AI layoff reversals: the 6% is where your business lives

The most useful AI story this month isn't a launch. It's a walk-back. A growing number of companies that cut staff and handed the work to AI are quietly reversing course, and the reasons are a field guide for anyone deciding what to automate. As CNBC reported, the AI didn't fail at the job — it failed at the part of the job that actually matters.

What actually happened

Commonwealth Bank of Australia laid off more than 40 customer-service staff and replaced them with an AI voice bot. Call volume went up, not down, because the bot couldn't resolve what customers were calling about. CBA reversed the redundancies and admitted it "did not adequately consider all relevant business considerations." IBM ran a similar experiment on its own HR function: the AI handled roughly 94% of routine requests but couldn't touch the other 6% — the ethical judgment calls, the exceptions, the messy human ones. IBM has since moved to triple its U.S. entry-level hiring, arguing an AI-first workplace needs more people who can handle what the machine can't.

Notice the shape of both failures. The AI aced the volume and broke on the edge cases. And the edge cases weren't a rounding error — they were the calls that kept customers, the decisions that carried risk, the moments a business is actually judged on.

Why it matters for your business

Small businesses don't get to run this experiment twice. If you fire your best support person to save a salary and the bot torches your worst 6% of interactions, you didn't cut costs — you cut the exact moments that determine whether someone stays a customer. The 94% the AI handles is table stakes. The 6% is your reputation.

So automate the routine ruthlessly, and design the handoff deliberately. The winning pattern we keep deploying: let AI take the volume — order status, password resets, FAQ, first-draft replies — and route anything ambiguous, high-stakes, or emotional to a human with full context. Keep the people who own the hard 6%; give them tools so they punch above their weight instead of drowning in the easy 94%. We've made this case on what AI actually automates and why demos aren't deployments. CBA and IBM just paid tuition to prove it.

Sources: CNBC — Employers who laid off workers for AI are reversing their decisions, Fortune — IBM is tripling Gen Z entry-level hiring after finding the limits of AI.