How Change Has Changed: Living Through a General Purpose Technology Pivot

Matt Alverson is a partner at IA Collaborative, a firm whose business is understanding emerging behaviours and building the future from them. His Nivan Live session refused to platitude the future. He offered a historical pattern for technology pivots, an explanation of why organisations structurally cannot keep pace, and a live demonstration of two agents negotiating on behalf of two companies.

Matt Alverson at Nivan 2026

Future ambivalence

Matt named the condition he thinks most practitioners are in. Ambivalence means feeling strongly in two opposite directions at once, and that is an accurate description of this period. His clients depend on him to have a defensible view of what is next, and he was honest that the view has never been more opaque, which is precisely why he refuses to fill it with reassuring generalities.

A general purpose technology pivot

His framing is that we are at a general purpose technology pivot point, a category that includes the printing press, the industrial revolution and computing. That matters because such moments are studiable. History does not repeat, but the pattern rhymes closely enough to be useful, and Matt went back through those eras looking for consistencies the way a researcher looks for patterns.

Era one, incrementalism

Every pivot opens with a reaction loop of hype, fear and critique. The printing press was going to reinvent knowledge and simultaneously destroy human memory, and there was a belief that mechanical reproduction removed the soul from the work, which is why the earliest presses imitated handwriting. Anyone reading current commentary about AI will recognise the shape immediately.

Era two, radical exemplars

Next come pockets of the future arriving before broad adoption, because the surrounding ecosystems do not exist and because risk is unevenly distributed. Matt cited a company that reached a billion dollar valuation with two people partly because it moved through territory a large corporation could not risk entering. The gap between what frontier models can do and what companies are willing or structurally able to bring forward is what he calls diffusion.

Era three, pent up potential

This is his favourite and the one that describes the room. Creative people hold excellent old tools that do not fit the new way of working, and they know it. There is new vocabulary to learn and new instruments to point themselves at. His supporting evidence is that each of these eras has invented a category of designer, from print design to industrial design to interface design, and his prediction is that intelligence design is the role this era produces.

The time we will not be given

The uncomfortable part of the model is the compression. The printing press had centuries. The industrial revolution had generations. Computing had decades. Matt was blunt that this era will not receive comparable grace, and that we are nowhere near the top of the curve where equilibrium returns.

Why organisations cannot keep up

The structural argument is the most important thing in the session. Technology moves exponentially and this is the most exponential case yet. Organisations can only change logarithmically, and when something complex is introduced they slow further. The consequence is that many organisations will never get past pent up potential, because the curve passes them while they are still adapting. Leading enterprises work above the curve rather than chasing it.

Prioritise new value over efficiency

His first instruction was to leave efficiency behind, which he acknowledged is a reversal of what he has advocated for years. You do not want to be the most efficient company in a market that time has passed. His example is a retailer that asked what AI genuinely unlocks and concluded it could turn customer support staff into interior design consultants, driving from a human need and taking the efficiency gains second.

The intent economy

The old economy was attention. The new one exchanges context for intent. His illustration used measured behaviour rather than theory. Someone using a language model to search spends roughly six times longer working out what they actually want, is around eighty percent more likely to stay with the decision they reach, and is roughly seven times more likely to transact than someone who ran a conventional search. That is the intent economy in a single sequence.

The no go zone

Matt was firm about the most dangerous thing a leader can do, which is to connect a language model to a pile of data and treat that as a capability. The model surfaces a handful of items and returns them convincingly, which is where much of the hallucination conversation actually originates. The problem is missing context engineering. His progression runs from mapping and defining data, to logic and reasoning rules, to output definition and guardrails, to evaluation, and back around before anything ships.

Owning a place in the agentic web

He described the shift from a web people enter and leave to a membrane around it, where agents cross on your behalf. Preparing for that requires understanding emerging behaviours, forefront applied technology, new economies and possible futures, which he noted are simply the fundamentals. His behavioural evidence came from healthcare research, including a person whose husband built an assistant modelled on a specialist because she was dissatisfied with her own physician. He calls the behaviour generative agency, where users report clarity, calm and control.

The demonstration and the closing instruction

Matt used the session to launch IA Collaborative's agentic experience, showing a client's agent contacting his firm's agent in real time about modernising an oncology department, followed by a generatively assembled response that also declines work outside their fit. Building it required a citation architecture, a constitution and trust principles, defined limits and handoffs, and a generative interface able to return across multiple surfaces. His closing advice was compact. Work on the new value you are creating, lean into the agentic web, and lead the forefront of whatever industry you are in.