Is Your ROAI Self-Sabotaging?

Most organizations measure return on AI investment the same way they measure every other investment: cost reduction, headcount eliminated, tasks automated. Fewer people doing the same work. More output per dollar of labor.

This is scarcity logic applied to an amplification technology. And it’s actively undermining the investment.

Here’s what’s actually happening in organizations that give workers real access to AI tools: people aren’t just doing their existing jobs faster. They’re doing more. An analyst who used to produce three reports a week is producing ten — and they’re better. A project manager who spent hours on status updates is now spending that time on the risk assessment and stakeholder alignment that used to get squeezed out. A customer service representative who used to handle routine inquiries is now resolving complex issues that previously required escalation.

This isn’t a problem. It’s undervalued gain. Workers are using AI to expand what they can accomplish — not just to subtract tasks, but to add capability. The people closest to the work are figuring out, on their own, how to use AI to do things that nobody planned for and nobody is measuring.

And what are some organizations doing in response? Limiting which workers can access AI tools. Restricting what they can use them for. Requiring approval before employees can experiment with new applications. Rationing an amplification technology as though it were a scarce resource.

This is foolish scarcity thinking. It’s like handing someone a telescope and telling them they’re only allowed to look at one star.

The business case built on “fewer people doing the same work” doesn’t just miss the point — it works against it. When you measure AI by how many people it replaces, you drive the organization to strip out the very human judgment that makes AI valuable. You create brittleness. You lose the people who know what the AI is missing. And you leave the real gains — the expanded capabilities, the new work, the judgment that was never applied before — completely unmeasured and unmanaged.

“You’re optimizing for less when the technology delivers more.”

The real return on AI shows up in places most business cases don’t look: outcome achievement that wasn’t possible before. Response times that dropped from weeks to hours. Capabilities deployed where they’re needed rather than locked inside functional silos. Workers contributing in ways their job descriptions never imagined.

If your ROAI calculation starts with headcount reduction, you’re measuring the wrong thing. You’re optimizing for less when the technology delivers more. And you’re signaling to every employee in the organization that AI is a threat to their livelihood rather than an amplifier of their capability.

That signal will be received. And it will determine whether your people embrace the technology or quietly resist it.

The question isn’t how many people AI can replace. It’s what your people can accomplish when you stop rationing the one technology that multiplies human judgment instead of subdividing it.