Why Replacing People With AI Backfires

I had a call last week with a director who wanted to "replace half the admin team with AI agents." He'd read a LinkedIn post about a company that sacked its customer service team and replaced them with a chatbot. Saved a fortune, apparently.

I asked him one question: "When the chatbot gets it wrong, who catches it?"

Silence.

This is the conversation I'm having every single week now. And BCG's latest research, published this month, backs up exactly why it's a proper nightmare waiting to happen.

What BCG actually found

BCG published research in July 2026 on the AI-powered transformation office, and one finding jumps out immediately:

Organisations positioning AI as a substitute for human roles risk diluting accountability, reducing review quality, and eroding trust.

That's not me being dramatic. That's BCG, one of the biggest strategy consultancies on the planet, saying plainly that swapping people for AI without keeping humans in the loop is a recipe for disaster.

And the numbers behind it are stark. S&P Global found that 42% of organisations abandoned most of their AI initiatives in 2025, up from 17% the year before. The average company now scraps 46% of its AI projects before they reach production. Nearly half. Gone.

Meanwhile, McKinsey's data shows only about 6% of companies are getting significant financial return from AI. MIT's research lands in the same ballpark — about 5% of task-specific AI tools ever make it to production with measurable impact.

Why replacing people is the wrong instinct

Here's what I see all the time with UK SMEs. Someone goes to a conference, hears a keynote about AI agents replacing whole departments, comes back and wants to "automate the team away."

The problem isn't the ambition. It's the mental model.

When you treat AI as a replacement for a person, three things happen:

• Nobody owns the output anymore. The person who used to catch errors is gone. The AI doesn't know it's wrong. Mistakes sail through.
• Review quality drops. The remaining staff don't trust the AI's work, so they either over-check everything (defeating the point) or stop checking altogether (dangerous).
• Trust erodes fast. One high-profile cock-up — a wrong invoice sent, a dodgy compliance report — and the whole project gets binned. I've seen it happen in weeks.

BCG's other key finding reinforces this. Their 2026 AI Radar research showed that decisive CEOs were twice as likely to deploy AI across an entire workflow — not just tinker with pilots. But the ones who succeeded weren't replacing people. They were redesigning the work itself, then layering AI on top with clear human checkpoints.

What actually works (and what to do Monday morning)

The companies getting this right aren't the ones with the biggest AI budgets. They're the ones who follow a dead-simple pattern:

• Augment, don't replace. Use AI to do the grunt work — data entry, first-draft reports, sorting invoices — but keep a human owner who reviews and signs off. The AI does the heavy lifting. The human does the thinking.
• Redesign the workflow first. McKinsey found workflow redesign is one of the strongest factors separating AI high performers from the rest. Before you add AI, map out the process, cut the dead steps, and decide where decisions actually get made. Then slot AI into the bits that are repetitive and low-risk.
• Start with one process, not the whole company. Pick something painful but contained — like purchase order processing or month-end reconciliation. Get it working. Prove the value. Then expand. The top mid-market companies in MIT's research went from pilot to production in about 90 days. The big enterprises that tried to boil the ocean took nine months and usually failed.

The bottom line

If your AI strategy starts with "how many heads can we cut," you're already on the path to joining that 42% who bin their projects.

The companies winning at AI aren't the ones replacing people. They're the ones giving their people better tools, clearer processes, and more time to do the work that actually matters. That's not as exciting as "we fired everyone and hired a robot." But it's what actually works.

And if you're an SME thinking about dipping your toe in, don't start with the AI. Start with your data and your processes. Get those sorted, and the AI piece becomes ten times easier — and ten times more likely to actually deliver.

References & Further Reading

1. BCG, The Future of the AI-Powered Transformation Office (July 2026)
2. BCG, Agentic AI Strategy for CIOs and CTOs in 2026
3. S&P Global Market Intelligence / McKinsey, Enterprise AI Agents Adoption Statistics 2026
4. McKinsey, Superagency: How Generative AI Can Empower People

Thinking about your first AI project?

I help UK SMEs get their data and processes sorted before they throw AI at the problem. No hype, no jargon — just practical steps that actually work.

Let's have a chat

Matty Hatton is the founder of Digital Adaption, an ERP and data consultancy based on the Wirral. He has spent 15 years delivering ERP transformations for manufacturers, including leading the data migration on a £4.5m consolidation of four legacy systems onto a single Infor LN cloud instance for a 220-user group. He holds an MSc in Digital Transformation and IT Strategy from Manchester Metropolitan University and is Microsoft PL-200 certified.

LinkedIn | Get in touch

Start with a 30-minute data risk call