, a trial with clear limits or a matched comparison; D, before-and-after, surveys or case studies. How studies are graded.
Studies we found but do not count
- Writing at the speed of hype: officers' post-experimental perceptions of AI report writing (Boehme et al. (2025), grade D). A survey after the Manchester trial counted through Adams et al., so not a separate study. Its abstract and table disagree on how many officers saved time (48% and 29%).
- Boosting police productivity with AI? An assessment of productivity and efficiency gains in report writing (Alda and Salazar-Tobar (2026), grade D). A re-analysis of the Manchester trial counted through Adams et al., and only the abstract could be read.
These studies are real and often cited, but they fall outside our rules for this rating. What counts as a study.
Does the rating hold up?
- Run without company-run studies1 studies: strength 0, verdict no good evidence, headline band no clear effect.
- Run without working papers and preprints1 studies: strength 0, verdict no good evidence, headline band no clear effect.
How up to date is this evidence?
Method →Evidence searched to 2 Nov 2026. Studies cover models released Mar 2023 to Mar 2023. Rating computed 5 Nov 2026 under Module 15 v0.3 (pilot).
How this page was made: AI research agents searched for studies and read each paper; a second pass checked every number on this page against the paper itself, with the quote and where it appears. The rating is computed by fixed rules, not chosen. No person outside AI Adoption has reviewed it yet. Editorial owner: AI Adoption editors. Spot a mistake? Corrections.
Disclosure: AI Adoption uses AI tools, including Claude from Anthropic, in its research and editorial work. No company pays for or has any say in a rating.