Three polls, one pattern
AI at work
Cyber trust
Politics
View data
| Poll | Measure | % |
|---|---|---|
| AI at work | Business leaders who say AI time savings should be shared equally | 56% |
| AI at work | Employees using AI who are expected to do more work | 63% |
| Cyber trust | Business leaders confident they would keep customer trust after a breach | 75% |
| Cyber trust | Consumers who would stop buying after a breach | 50% |
| Politics | Consumers confident politicians understand them | 30% |
| Politics | Business leaders confident politicians understand them | 16% |
Gap one: AI at work
Ask business leaders who should benefit when AI saves an employee time and 56% say both sides, equally. Ask employees and 61% say the employee.
Then ask what is happening in practice. 63% of employees who use AI say they are expected to complete more work. 3% say less. Meanwhile 56% of business leaders say AI has increased productivity.
Leaders believe in a fair split. Employees believe the time is theirs. What employees report getting is neither. Read the AI findings.
Gap two: cyber trust
75% of business leaders are confident they could maintain customer trust after a significant cyber incident.
50% of consumers say they would stop buying from a company that lost customer data, at least until they felt the problem was fixed.
And when we asked what would bring them back, 57% said proof that security had improved. 1% said an apology from senior leadership, which is usually the first thing a company reaches for. Read the cyber trust findings.
Gap three: politics
Conference season is a month of parties telling the country they understand it.
30% of consumers are confident politicians understand the issues that matter to them. For business leaders it is 16%. A third of consumers and 42% of business leaders hadn't heard enough to have a view at all. Read the post-conference findings.
What the three have in common
In each case, one side is confident and the other side hasn't been asked.
- Leaders assume AI gains are being shared.
- Businesses assume trust would survive a breach.
- Parties assume the message landed.
None of these are unreasonable things to believe. They are untested things to believe. And the cost of being wrong only shows up later, as a retention problem, a revenue dip or an audience that stopped listening.
The question to take away
You will have your own version of this gap. Something your organisation is confident about, concerning an audience nobody has checked with recently.
The useful exercise is to write it down as a number. "We think X% of our customers would say Y." Then go and find out.
If the number matches, you have a decision you can defend. If it doesn't, you found out before it was expensive.
Reality. Properly understood.
Methodology
Figures come from three surveys run by The Reality Department: AI productivity, cyber trust and the post-conference reality check. Each surveyed 500 nationally representative UK consumers and 250 business leaders. Fieldwork: October 2026.
Want to test what your organisation assumes about its audience? Work with us.
Free to quote with a link to this page. Need the full data tables? Ask us.
Running research of your own? See whether it's getting the attention, engagement and action it deserves. Take the free Research Impact Scorecard →