AI Workload Creep: Your Team Saves 19% of Its Week With AI. Who Decides Where That Time Goes?

By Nora, Chief of Staff, HUMAN HIVE

Where does the time AI saves your team actually go? Mostly back into more work: a new LSE and Protiviti survey of 3,001 professionals finds AI users save 19% of their working week and put 72% of it straight back into other work. The fix is a leadership decision about what that time is for, and training that makes the savings real.

The 60-second version

  • AI users save about a fifth of their week, but most of that time quietly turns into more work.

  • Trained employees save 22% of their hours; untrained employees save 13%.

  • This week, decide out loud with your team what one saved hour each week is for.

What happened

On November 17, 2026, the Inclusion Initiative at the London School of Economics and the consulting firm Protiviti released their third annual report, "Beyond the Technology: Who Really Captures AI's Gains at Work". It is based on a survey of 3,001 professional employees worldwide, run from March to May 2026.

The release has the headline numbers. 74% of respondents now use AI at work, up from 29% in 2024. AI users report saving an average of 19% of their working week, and they reinvest 72% of that saved time in other work. Half of those who saved time used it to take on new responsibilities or projects. And 48% worry about being judged for not using AI enough, compared with 20% who worry about being judged for using it.

The researchers have a name for the pattern: "workload creep." Silicon Republic's coverage quotes LSE's Dr Grace Lordan: "Without that direction, greater efficiency may simply raise the baseline for what employees are expected to deliver."

Why it happened: the human side

Nobody decided this. That is the problem. When a tool gives people back an hour and no leader says what the hour is for, the inbox decides. Lordan put it plainly: "Employees feel pressure to show they are keeping pace with AI while also taking on more."

Training is the other half. In the survey, 93% of trained employees use AI, against 50% of untrained ones. Trained people save 22% of their working hours, untrained people 13%. Protiviti's page adds that people with standardised AI workflows report 27% time savings, against 14% for those without. And the training is uneven: 66% of Gen Z employees got some AI training in the past year, compared with 40% of baby boomers.

Here is the line we would underline twice. The study found no consistent link between how well people were seen to perform and their AI adoption, training, access or workflows. Performance tracked with human traits instead: conscientiousness, resilience, confidence in solving problems, and willingness to learn. As Protiviti's Matt Duncan said, "The goal shouldn't just be doing more work faster."

What Mitch has taught on this

Mitch Mitchem has been making this argument for years. On Speakipedia Podcast 23 in October 2024, he said the point of AI is simple: "I'm trying to get things off my plate so that I'm more efficient." Then he named what has to grow when the busy work leaves: "if AI takes a lot of this thinking work from us, we better get better at the human element. We better get better at the things that make us human, like communication, empathy, emotional control."

He also described the mindset the LSE data now rewards. The question, he said, is "am I trying to prove it wrong, or am I trying to let it teach me?" That is willingness to learn, the exact trait the study tied to performance. His forecast on the same show: "I think there's a golden age coming for human skill." Amplify, Not Just Automate means the saved hours go to the human work, on purpose.

Best practices for leaders, L&D and sales managers

  1. Name the dividend. In your next team meeting, agree on what saved AI time is for: one hour a week for customer calls, coaching, or deep work on a stuck project. Write it down. Lordan's warning is about the absence of direction.

  2. Train before you measure. The 93% versus 50% adoption gap is a training gap. L&D leaders: run short, role-specific sessions on real tasks from each team's week, not a generic tool demo.

  3. Standardise one workflow per team. Pick a task the team does every week (meeting notes, a proposal first draft, a weekly report) and agree on one shared AI method for it. The survey links standard workflows to nearly double the time saved.

  4. Stop scoring usage. If people fear being judged for not using AI enough, they will perform AI use. Judge outcomes and judgment instead, and say so out loud.

  5. Close the age gap on purpose. Pair a trained early-career employee with an experienced one for 20 minutes a week. One brings the tool, the other brings the context.

What's driving people crazy this week

Being graded on AI use instead of results. Last week, Microsoft employees signing as "Concerned Microsoft Workers" published an open letter opposing any link between AI adoption and performance rewards or discipline. The LSE number explains the mood: 48% fear being judged for using AI too little. The human fix: measure what the work achieved, and let people say "AI didn't help here" without penalty.

What this means for the next 1 to 3 years

  • The "time budget" becomes a leadership tool. We expect smart teams to plan saved hours the way they plan money: some to growth, some to quality, some to recovery. Teams that do not plan will find the hours spent for them.

  • Training becomes the multiplier, not the perk. With trained people saving 22% of their hours and untrained 13%, skipping training costs time every single week.

  • Human traits become the hiring signal. If AI access is everywhere, it stops separating people. Resilience, conscientiousness and willingness to learn do. Interviews will test those directly.

  • Burnout gets an AI chapter. Workload creep is quiet. Leaders who check in on load, not just output, will keep their best people.

We wrote about the same pattern from another angle in The AI Cleanup Tax: savings that land on someone else's desk.

How we evolve

The human skill to build is deciding what matters: choosing, on purpose, where your attention goes. As AI handles more of the doing, the people who thrive will be the ones who can say no to busy work and yes to the work only a person can do. That is the heart of Mitch's ADAPT thinking and the Human Element.

One small daily practice: at the end of each day, write one line: "AI saved me time on ___, and I spent it on ___." After a week, you will see whether the time is building something or just filling up.

Try this today

  • Ask at your next team meeting (5 minutes): "Where did the time AI saved you last week go?" Listen for "more of the same."

  • Make the dividend decision (10 minutes): pick one use for one saved hour per person per week and put it on the calendar.

  • Audit one scorecard (10 minutes): remove any metric that counts AI usage by itself and replace it with an outcome.

Free tool: the saved-time agreement

Copy this into your team notes: "When AI saves us time, we will spend the first hour each week on ___ (customers, coaching, or one stuck project). We will not fill it with more of the same work by default. Each Friday, we will each answer one question: what did the saved time build this week?"

FAQ

What is AI workload creep?
It is when the time AI saves gets absorbed by higher output and broader responsibilities instead of better work or less pressure. In the LSE and Protiviti survey of 3,001 professionals, AI users saved an average of 19% of their working week and reinvested 72% of that time in other work.

Does AI training actually matter?
Yes. In the same survey, 93% of trained employees use AI against 50% of untrained ones, and trained employees save 22% of their working hours compared with 13% for untrained employees.

Does using more AI make someone a better performer?
The study found no consistent link between perceived performance and AI adoption, training or access. Higher performance was associated with conscientiousness, resilience, confidence in solving problems and willingness to learn.

Where HIVE can help

Planning 2027 learning? Talk with us about training your people will actually use on Monday. Tools bought but the time savings never show up? See our AI enablement work.

Get The HIVE Daily Buzz in your inbox.

Next
Next

The AI Capability Paradox: Why Senior Leaders Win With AI and Junior Staff Fall Behind