Why won't the next decade of business resemble the last century?

Over the last two years, studies have shown that generative AI reliably boosts task-level productivity, with typically between 15 and 40% gains. These gains happen when a knowledge worker uses a tool, like ChatGPT, to draft an email or write a piece of code. They are real improvements, but don’t automatically translate into company-level earnings. Most organizations are still using AI as a tool inside yesterday’s workflows.

Our experience points to a different path. When companies intentionally redesign workflows around AI, the gains stop being incremental. They unlock structural, not incremental, scale and enable entirely new ways of operating, not just the same steps done a little better.

The bottleneck to achieving this future is no longer the technology or model intelligence — it’s the 'context gap' – the distance between generic intelligence and the company-specific expert judgment required to run high-stakes enterprise workflows. As AI absorbs more execution, the marginal cost of doing work collapses. The scarce resource shifts from doing the work to subject-matter experts capable of directing, governing and the improving work.

Take, for example, a leading telecommunications provider serving more than 100 million customers, which relied for decades on frontline agents resolving problems manually. When a top agent solved a complex issue, the reasoning behind that solution stayed with the individual. Outcomes were reactive, uneven and hard to scale.

AI changed this dynamic. By training systems on the reasoning patterns of top performers, service calls closed much faster because the company turned individual expertise into a shared asset for the entire team. While humans were previously constrained to reacting to calls, the AI systems scales to proactively anticipate individual customer needs, enabling the company to turn its care operations from a resolution hotline into a revenue-generating concierge.

In effect, the company transformed the judgment of its best agents into shared infrastructure — and then built new economics on top of it.

In another example, a national healthcare provider faced a familiar constraint: clinician capacity was being eaten up by administrative work. Routine tasks — authorizations, documentation and coordination — were consuming time that should have gone to patient care. The result was a system stuck in reactive mode: the sickest patients absorbed most of the available bandwidth and scaling preventative interventions was economically unrealistic. By capturing and automating these administrative workflows with AI, the provider frees substantial clinical time and redesigned how providers and patients engaged. Authorizations became faster and more consistent, care teams could focus on higher-value decisions and preventative outreach became economically viable at scale — a structural shift, not a productivity gain.

These examples point to the same lesson: technology is the catalyst, but value comes from structural change. Once intelligence becomes reusable and improves continuously, the boundary between traditional roles start to blur and new ones emerge. Organizations with redefined roles, reconfigured workflows and real-time learning systems are already pulling ahead.

AI-native enterprises will increasingly run on continuous feedback loops, where operations, decision quality and customer outcomes can be measured and improved in near real time. Planning cycles won’t disappear, but they will compress, because capabilities and constraints can be learned faster than ever

Based on what we’re already observing, by 2030, we will see AI-native public companies share common characteristics – flatter structures, expertise-driven operating models, real-time decision-making and software that behaves like a living system, rather than one that requires periodic upgrades to function well. These companies will have integrated and adopted AI and rearchitected around completely new business principles.

While full transformation takes time, leaders can take specific actions now that will meaningfully accelerate their progress.

Many great enterprises fail because AI is introduced into rigid systems whose governance models were built on yesterday’s static business practices. Today’s leaders have to reverse that trend. This starts by working backwards from desired outcomes with partners who are operating in an AI-native way, not projecting the past onto your systems.

We know that AI systems learn from the context we humans encode into them. Without expert reasoning, however, the systems degrade and value is left on the table. Subject matter experts must be closest to the transformation. Make them 'intelligence managers', responsible for shaping how AI reasons, decides, reports and evolves. This shift is non-negotiable – transformations stall without active SME involvement and incentivization.

Start by bringing technology and the business closer together. Given the number of moving parts – technology, operating model and people – there is a need for different silos to work as part of a joint cross-functional team. This enables rapid iteration by removing the barriers between business, people and technology teams.

A small, but growing, number of enterprises are moving from experimentation to structural change. The difference between them and everyone else is that they see the value only AI can unlock and are redesigning their organizations for that future. Those still on the sidelines must choose to evolve for an AI-native world; otherwise, they’ll find themselves building for one that no longer exists.