AI is saving you hours. Here's how to make them count at review time
Most people who used AI at work last week saved time. Did anyone notice?
Government data says most workers who used AI last week saved time, often an hour or more. But goals and reviews mostly aren't counting it, and about a third of workers hide or downplay their AI use. Here's a simple system: log your net hours weekly, say what you did with them, and get AI into your goals before review season.
By Preston KirschnerFounder, AI Career HubUpdated
- Why AI time savings go unnoticed
- A weekly AI wins log
- Honest time-saved math and the reinvestment line
- Self-review bullets and a 1:1 script
- Safety and policy checklist
- A monthly review-ready one-pager
- U.S. workers who have used AI on the job for at least one of 11 tasks.
- 56%
- U.S. Census Bureau, Aug 11, 2026 (rev. Sep 24)
- Of last-week AI users said AI saved them 1–2 hours. Another 15% said 3–4 hours, and 15% said more than 4.
- 31%
- U.S. Census Bureau, Aug 11, 2026 (rev. Sep 24)
- Workers who hide their AI use / downplay AI's help.
- 32% / 33%
- Vendor survey of 6,000 digital workers.
- Glean Work AI Index 2026, Jun 10, 2026
- Managers who require AI use and tie it to performance reviews. Fewer than 1 in 10.
- About 7%
- Vendor survey.
- Section, AI Proficiency Report, Jun 2026
- Of AI time savings lost to rework.
- ~40%
- Vendor survey of 3,200 employees who use AI.
- Workday, Jan 14, 2026
In this guide
Why your AI hours are invisible
About 56% of U.S. workers have used AI on the job for at least one of 11 tasks (U.S. Census Bureau, March 2026 survey, published August 11, 2026 and revised September 24, 2026). Among those who used it in the last week, 31% said it saved them 1–2 hours, 15% said 3–4 hours and 15% said more than 4 hours. A quarter saved less than an hour.
But the systems that shape your review mostly aren't counting it. Only 42% of organizations reflect AI expectations in goals (Betterworks, March 2026). Only about 7% of managers require AI use and tie it to performance reviews, fewer than 1 in 10 (Section, June 2026).
Meanwhile, 32% of workers hide their AI use and 33% downplay AI's help (Glean Work AI Index, June 2026). Glean's explanation: finish faster and you get "rewarded" with more work, and if the machine gets too much credit, you start to look replaceable.
Saved time also leaks away. Workday found that nearly 40% of AI time savings are lost to rework, and that 32% of companies simply increase workload with the time saved (January 2026).
None of this means you should hide your AI use. It means you need a record.
Keep a weekly AI wins log
Five minutes every Friday. Keep it in a private doc first, then share the cleaned-up version.
Each row answers one question: what did AI change this week, and what did you do with the time?
Date | Task | Baseline time (how you know) | Time with AI (incl. review + fixes) | Net hours back | What I did with the time | Outcome / evidence | My judgment call | Reusable? (shared Y/N) [Fri date] | [e.g. weekly vendor report] | [e.g. ~3 hrs, last 4 reports in my calendar] | [e.g. 1 hr: 40 min draft + 20 min checking] | [e.g. ~2 hrs] | [e.g. started the Q4 supplier analysis] | [link, metric or artifact] | [e.g. caught 2 wrong totals in the AI draft] | [Y/N]
In Google Sheets, paste into column A, then use Data → Split text to columns with a custom separator of |
Quantify the time saved honestly
Use the question government researchers use. The Census Bureau and the European Central Bank both ask, in essence: how many additional hours would you have needed to finish this work without AI? Answer it per task, not as a gut feeling.
- Get a baseline: time 2–3 runs of the task the old way, or use calendar entries and past tickets.
- Time 2–3 runs with AI, including prompting, reviewing and fixing.
- Subtract rework. Workday found nearly 40% of AI time savings get eaten by rework, so if your net number looks too good, check it again.
- Report a range and round down: "2–3 hours a week", not "3.4 hours".
- Convert to capacity: about 3 hours a week is roughly 7–8% of a 40-hour week. For reference, the ECB found the median euro-area AI user saves about 3 hours a week, about 7.7% of median working time (August 2026).
- Prefer outcome metrics over raw hours when you can: turnaround time, cycle time, error rate, volume handled.
Turn hours into outcomes with a reinvestment line
Saved time that nobody claims gets absorbed. Glean's report says low achievers' gains "get reabsorbed into more of the same work." Decide where the time goes before someone else decides for you, and tell your manager.
AI saves me about [N] hours a week on [task]. I'm putting that time into [project / deeper analysis / customer time], which should move [metric] by [date].
Talk about it without sounding replaceable
The fear behind downplaying is real. The fix is to lead with your judgment, not the tool.
- Lead with ownership: "I own the analysis. AI drafts the first pass."
- Show what you caught: the errors you spotted, the claims you verified, the draft you rejected.
- Show the multiplier: a workflow your teammates now use.
- Don't oversell either. Glean found 33% of workers exaggerate their AI skills, which backfires the moment someone asks you to explain the work.
Instead of: "AI wrote the report." Say: "I rebuilt the reporting workflow. AI drafts the first pass; I check every number and write the recommendations." Instead of: "I'm basically an AI expert now." Say: "I have three AI workflows I trust, and I can show you how I check each one."
Self-review templates
Fill in the brackets with numbers from your log. Use the tool your company approved, and name it.
Impact: Rebuilt [workflow] with [approved tool]; cut [task] from [X] to [Y] hours a week, net of review; reinvested the time in [Z], which [result]. Quality: Introduced a check for [failure mode] in AI drafts; caught [N] issues before [stakeholder] saw them. Team: Documented and shared [workflow]; now used by [N] teammates, saving the team about [M] hours a month.
The 1:1 script and a sample AI goal
Raise it mid-cycle, not the week reviews are due. Bring your log.
Getting your manager in the loop matters. In Gallup's data (as of May 2026), employees who strongly agree their manager actively supports their AI use are 8.7x as likely to strongly agree that AI has transformed how work gets done in their organization. That's a correlation, not proof of cause, but it's a good reason to start the conversation.
I've been tracking where AI saves me time. Over the last [4] weeks it's about [N] hours a week, net of checking, mostly on [tasks], and I've put that time into [project]. I'd like to make this part of my goals next cycle so we're measuring it on purpose. Here's one I drafted. Can we shape it together?
Goal: Document 3 AI workflows that save the team [N] hours a month, each with a written quality check, and share them in [team channel / wiki] by [date]. How we'll measure it: hours saved (net of review), issues caught by the checks, and how many teammates use each workflow.
Safety and policy checklist
A win that breaks policy isn't a win at review time. Check these before you log anything.
- Use the AI tools your company has approved.
- Keep confidential, customer and personal data out of public AI tools.
- Read your company's AI policy. If there isn't one, ask your manager what's okay.
- Note in your log which tool you used and what data it touched.
- Check outputs before they reach anyone else, especially numbers, names, and legal or medical claims.
Roll it up monthly into a review-ready one-pager
Once a month, turn four weekly logs into one short summary. Keep a running file, and review season takes 20 minutes instead of a weekend.
[Month] AI roll-up Net hours back: about [N] (range [low]–[high]) Where the time went: [project / outcome] Best win: [one sentence with a result] Quality: [issues caught / checks added] Team: [workflows shared / teammates using them] Next month: [one AI goal]
Make your AI hours count
Know what's out there before review season
Create a free account to save AI roles worth comparing against your current job, and get alerts when new ones post.
Looking for a team that values AI work?
Fresh AI, ML and data jobs, all posted in the last 30 days.
Browse AI jobsSources
- 1.About a Third of Workers Who Used AI in the Last Week Said They Completed Tasks One to Two Hours FasterU.S. Census Bureau (Renee Stepler) · Aug 11, 2026 (figures revised Sep 24, 2026) · Household Trends and Outlook Pulse Survey, March 2026. Time-saved figures are among workers who used AI at work in the last week.
- 2.AI adoption and the productivity promise: what workers reportEuropean Central Bank Blog · Aug 26, 2026 · Euro area, about 20,000 people across 11 countries.
- 3.Work AI Index 2026Glean Work AI Institute · Jun 10, 2026 · Vendor survey: 6,000 full-time digital workers in the US, UK and Australia.
- 4.Beyond Productivity: Measuring the Real Value of AIWorkday · Jan 14, 2026 · Vendor survey by Hanover Research: 3,200 full-time employees who use AI.
- 5.2026 State of Performance EnablementBetterworks · Mar 3, 2026 · Vendor survey of HR leaders, managers and employees.
- 6.The AI Proficiency ReportSection · Jun 2026 · Vendor survey of 5,026 U.S. knowledge workers. The report gives both 7% and 7.7%, so we say about 7%.
- 7.AI in the Workplace indicatorGallup · Data as of May 2026 · Correlational.
- 8.38% of AI users hide techniques from colleagues to protect their edgeUse.AI (press release) · Oct 1, 2026 · Vendor survey; fieldwork dates and method not disclosed.
Facts last checked Oct 20, 2026. Vendor surveys are labelled as such, and our own recommendations are marked as ours. Spot something out of date? Tell us.