From where we’re sitting now, the idea of post-AI seems impossibly far off. This new technology has gripped our collective consciousness with unprecedented speed. We are, perhaps, four years into this journey, and yet nobody is dismissing it in the way they did the internet (and Wi-Fi) as the “CB radio of the 90s”. We have learnt our lesson about emerging technologies: it’s better not to dismiss them too quickly….

AI is here to stay – but how will the world look once it is deeply integrated in the way we live?

This is a question we asked of digital, 14 years ago: What will the world look like when digital technology is no longer remarkable, but ubiquitous? That is the world we live in right now. And so our new question becomes: What does the world look like when AI is ubiquitous and unremarkable?

However hard that might be to imagine, it’s a question worth asking ourselves as we seek a compass to the future.

Don’t use AI for existing processes – replace them

The challenge for our imagination is that we’re still in the earliest stages of this technology’s adoption. Back in the early 2000s, a consultant working at the intersection of technology and real estate explained to me that he always saw digital adoption come in two waves. First, the companies tried to digitise existing processes. And then, eventually, they realised that the true gains were to be had by doing things differently; by rethinking their processes and services in light of the new technology. His job was to try to minimise or eliminate the journey between the two steps.

And that’s where we are right now: we are trying to “AI-ify” existing processes, to make them cheaper, and more efficient. It’s not panning out terribly well. Research from MIT suggests that only 5% of businesses are seeing returns from AI investment. Now, there are good reasons to be sceptical of some of the more breathless reporting of that figure. Exponential View did a good job of digging into the weaknesses of what is, after all, a very early study. But, quibbles aside, there is evidence in this and other research that AI is not yet piling undeniable productivity gains into the economy.

But, to understand why, you need to dig into the nuance of the report, beyond that headline figure:

Only 5% of organizations are translating AI pilots into real operational or financial impact. The core issue isn’t talent, infrastructure, or regulation—it’s the lack of learning, integration, and contextual adaptation. Tools like ChatGPT are widely used by individuals, but enterprise-scale implementations fall flat due to brittle workflows and misaligned deployment strategies. Yet a few buyers and vendors are succeeding—by focusing on adaptive, feedback-driven, and tightly integrated systems.

Yes, it’s those words from the 2000s floating back to us. You don’t make the most of AI by applying it to existing processes; you make the most of it by adapting what you do to the reality of AI. As a recent piece in The Economist on this issue put it:

All this signals a deeper flaw in the argument that AI is powering a productivity boom. Such improvements are usually made not just when workers use a new tool more often, but when firms reorganise production around it. Early factories became only a little more efficient when steam engines were replaced with electric motors; the real revolution came decades later after floor plans were redesigned to make the most of electric power. More recently, productivity growth was a disappointment for years after personal computers became widespread. It accelerated only once firms implemented business models that exploited the technology to its full potential.

The post-AI world works differently

And that’s our first clear glimpse of the post-AI world: things will be done differently. We aren’t – yet – able to say how, but the way business transacts in the future will be as fundamentally different from now as digital business was from the world it replaced. Few of us can now imagine how work was done before we had computers on our desks – or in our pockets. But I remember my Dad getting his first computer at work (a pre-Windows IBM PC), and only ever using it once in a while for spreadsheets. We have been through the adaptation phase in the decades since, and have built a new way of working to accommodate it.

We’re about to go through it again, but the recency of the last shift may be hampering us. The digital transformation was based on computers that were accurate and predictable. The AI transformation requires us to adapt to machines that guess, and we are not well-prepared for that.

Changing our mental model of AI

Yes, calling AI guessing machines is an oversimplification of a technology that is deeply rooted in probabilistic rather than deterministic output. But it is a useful one in reminding us just how different this technology is. That gap in understanding means that some existing uses of AI we’re attempting might not pan out. Are AI overviews accurate enough to trust? Many have found not, with sometimes disastrous results:

In one case that experts described as “dangerous” and “alarming”, Google provided bogus information about crucial liver function tests that could leave people with serious liver disease wrongly thinking they were healthy.

Generative AI can produce new material based on patterns found in existing material. But that is not the same thing as extracting answers. Because it works probabilistically, it gives likely answers rather than accurate ones. And in some situations, that can be dangerous. If that proves to be an inherent characteristic of this technology, rather than a bug that can be ironed out, then we’re moving towards a world in which we deploy this technology where rapid determination of probable outcomes is good enough, and only there. Other tasks will be better left to more deterministic – or even human – approaches.

What will be AI-ified will be AI-ified

And that’s our second sign of the post-AI future: what can be AI-ified will be AI-ified, just as everything that could be digitised was (and is being) digitised. And that’s the process we’re working on right now. The pressure to use AI is everywhere:

Having already coaxed true believers into trying out AI tools, executives at some of the world’s biggest companies are now turning to more aggressive tactics to boost uptake.

Last month, Julia Liuson, president of Microsoft’s developer division, warned staff that “using AI is no longer optional”. Liuson said in an internal email that AI use should be factored into “reflections on an individual’s performance and impact”.

Many of those initiatives will fail. AI is not – yet – such a general technology that we can apply it with equal success to every task. But a process of experimentation will prove where it has value, and where it doesn’t. In our conceptual post-AI world, this will seem obvious. But from this temporal side of the process, it remains a massive question mark hanging over every business.

Accelerating the journey to the post-AI world

And here’s where the idea of the AI hype bubble bursting might actually help accelerate transformation. Many of the current experiments will prove to be ill-conceived, and driven more by fear than by opportunity. As HBR, in an analysis of the MIT finding about the lack of impact of AI projects, suggests:

It sounds obvious, but by framing AI as radical and disruptive we often lose sight of the connection to the most fundamental objective of business: to solve problems for customers. The way out of this trap is to 1) understand this AI moment in the larger arc of transformation, 2) focus on AI’s potential to help better serve customers, 3) experiment with a focused set of opportunities to prove them out (with an eye toward scaling), and then 4) scale them up.

However, we do already know some things about generative AI. It’s great at extracting patterns in large data sets, and it’s pretty good at creating new things based on those existing patterns.

So, we can make some reasonable predictions about our post-AI future. For one, AI will make today’s glut of content online seem like a mere molehill, when compared with the looming mountain range that is already building. When content is abundant, attention becomes scarce – and will become even more scarce in comparison to the value of content. And so, generating content will not, in any way, be a competitive advantage. Finding a way to prove its value to people will be. There’s likely to be multiple solutions to that, from generative content highly aligned with customer needs, to a premium on high-quality human content that rises above the slop. And successful businesses will scaffold that by a relationship between creator and consumer, which AI can’t have.

Humanity in the post-AI world

That, in turn, might see a return to more experiential, human products. It’s likely to be ones which are, by their nature, limited in scale and scope that thrive. Artisanal experience, just as artisanal products find a place in a world of mass production. You can only fit so many people into a conference venue and still deliver a meaningful experience, for example.

In fact, one of the more likely predictions about the post-AI world is the end of the process of trying to make human beings behave like machines. The process that began with the Gilbreths over a century ago has reached its logical conclusion. The Industrial Revolution mechanised many formerly human-performed tasks. Time and motion stgdies-infomerd corporate cultures sought to apply machine-like principles to other human activity, and office work in particular. But now, much of that work can be done by machines, too.

What’s left for humans? As David Mattin put it at NEXT25, we need to focus on what is uniquely human in terms of connection and empathy.

Dystopia or renewed focus on humanity?

So, our future post-AI world exists in a Schrödinger’s Cat-like situation. We could end up with a dystopian future, in which the benefits of AI accrue to a shrinking number of tech executives. This is a nightmare world where the rest of us struggle to find work. Or, we create a world where AI tools abstract away much of the most menial information work – including AIs with a physical awareness of the world. Then we find new purpose in human connection and relationships.

The collapse of that waveform – the opening of the box – is not predetermined. How we handle the next decade will shape that decision. The technology’s implementation is not pre-destined. It’s in our hands.

The digital world created a situation where we mechanised humans, but the AI world is reversing that: we’re humanising machines. And machines are better at doing some of the work of humans, than we are at doing the work of machines. And that suggests a deeply transformative model of work.

What will that look like? And what will the liminal space between the post-digital world and the post-AI world look like? Well, that’s what we’ll be exploring next.

Photo by Sascha, Adobe Stock number 574937307.