Your team already knows which AI strategy you're running. New research shows what happens next.
Harvard Business Review published research this month tracking what happens inside organisations when leaders make their AI strategy about cutting headcount versus building human capability. The findings apply directly to every UK SME making this choice right now — including yours.
The fork in the road
The HBR researchers surveyed 1,294 full-time workers across the US, Canada and the UK and asked one question: does your organisation use AI to augment employee capabilities, or to automate work and reduce costs?
62% of employees believe their organisation is on the augmentation path. 34% believe they are on the automation path. That sounds encouraging until you look at the breakdown by seniority.
81% of senior leaders think their organisation is all-in on augmentation. Only 53% of individual contributors agree. In professional services and retail the numbers are sharper still — 40 to 50% of employees in those sectors suspect AI is being used to undercut their job security.
Less than half of all organisations have formally communicated anything about their AI plans. The researchers are blunt about what fills that silence: employees draw their own conclusions, and those conclusions shape everything that follows.
Why the automation path looks right and ends badly
The automation case is logical. Fewer people doing the same output means lower costs. In the short term that shows up in the numbers, and it is why Jack Dorsey cut nearly half of Block's 4,000-person workforce in February, citing AI as the justification.
The research maps out what follows. Staff who sense their jobs are at risk disengage. Productivity drops — the same researchers found in earlier work that happy workers are 13% more productive, and that relationship runs in reverse too. Leaner, overloaded teams start filling the gaps with AI they have not been trained to use well. The researchers have a term for the output: "workslop." Low-effort, low-quality AI-generated work that looks like progress but is not. Employees who feel forced to adopt AI produce it at a rate 65% higher than those who feel genuinely empowered.
Then the high performers leave first. They always do. Institutional knowledge walks out with them. The talent pipeline thins. The employer brand deteriorates. The business becomes progressively harder to run and harder to grow, compounding quietly across six stages the researchers map in detail.
Large organisations can absorb this process for longer before the damage becomes visible. Most UK SMEs cannot. You do not have the bench depth to survive losing three high performers in eighteen months because your team read the signals correctly and made rational decisions about their futures.
What the augmentation path actually requires
The augmentation path has a deeper initial dip. It requires investment in training, process redesign and a genuine, credible commitment to people. Brynjolfsson's research — cited in the HBR piece — suggests that the organisational rewiring needed to get real value from AI requires roughly ten times the investment of rolling out the technology itself. That number is uncomfortable. It is also honest.
What the research shows is that it compounds. When people believe AI is making their work better rather than threatening their position, they engage with it properly. They become pilots rather than passengers — using judgment about where AI helps and where it does not, producing better work, staying longer and building the institutional knowledge that makes a business genuinely hard to compete with.
Aon committed to its 60,000-person workforce that AI would expand opportunity rather than erode it, and that commitment is credible because they have a track record — they protected jobs through Covid when cutting would have been simpler. Microsoft rebuilt its culture from "know-it-all" to "learn-it-all" under Satya Nadella rather than cutting its way to efficiency. The researchers argue this is the path that produces compounding gains. The augmentation advantage builds on itself precisely because it is rooted in human behaviour rather than cost arbitrage.
"The question is not what AI can replace. It is what your best people could do if AI took the work that does not need their judgment."
What this means if you run a UK SME
The HBR paper is written for enterprise CEOs. The mechanism it describes operates at any size.
If your staff were asked today whether AI is being introduced to make their work better or to reduce the wage bill, what would they say? If you are not certain of the answer, that is itself an answer. The research is clear that perceptions of intent form quickly, travel fast and are difficult to reverse once established.
Augmentation does not require a culture change programme or a strategy document. It starts with a clear, honest answer to one question: where are your best people spending time on work that does not need their judgment, and what would they do differently if AI took that work from them?
The businesses that answer that question well, and build around the answer, are the ones the research suggests will be in a structurally different position in three years. The ones that do not will have bought some tools, saved some costs in the short term and spent the intervening period wondering why none of it is working.
Which path is your business actually on?
The free AI Readiness Audit gives you an honest read of where your business sits and what to address first. No pitch. No obligation. You get the output regardless of whether we work together.
Start the Free AuditSource: De Neve, Hancock and Niederhoffer, Why Companies That Choose AI Augmentation Over Automation May Win in the Long Run, Harvard Business Review, April 2026. Read the full article.