Reading Time: 6 minutes; Soundtrack: Apocalypse - by Bear McCreary
For the past several years, the defining skill of anyone in a digital leadership role has been the ability to mediate the unavoidable tension between three things that never quite want to cooperate: brand communication, commercial results, and the rigid constraints of digital technology — which stubbornly refuses to evolve with the same fluidity as the other two.
All of that is already changing, driven above all by Generative AI — and plenty of people are starting to smell a mass extinction event for anyone whose livelihood is tied to digital technology. Is it true? I’ll try to work through it in this #postNerd.
Where Are We Headed?
It seems fairly obvious that GPTs are making us more productive but somehow dumber, so it’s no surprise that OpenAI — by acquiring Jony Ive’s startup — wants to build a voice-enabled device.
Picture a wearable that requires neither an opposable thumb nor a screen to show you a result — something that speaks to you in your own language and can buy you a shirt on your behalf.
Got it? Good. Now close your eyes and make a wish…
Welcome to the post-digital era.
You might wonder what whatever comes after digital actually looks like. On the surface, it won’t be entirely unlike the post-capitalist era we’re already living through — the one Jeremy Rifkin described rather well in his 2014 book “The Zero Marginal Cost Society”.
Generative AI itself — which Rifkin didn’t predict — has triggered a deep, non-incremental phase of change, precisely because once you’ve absorbed the upfront cost of training the models, deploying them delivers value at near-zero marginal cost.
Can’t do SEO? Ask GPT — it handles (almost) everything. This will be even more true once interaction shifts to voice assistants capable of:
- Creating marketing campaigns and narrating them aloud
- Helping build software interfaces by talking directly with developers
- Generating product imagery after you’ve described your ideas
- Guiding customers to products via home-device voice assistants that are marginally less dim than Alexa and Siri
- Choosing products on behalf of customers, based on their stated preferences
- Actually completing the purchase for them
- Handling customer support requests, verbally
In this scenario, the brand website will no longer sit at the centre of the experience. Digital tools will become less visible — still important for communicating products to AI engines, but at a semantic level.
Branding will change too. Logo and brand name will no longer be enough. You’ll need names that are easy to pronounce and recognise. Asking an LLM verbally to find “a Zara dress” is obviously simpler than asking for “a C&A outfit” — how is a voice assistant supposed to know the right pronunciation when two native speakers say the same acronym differently? (For the record: it’s “Tse-und-Ah”, German-style.)
Whoever can structure their catalogue around the experiences users actually search for will have the edge. In a voice context, queries like “find me lightweight trainers for trail running” only work if you’ve encoded into your catalogue the information tied to actual or aspirational use — the context you hope the user has in mind when they ask the LLM.
It’s easy to see how, as many of the traditional structures that have defined digital for the past 15 years begin to erode (e-commerce first among them), demand for those roles will fall — especially since many of the more technical digital jobs can already be replaced by GPTs, at near-zero marginal cost.
The naturalness of this shift becomes even clearer when you consider that the running cost of GPTs, where decentralised and powered by renewable energy, often delivers energy savings on top of economic ones — provided the automated task has the right degree of repeatability. ESG for the win, but not for digital managers.
With the right technical architecture in place (Headless Commerce, API-First), LLMs won’t just act as payment service providers — they could become “natural language transaction providers” (NLTPs), cutting out large swathes of the digital economic chain.
What survives?
- Product manufacturing (if you make the thing — though for pure distribution, AI can already help you run dropshipping; the most extreme cases are today’s mega-influencers)
- Warehousing and pick & pack (for mid-to-large brands, fully unmanned pick & pack isn’t realistic yet)
- Brand communication (still necessary, but increasingly mediated through LLMs — and bottom-funnel comms will likely be absorbed away from Google before long)
- Last-mile delivery (until autonomous vehicles like Waymo disintermediate that too)
- Non-trivial customer care (the trivial stuff already belongs to AI)
- Pre- and post-sale consulting (where it’s genuinely high-value — otherwise, easily disintermediated)
And here’s the kicker: the moment NLTPs manage to disintermediate payments, it becomes rational for them to move into logistics as well — just as Amazon did — keeping customer data for themselves and reducing brands to pure production and distribution structures, whose marketing serves nothing but raw awareness.
Mapping a Shifting Territory
“Know yourself and you will know the world,” said Socrates — and the ancient Greeks broadly held that moving into the future was a bit like walking backwards, eyes fixed on the past. That still seems to be the case. The difference now is that instead of walking, we’re sprinting.
Since we can only look backward and inward, what can we learn from that vantage point to anticipate what’s coming? A few generalisations:
-
Transformations happen gradually — some faster than others, and today very fast — but they still take time, and time is our best asset for aligning with future scenarios. The real question is: how much time do we have?
-
Transformations are never total. Digital technology will become less visible, but it won’t vanish. Just as agriculture didn’t disappear with industrialisation, and paper didn’t disappear with the internet, our roles won’t disappear — they’ll progressively lose their centrality.
-
Transformations hit less specialised roles first — the ones most easily automated or disintermediated. Programmers are still significantly better than LLMs — not my claim, but Antirez’s, creator of the world’s most widely used NoSQL database and a certified Sicilian. In fact, the best programmers are amplifying their own capabilities through LLMs.
-
Transformations don’t unfold the same way across all industries. At any given moment, some sectors are more exposed than others. If you’re an e-commerce manager selling unbranded coffee pods, you have considerably more reason to worry than the creative director of a luxury brand — where the product’s strong identity and the inherently 1:1 nature of the commercial relationship make disintermediation much harder. AI disintermediates low-differentiation products and services most easily.
In the post-digital era, some roles will transform and some will disappear through falling demand. This will arrive at different times and in different ways for each of us — depending on the industry, the product or service, and the individual.
What are the critical capabilities we need to develop to stay relevant? A recent Harvard Business Review article on leadership in the age of artificial intelligence offers some useful pointers.
First, we’ll need to become systems thinkers who can execute with AI support — people capable of seeing the connections between market, technology, and people, and weaving them into semantically rich, experientially satisfying user journeys, even when those journeys are filtered through LLM-based gatekeepers.
We’ll need to develop our capacity for empathy. After all, who holds together a team of strategic-operational geniuses while the AI handles the low-level tasks? Who provides the connective tissue between the big picture and the day-to-day? Who is antifragile enough to navigate the chaos of a new era while keeping teams and stakeholders calm?
We’ll need to know how to do things, not just how to be things. In a post-digital world, yes — many operational tasks will fall to middle managers. But senior managers will also need to get their hands dirty in a zero-marginal-cost world, or they’ll be incapable of transmitting the AI-first mindset their teams need to compete.
How much time do we have before the change catches up with us? Some of us are already in trouble. For others, it’s three to five years away. For a few, perhaps a little longer. Within a decade, the transition will have set its course.
It seems clear to me that tomorrow’s most relevant managers won’t be those who identify with a role — they’ll be those who already understand their own evolutionary trajectory and can move fluidly between roles: sometimes tactical, sometimes strategic, depending on the company, the product, the state of the market.
In a world where even sophisticated tasks are progressively handed off to LLMs, the future belongs to those managers who are already cultivating meaning, not just function. Everyone else will need to adapt to survive — and the only viable evolutionary response is continuous evolution.
Interesting times ahead…