Is AI Killing Productivity, Or Are We Waiting For It To Save Us?

We often tell ourselves that the next wave of automation will finally change everything. ERP once promised it, then CRM, then workflow platforms. Now AI is in the spotlight. Each tool is real, and each does make a difference in practice. But at the level of whole economies, the numbers remain stubborn. The UK still produces around 20% less per hour than the US, despite both using the same shelves of technology. Countries like Germany, Denmark and the Netherlands sit much closer to the US benchmark.

The promise of automation

Every decade brings its miracle tools. In the 90s it was ERP. In the 2000s, CRM and workflow platforms. Now, AI has centre stage. Each time the story feels the same: soon work will be effortless, output will soar, we’ll need fewer hours, and life will get easier.

The story is tempting. The risk is clear. We over-index on potential, we under-index on adoption, discipline, and grind.

The tools are real. CRM systems really do help sales teams manage pipelines. Automation in invoicing really does cut errors. AI coding assistants really do speed up developers. Yet at the level of whole economies, the numbers are stubborn.

On GDP per hour worked, a widely used measure of labour productivity, the UK sits around 20% below the US. That gap has persisted for decades, despite both economies having access to the same shelves of technology. Germany sits closer to the US benchmark. Countries like Denmark and the Netherlands often appear near the top of the tables (Commons Library; OECD data).

A simple comparison

To make the point concrete, here’s a snapshot of 2023 estimates (PPP-adjusted, OECD basis).

Country	GDP per hour worked (USD PPP, 2023 est.)
United States	77
United Kingdom	62
Germany	72
Denmark	80
Netherlands	78

The shelves of technology are shared. The results are not.

What do we mean by productivity?

Labour productivity is defined simply: output per hour worked. The US Bureau of Labor Statistics explains that it is different from total factor productivity, which also accounts for capital and other inputs.

It is a straightforward ratio. More output for the same hours means higher wages, stronger firms, and better living standards. Less output for the same hours means the opposite.

The puzzle that won’t go away

The UK’s productivity gap has been studied for years. Reviews by the Economics Observatory highlight the same recurring themes: under-investment in capital and skills, weak diffusion of best practice from leading firms to the long tail, and fragmented policy that fails to reinforce adoption consistently.

This is the so-called productivity puzzle. The puzzle is not why the UK lags the US by around a fifth. It is why, after decades of debate and repeated technological waves, the diagnosis has changed so little.

Same shelves, different results

Here’s the heart of it. The shelves of technology are global. ERP, CRM, workflow software, cloud, data, AI — they are all accessible to firms across advanced economies. If some countries consistently turn those tools into higher output per hour, the difference is not access. It is habits.

At Yopla, we explore this in our digital mapping work. What we often see is not a shortage of tools, but a shortage of visibility: firms struggle to see how work actually flows. Without that clarity, adoption becomes patchy, and productivity gains stall.

And when you step back, this same pattern, over-indexing on potential, under-indexing on adoption, discipline, and grind, shows up in digital maturity too. Our guide to digital maturity levels makes the point directly: maturity is not about what’s in the toolbox, but about how far organisations embed those tools into daily practice.

The danger of waiting

AI risks repeating the same cycle. In boardrooms and team meetings the message often goes: “Let’s wait for the next model. Once the platform matures, then the gains will come.”

But waiting is costly. Workflows remain unchanged. No one measures progress. The opportunity to learn is lost.

Contrast that with the United States. Not every firm is a paragon, but enough adopt early and operationalise consistently to shift national figures. US labour productivity growth has ticked along at roughly 2% annually since the pandemic (Reuters). That is not spectacular, but compounded, it is decisive.

Germany, Denmark, and the Netherlands also maintain high productivity levels. These examples underline the point: the difference is not tools, but adoption and spread.

What to do on Monday

This is not about national policy alone. It is about what firms do tomorrow morning.

Take one workflow, claims, onboarding, reporting, invoicing. Document the steps, set a 90-day target for saving time or reducing errors, and deploy the mature tools you already have, whether that’s CRM modules, billing automation, or document AI.

Don’t bolt new tech onto old routines. Redesign the workflow. Measure the before and after. Publish the results. And when it works in one team, spread it to three more. Standardise what you can.

That is adoption. That is diffusion. And that is how productivity actually moves.

The simple claim

Automation does not kill productivity. Waiting for it to rescue us does.

Whether it’s ERP, CRM, workflow software, or AI, the difference lies in adoption, discipline, and grind. The firms and economies that operationalise early, measure honestly, and spread what works will keep pulling ahead. The rest will keep waiting for the nirvana that never arrives.

The story is tempting. The risk is clear. We over-index on potential, we under-index on adoption, discipline, and grind.

Sources

Let's talk