Copy article

New Forrester Research on AI Uptake and Employee Qualms

ended 30. March 2026

Forrester's report advice on low workforce AI adoption is more training and better communication. That's a reasonable prescription for a skills gap. It doesn't touch the actual diagnosis. 

When 51 percent of UK business leaders publicly describe AI as a staff-reduction tool, and 43 percent plan to shrink entry-level roles, employees who resist AI adoption aren't anxious or underskilled. They're reading the room correctly.

Training programmes don't fix broken trust. They can't, when the person running the session reports to someone whose KPIs include headcount reduction. The workers Forrester wants to upskill are being asked to accelerate a process they have every rational reason to slow down.

We'd like your views:

  • If employees know their employer views AI primarily as a headcount tool, what would a credible "AI as opportunity" message actually look like, and who in the organisation could deliver it with any authority?
  • Forrester says social learning outperforms formal training for AI adoption. What does that mean in practice for a mid-sized firm with no internal AI community to speak of?
  • Should regulators require employers to disclose their AI workforce strategy to staff before rolling out AI tools, in the same way TUPE rules require disclosure during business transfers?

3 responses from the Newspage community

Copy all

Star Quote
Copy

Colette Mason
AI Ethics Consultant at Clever Clogs AI
Asking frightened employees to embrace AI while their CEO is on record saying it will cut headcount isn't an engagement failure. It's a common decency issue.

Price Waterhouse Coopers found that more than half of CEOs have seen neither revenue gains nor cost reductions from AI investment. MIT's NANDA research puts a harder number on it: 95% of enterprise AI pilots are delivering no measurable impact on the bottom line. So the productivity argument isn't landing, the trust argument isn't landing, and the training isn't landing.
At some point the variable worth examining isn't employee readiness. It's whether organisations are asking people to become compliant in their own replacement by systems that, so far, aren't obviously outperforming them.
Copy

Rohit Parmar-Mistry
Founder at Pattrn Data
Training is the easy bit. Trust is the hard bit.

If staff think AI equals redundancy, “upskilling” lands as: help us automate you out. That is not a skills gap, it is an incentives gap. The only credible message is one that trades something real: transparent workforce plans, guardrails on how AI outputs are used in performance management, and a commitment to redeploy before you remove roles. Otherwise people will keep slowing the rollout, quietly and rationally.

“Social learning” works when it is safe to admit mistakes. In practice that means time carved out for peer demos, shared prompt libraries, and lightweight governance so experimenting is not a career risk. You do not need an “AI community”, you need a protected space and a leader who is not measured on headcount reduction. Regulators should treat workforce impact like any other material change: disclose the intent before the tools arrive.
Copy

Martin Port
Founder at Build Concierge
If AI is positioned internally as a way to slash headcount, no amount of training will drive adoption.
Leaders need to be explicit that productivity gains will be reinvested into growth, better roles, and higher-value work, not just cost reduction. And critically, that message cannot come from HR or L&D. It has to come from the CEO and CFO, because they own the outcome and the trust.
Social learning, in practice, means moving away from formal programmes and embedding AI into real workflows. Small groups using AI on live tasks, sharing outcomes, and proving time saved or revenue gained. Adoption follows evidence, not instruction.
On regulation, transparency is inevitable. If companies are deploying AI that materially changes workforce structure, employees should be informed upfront. Not necessarily through heavy regulation, but through clear disclosure of intent and expected impact. Without that, you don’t just get resistance, you get disengagement, which is far more damaging.