Rare Signal | The Exponential Age Is Not Coming. It Is Here.
For years, organisations have spoken about AI in the future tense.
When it matures. When the tools become affordable. When the business case is clearer. When we have time to prepare. That window has closed.
The exponential age is not defined simply by better chatbots. It is defined by the collision of three curves- intelligence getting more capable, the cost of using it collapsing, and the ability to connect it (and doing so by architecting it well and training it carefully and equally so) directly to real work expanding at machine speed.
Stanford’s AI Index reports that organisational AI use rose from 55% in 2023 to 78% in 2024. At the same time, the inference cost of GPT-3.5-level performance fell more than 280-fold between November 2022 and October 2024. Hardware costs declined by around 30% annually, while energy efficiency improved by around 40% annually. This is not normal technology progression. It is industrial capability becoming radically cheaper, more available and more embedded. Stanford AI Index 2025
New Zealand is already inside that curve. Datacom’s 2026 survey found 91% of surveyed larger organisations using some form of AI. Yet 81% remain in exploration or implementation, only 15% report organisation-wide scaling, and just 4% say AI has transformed core operations. Datacom State of AI Index 2026. That is the signal.
It was once about digital literacy. Many organisations missed that boat—or treated it as a systems upgrade rather than a change in skilling to inform or improve how work, information and decisions were made. The next vessel is faster, less visible and far less forgiving- AI literacy. Not merely knowing how to prompt a model, but understanding where AI can be trusted, its architecture, where it must be challenged, how it changes roles and workflows, and how to retain human judgement, accountability and organisational control in all of its design.
AI adoption is accelerating. Organisational transformation is not. Shadow-AI is crippling.
The danger is that leaders mistake access to intelligence for an intelligent organisation.
A workforce with ye old copilots, subscriptions and experimental agents may look modern. It may even be more productive, or give the appearance ofile, being so for a while, in isolated moments. But if the underlying organisation remains fragmented—poor data, duplicated platforms, unclear authority, broken hand-offs, unmanaged vendors and obsolete workflows—AI does not resolve the disorder. It accelerates it.
That was the core warning in Rare’s recent pieces on Shadow AI, Experience-as-a-Service, and Agents of Chaos. Shadow AI creates an unofficial intelligence layer, not necessarily a useful one. SaaS sprawl turns technology into sediment. Agents or users, many isolated and untrained, access or connect fragmented systems, data and decisions with the capacity to act and sometimes in isolation.
The result can be an organisation that becomes AI-rich and strategically stupid- faster at producing work, but less capable of understanding whether the work is right, safe, necessary or commercially valuable. The global evidence is now catching up with that distinction. McKinsey’s 2026 survey found that nearly nine in ten respondents use AI regularly in at least one business function, and 44% report scaling it enterprise-wide. Yet only 37% report a positive enterprise-level EBIT contribution. Only 6% qualify as AI high performers—organisations attributing at least 5% of EBIT to AI and reporting significant value. McKinsey, State of AI 2026
The contrast is brutal. 80% of respondents say AI improves their personal productivity. But personal productivity is not the same thing as enterprise value.
The high performers are not winning because they have the cleverest prompts. They are more likely to redesign workflows, pursue growth and innovation alongside efficiency, measure impact, involve senior leadership and actively manage AI-related risk. Nearly three-quarters report fundamentally redesigning workflows around AI, compared with only one-quarter of other organisations.
That is the dividing line of the exponential age. The question is no longer: Which AI tool should we buy?
It is-
Which work should disappear, simplify or be rebuilt?
What data, identity and permissions foundation is required?
Where can an agent read, recommend, decide or act—and where must it stop?
Who remains accountable for the outcome?
How do we measure realised value rather than activity?
What happens when the system is wrong, compromised or unavailable?
These are operating-model questions. They belong with boards, chief executives, operational leaders, technology leaders and risk owners—not solely inside an IT pilot or vendor demonstration.
The agentic layer intensifies the issue. McKinsey reports that 40% of large organisations are scaling agents in one or more functions, compared with 22% of smaller organisations. Nearly one-third of respondents say they have decided against buying at least one software product or feature because agentic coding tools let them build it internally. That is a structural shift in technology economics: software can increasingly bend around the organisation, rather than the organisation bending around software.
But capability without control is exposure.
As Rare’s Agents of Chaos argued, the risk is not an AI suddenly “going rogue.” It is organisations giving capable agents too much authority across finance, cloud, customer, data and operational systems—without clear limits, human approval, audit trails, separation of duties or the ability to stop and recover.
The exponential age rewards organisations that can learn and adapt faster than conditions change. It punishes those that simply add more tools to inherited complexity.
The winners will not be those with the most AI licences. They will be those with the clearest strategic state, the cleanest operating foundations, governed agentic capability and the discipline to measure what changed.
That is no longer a future-state ambition.
It is the control problem of now.
The missing discipline is strategic choice. In the exponential age, strategy cannot be a static plan, a technology roadmap or a collection of transformation projects. It must be a living view of the organisation’s actual position: what it is trying to become, where its operating model is drifting, which capabilities matter, which dependencies are becoming dangerous and what must be controlled next.
This is the role of Rare Strategy. It looks across strategic fit, the broad and operational environment (now and into the future), people, workflows, systems, data, vendors, cost, risk and governance—not as separate workstreams, but as one operating reality.
It asks the harder questions: Are we doing this correctly? Are we doing the right thing? Should this activity, platform, process or dependency exist at all?
That matters because AI multiplies choices as quickly as it multiplies capability. More options can create more noise, more pilots, more supplier pressure and more plausible-but-wrong decisions. Strategic clarity provides the counterweight: a decision-grade understanding of where value is created, where complexity is accumulating and where leadership should deliberately say no.
The organisations that endure will not merely move faster. They will make better choices about what deserves acceleration.
Comments