Free resource · Word + Excel kit

AI-powered time-saving kit: decide as a team what to do with the freed-up time.

AI saves time, but this time is almost always swallowed up by more work, and often by more intense work. This kit gives managers a method to measure the real gain, decide with the team how it is used, protect the team from intensification, and track the workload, with a 90-minute workshop, a team charter, and seven ready-to-use tools.

📄 Word, 16 pages📊 Excel Workbook, 7 tools🗓️ Verified on October 16, 2026🆓 Free

Download the kitSee the five-step method

According to a 2024 study by the Upwork Research Institute, 77% of employees using AI say it has increased their workload. A Danish study by Humlum and Vestergaard, conducted on 25,000 employees, shows that the time saved is reallocated to other tasks by approximately 80% of users, and to breaks by fewer than 10%. And according to Gartner, only 7% of organizations provide guidelines on how this time should be used.

What happens to the time saved thanks to AI? More work, or also time for recovery, building relationships, and gaining perspective? If no one asks the question, the default answer is always the same. This toolkit turns it into a team decision, made by the manager with their team, protected by a few simple rules, and reviewed monthly.

1. The paradox of time saved

Generative AI accelerates specific tasks: writing, summarizing, searching, formatting. Yet, most teams don't finish their days earlier, nor are they more rested. Researchers Aruna Ranganathan and Xingqi Maggie Ye followed 200 employees for eight months: they increased their pace, broadened their tasks, and extended their work hours into the day, often without being asked to do so.

Finding	Source
77% of employees using AI say it has increased their workload; 39% spend more time reviewing what it produces	Upwork Research Institute, July 2024
Approximately 80% reallocate the time saved to other tasks, less than 10% to breaks.	Humlum and Vestergaard, 2025
Only 7% of organizations say what to do with the time saved	Gartner, October 2025
14% of AI users suffer from mental fatigue related to monitoring the tools, with 39% more major errors.	BCG and UC Riverside, Harvard Business Review, March 2026
An interruption every 2 minutes on average, or about 275 per day	Microsoft Work Trend Index, June 2025

The question is therefore not merely an individual one. What do we do with the time saved? This choice cannot be left to each employee alone: ​​it is a managerial decision.

2. What AI really wins

The published figures vary greatly depending on the method. On average, serious studies converge on a gain of 3 to 6% of working time, or one to two hours per week per person, much more on certain tasks and among experienced users.

Étude	Measured or declared gain
Federal Reserve Bank of St. Louis, February 2025	5,4% of users' hours, or approximately 2 hours and 10 minutes per week
Humlum and Vestergaard, Denmark, 2025	2,8% of hours worked on average
UK Government Copilot Trial, 2025	26 minutes per day reported, out of more than 20,000 employees
Cegid and OpinionWay, France, September 2026	Approximately 3 hours per week declared
Workday, January 2026	Approximately 40% of the time saved is lost correcting low-quality content

Hence the first rule of the kit: calculate the net gain, once the time spent on verification and rework has been deducted. The gain is individual, but the burden is collective: without team decision-making, the time saved by one person often becomes an additional burden for another.

3. The four traps that turn gain into exhaustion

Trap	What is happening	The signal to spot
1. Silent expansion	The time saved is filled with tasks that no one has decided on.	Perimeters that are expanding without decision-making meetings
2. The high-intensity day	AI takes over the simple tasks, the ones that used to be easy; only the demanding tasks remain.	Breaks that disappear, fatigue at the end of the day
3. Multitasking and urgency	Multiple tools and agents operating in parallel, immediate responses that call for immediate responses	Employees who no longer know what they worked on
4. The verification debt	Content produced quickly and poorly proofread shifts the burden onto the recipient.	Rework of deliverables, tensions regarding quality

According to BCG, 67% of employees say that AI has taken over the simple tasks, leaving them with the more complex ones. Researchers from BCG and the University of California, Riverside, also observed that perceived productivity peaks with three tools used in parallel and drops beyond four. And "workslop"—poorly reviewed AI content—costs nearly two hours of rework per incident, according to Stanford and BetterUp.

4. What the law requires of the employer and the manager

The intensification of work related to AI is an occupational risk that the employer must prevent, and the manager is the first point of contact.

texts	What it implies
Labour Code, Article L4121-1	Safety obligation: to protect physical and mental health, to adapt measures when the situation changes
Item R4121-1	Risks, including psychosocial risks, are recorded and updated in the single document (DUERP).
Article L2242-17, 7°	The right to disconnect, negotiated or established by a charter, with training for managers
Articles L3121-60 and L3121-64	Daily rate: regular monitoring of a reasonable load and maintenance on the load
Article L2312-8	Consultation with the Works Council on the introduction of new technologies that will change working conditions

The courts are taking up the issue: according to the specialized press, the Créteil court reportedly suspended an AI-based reorganization in September 2026, notably because psychosocial risks had not been assessed and the time required for human intervention had not been factored in. The team decision, charter, and rituals in this toolkit are concrete preventative measures that should be documented with your HR department. AI and Social Dialogue Kit details the consultation of the CSE.

5. The five-step method

Temps	The goal	Binder tool
1. Measure	Knowing the team's net profit and where it goes today	Time Saved Journal, Summary
2. Decide	Distribute the time saved among the four uses.	Allocation decision
3. Protect	Establish rules that prevent escalation	Team Charter
4. Ritualize	Making the decision last over time	90-day plan
5. Follow	Identify the overload early and adjust	Load barometer, indicators

The core of the method is a grid of four uses, to be completed as a team. Every hour saved can go towards the Performance (to produce more or better on priorities), the Recovery (keeping tasks light, finishing on time, taking real breaks), the Relation (clients, mutual support, integration, feedback) or the Recoil (training, monitoring, process improvement). There is no universally good allocation: what matters is that it is decided, explicit and reviewed.

6. The team workshop and the "AI without overheating" charter

After two weeks of data entry, the team meets for 90 minutes: energy outlook, review of team figures, discussion on how AI has changed things, decision on the target allocation, choice of rules, and commitments. The kit provides a timed agenda and questions to facilitate the meeting.

The team charter is based on ten model rules, of which the team keeps five to seven: the scope does not grow silently, we count the net gain, we reread before transmitting, three tools or agents in parallel at most, no solicitation outside of agreed hours (including messages programmed with AI), we keep breathing tasks, a protected concentration period per day, urgency is justified, performance is not measured by the intensity of AI use, we talk about it every month.

7. Rituals and Indicators

Rhythm	Duration	What we do
Weekly	10 min	Load weather and "what overflowed this week?"
Each month	20 min	"Where has our saved time gone?": actual distribution versus target, barometer, adjustment
Each quarter	45 min	Review of the charter and summary to be submitted to management and HR
Each interview	10 min	Charging, disconnecting, and using the time saved with each collaborator

The workload barometer consists of five anonymous questions scored from 1 to 5; the system triggers an alert when the score falls below 3 or drops by half a point in a month. The indicators are always interpreted at the team level, never to evaluate an individual.

8. What the kit contains

The kit includes a 16-page Word guide in 13 chapters and an Excel workbook with 7 tools. Sections to be completed with your team are highlighted in yellow, both in the guide and the workbook.

Chapter 1. The paradox of time saved

Studies that show the gain is reabsorbed, and the resulting managerial question.

Chapter 2. What AI Really Wins

The published orders of magnitude, the net gain and the collective gain.

Chapter 3. The Four Traps

Silent expansion, high-intensity day, multitasking and urgency, checking debt, and their signals.

Chapter 4. What the law requires

Safety obligation, DUERP (Single Document for Risk Assessment), right to disconnect, fixed number of days worked, CSE (Social and Economic Committee) and recent case law.

Chapter 5. The Five-Step Method

Measure, decide, protect, ritualize, follow, with the grid of the four uses.

Chapter 6. The 90-minute team workshop

The timed schedule and questions to facilitate discussion.

Chapter 7. The "AI without overheating" charter

Ten model rules to adapt.

Chapter 8. The Rituals

Weekly, monthly, quarterly and individual interviews, and the manager's role as an example.

Chapter 9. Indicators

Six indicators with their alert threshold and the load barometer.

Chapter 10. Reframing an Excessive Demand

Statements made in front of his superiors, his team, and a colleague.

Chapter 11. The 90-Day Plan

The actions week by week.

Chapter 12. The Manager's Checklist

Twelve items to tick.

Chapter 13. Resources and Sources

The studies and texts cited, with their links.

9. Frequently Asked Questions

Does AI really save time?

Yes, but less than you might think, and not always where you expect it. Serious studies converge on an average of 3 to 6% of working time, or one to two hours per week per person, mostly spent on writing and emails. This saving is reduced by the time spent checking and revising, which the kit teaches you to account for.

Why does AI often increase workload?

Because the time saved is swallowed up by new tasks that no one has decided on, because AI first removes the simple tasks that provided respite, because it encourages multitasking, and because poorly proofread content shifts the workload onto the recipient. Without an explicit decision, the default response is always "more work."

What should be done with the time saved thanks to AI?

The kit suggests dividing the time among four uses within the team: performance, recovery, relationships, and reflection. There is no universally correct distribution; a commonly used starting point is 50% performance, 20% recovery, 15% relationships, and 15% reflection, to be adjusted and reviewed monthly.

Is the manager responsible for the workload related to AI?

The employer has a duty of care, which includes protecting the physical and mental health of employees and adapting their measures when the organization changes. Managers are the first point of contact: they monitor workload, report issues, and can implement team-based prevention measures, such as those in the safety kit, in conjunction with HR.

Can we measure the time saved without monitoring employees?

Yes. The kit is based on a self-reporting log kept for two weeks and an anonymous barometer, read at the team level. If you use activity measurement tools, stick to aggregated views, inform employees, and check with HR that the Works Council has been consulted.

How does this differ from the manager's kit for dealing with AI?

The AI ​​Manager's Toolkit helps kickstart AI adoption within the team: kick-off meeting, objections, usage expectations, 90-day plan. The Time Saver Toolkit comes into play once AI is adopted: it helps decide what to do with the freed-up time and avoid overuse. The two are complementary.

What does the Excel workbook contain?

Seven tools: a time saved log that calculates the net gain, an automatic summary with the actual distribution between the four uses, a distribution decision sheet, the team charter, a monthly workload barometer with alert, indicators with thresholds and a 90-day plan.

Is the kit free?

Yes. The Word document and Excel workbook are free, in exchange for a professional email address which will only be used to send you the download link. No unsolicited emails.

10. Going further

Determining the time saved assumes that AI has already been integrated into the team's workflow. To initiate this adoption, the Manager's toolkit for dealing with AI provides the kick-off meeting, answers to objections, and a 90-day plan. AI Measurement and ROI Kit helps to measure gains across the organization, the AI and Social Dialogue Kit framework for information and consultation of the CSE, and the AI charter template sets the company's rules of use. To train your managers to manage AI within their teams, see the learning path. Managers Catalyst and the AI training for managers.

Read also : AI does not reduce the workload, it increases it. et Cognitive load in the age of AI: the paradox that is exhausting your teams.

The entire series of free kits. To drive adoption: AI charter template, AI training plan template, AI Ambassadors Kit, Manager's toolkit for dealing with AI, Adoption Roadmap Kit, AI maturity grid, AI Measurement and ROI Kit, AI Governance Kit, Gift Sets Cowork, Copilot Studio kit, Claude Skills kit et Vibe coding kitFor the AI ​​project sponsor: AI Act kit, GDPR kit, Security and data kit, Shadow AI kit, AI and Social Dialogue Kit, frugal AI kit, AI Sovereignty Kit et responsible AI kitBy profession: AI HR kit, AI marketing kit, AI finance kit, AI kit for lawyers, AI kit for salespeople, AI toolkit for project managers, AI innovator kit et communicator's AI kit.

11. sources

The information cited on this page and in the kit is based in particular on the following sources, verified on October 16, 2026. The kit provides the complete list in its Resources chapter.

• Harvard Business Review: Ranganathan and Ye, AI Doesn't Reduce Work: It Intensifies It (February 2026); When Using AI Leads to Brain Fry (March 2026); AI-Generated Workslop Is Destroying Productivity (September 2025).
• Upwork Research Institute, study on workload and AI (July 2024); Humlum and Vestergaard, Large Language Models, Small Labor Market Effects (2025).
• Gartner, survey of HR directors (October 2025); Microsoft, Work Trend Index, the infinite workday (June 2025); BCG, AI at Work 2026.
• Federal Reserve Bank of St. Louis, impact of generative AI on productivity (February 2025); British government, Copilot test (June 2025); Workday (January 2026); Cegid and OpinionWay (September 2026).
• Labor Code : item L4121-1, item R4121-1, item L2242-17, item L3121-64, item L2312-8.
• INRS, the right to disconnect ANACT talking about work to improve it: the key role of managers ActuIA, The courts suspend an AI-based reorganization (September 2026).