AI arrives as a technology project, so it comes with a technology plan. That plan is usually good. It just isn't finished.
It says what the tools will do, and very little about what happens to the people using them. So people fill the silence themselves. They use AI quietly and don't mention it. They worry and don't say so.
This is the short version of the first Humore Report: seven findings, 24 questions to ask yourselves, and a plan for the first 90 days. The full report is at the end.
1Finding 1 of 7
AI has arrived faster than anything before it.
Generative AI reached 53% adoption in three years, faster than the PC or the internet. In the UK, the share of businesses using AI doubled in a year.
of UK businesses use AI, up from 16% a year earlier.
GOV.UK AI adoption survey 2026. Adoption speed: Stanford AI Index 2026.
2Finding 2 of 7
People are ahead of the plan.
More UK employees use AI at work than there are businesses that say they've adopted it, and most have used tools their employer never approved.
ONS, July 2026. Firms with ten or more staff.
of UK employees have used AI tools their employer hasn't approved.
Microsoft UK and Censuswide, October 2025.
3Finding 3 of 7
Adoption is wide, not deep.
Nearly nine in ten organisations use AI somewhere. Far fewer have changed how work gets done, and only 37% see an effect on profit.
of UK firms using AI use it extensively.
ONS, July 2026.
Organisations worldwide. McKinsey, The state of AI, 2026.
4Finding 4 of 7
Britain is anxious, and quiet about it.
Britons are more nervous about AI than the world average. At work, most feel overwhelmed, and nearly a third quietly work against their company's AI plans.
of Britons feel nervous about AI, against 50% globally.
Ipsos AI Monitor 2026.
5Finding 5 of 7
The change is in hiring, not firing.
The Bank of England finds AI is reducing hiring rather than increasing firing. It shows first at the bottom of the ladder.
Not everywhere. Among large employers, graduate hiring rose 4% in 2025-26, according to the Institute of Student Employers.
UK graduate vacancies, July 2026 against a year earlier.
Adzuna. Bank of England Monetary Policy Report, November 2026.
6Finding 6 of 7
The plan for people is missing.
Organisations have rollout plans and policies. Few have decided, out loud, what happens to their people, their jobs and the time AI saves.
A policy tells people what they mustn't do. A plan tells them what will happen to them.
of frontline AI users get little or no guidance on using the time it saves.
Frontline workers who use AI regularly. BCG, AI at Work 2026.
7Finding 7 of 7
What happens next is a choice.
The transition can settle humanely or not. It depends on what organisations protect now.
things worth protecting, set out in the Humore Charter.
- AuthorshipAI can draft, suggest and summarise. People still decide.
- PresenceTime to think counts as work.
- FrictionNew starters still learn the hard parts.
- KinshipTeams still meet, and know each other.
- AccountabilityA named person behind every decision about a person.
- RestSome of the time AI saves comes back to people.
Two organisations, one building.
Leaders and their people are living through different transitions. Leaders see productivity and possibility. Their people see uncertainty, and much less guidance than their leaders believe they've given.
Excited about what AI will do to jobs
King's College London Policy Institute and Opinium, April 2026: 506 UK employers, 1,215 UK workers.
Fear what AI means for their own job
Access Group and YouGov, February 2026: 503 HR decision-makers, 1,000 employees.
Skills training: called a priority, and within reach
Workday, January 2026: leaders who call training a top priority, and the employees fixing the most AI output who say they can get it.
Each row sets two related questions side by side. They weren't always asked in the same words, so read the distance, not the decimal point.
24 questions to ask before your next AI announcement.
Work through them as a leadership team. Tick the ones you can answer yes to with confidence, and be honest about the rest.
The first 90 days.
You don't need a perfect plan to start. You need to listen first, decide out loud, and keep checking how people feel.
Listen
- Run a confidential survey, and a handful of interviews across levels
- Include team leaders and the most junior people
- Map where AI is already used, officially and not
- Find the three biggest worries, and the three biggest hopes
Decide and say it
- Decide what stays human, what changes, and what you don't know yet
- Brief team leaders first, then everyone, in person where you can
- Name an owner for every AI decision about people
- Decide where the time AI saves will go
Build and measure
- Give people time and training to practise, in working hours
- Redesign one entry-level role so it still teaches the job
- Set up a safe way to appeal, and to report AI mistakes
- Measure how people feel, not only usage, and share what you learn
Ninety days from now, your people could be part of the plan instead of on the receiving end of it.
What to say on Monday.
People read every word of an AI announcement twice. Plain, honest language does more for trust than any amount of enthusiasm.
Not: “There's nothing to worry about.”
Say: “We don't have every answer yet. Here's what we know, what we don't, and when we'll tell you more.”
Not: “No roles are affected at this stage.”
Say: “Some roles will change. We'll talk to the people in them first.”
Not: “We're excited to announce our AI journey.”
Say: “We want to hear how you feel about this before we decide what happens next.”
Not: “AI will free you up for higher-value work.”
Say: “Here are the decisions that will always stay with people.”
Not: “Please use the tools responsibly.”
Say: “If the AI gets something wrong, tell us. You won't be in trouble.”
Not: “Training is available on the intranet.”
Say: “We'll give you time to learn this properly, in working hours.”
Not: “This will make us more efficient.”
Say: “When this saves time, here's what we'll do with it.”
The lines on the left are examples of the kind, not quotes from real announcements. Only say what's true where you work.
Everything here, with the evidence behind it.
63 pages and around 60 sources. One page for each of the six Charter articles, the 24 questions with their score, and every source in full.
The People Perspective is built from published research from 2025 and 2026, not from a Humore survey. Every figure names its source. Some studies come from companies that sell AI or HR software, and the report says so where it matters. If your organisation is rolling out AI and you'd like to know how your own people see it, that's what The Human Transition is for.
- The plan we haveAdoption, investment and how deep it really goes.
- The people it lands onHow Britain feels, and what people do quietly.
- The work itselfHiring, the first rung of the ladder, and where judgement comes from.
- The missing planThe gap between leaders and their people, and where the time goes.
- What stays humanThe six Charter articles, each with its evidence and a workplace test.
- What it's forWhat we want to keep.
- The other planEight principles, the 24 questions, the first 90 days.
- How Humore helpsThe Human Transition, and what you get.