Over the past two years, the conversation around artificial intelligence within organizations has been dominated by one question: Which is the best model?
Every new release of ChatGPT, Claude, Gemini, Llama, Mistral, Kimi, Qwen or DeepSeek generates headlines, comparisons and another race to adopt the “smartest” model. Yet these headlines can also confuse the discussion, drawing our attention to the branches and leaves while making it harder to see the forest.
For most organizations, the question of which model they are using is beginning to lose strategic relevance.
The reason is relatively simple: artificial intelligence is entering a process of commoditization.
From a management perspective, we have seen this phenomenon before.
Today, dozens, perhaps hundreds, of models are available, ranging from proprietary commercial offerings to open-source alternatives. Some run in the cloud; others can be deployed within an organization’s own infrastructure, and increasingly even on individual workstations.
Competition among providers has accelerated innovation to the point where differences between models, while still important for specialized or highly demanding use cases, are becoming less decisive for a large proportion of everyday business activities.
One of the many frontiers currently being explored is how to run truly large models, such as GLM 5.2 with its 744-billion-parameter Mixture-of-Experts architecture, on desktop-class hardware. Proofs of concept may still be relatively slow, but the ability to run increasingly sophisticated models locally, approaching the capabilities required for many tasks currently handled by commercial cloud-based LLMs, is rapidly moving closer to practical reality.
From a management perspective, we have seen this phenomenon before. As technologies mature, the technology itself gradually stops being the primary competitive differentiator and becomes part of the underlying infrastructure.
- It happened with electricity.
- It happened with the Internet.
- It happened with cloud computing.
- And it is beginning to happen with artificial intelligence.
This does not mean that all AI models are equal. Frontier models will continue to lead in complex capabilities and push the boundaries of what is technologically possible. But for thousands of business tasks, many models already provide performance that is more than sufficient.
In other words, intelligence is becoming abundant.
And when a resource is no longer scarce, the source of competitive advantage begins to shift. The question for a CEO or Board of Directors therefore becomes less about:
Which AI model are we using?
The more strategic question becomes: How can we enable our people to work better by taking advantage of this abundance of intelligence?
The distinction may appear subtle, but it can fundamentally change where organizations choose to invest. Many companies are discovering that one of the greatest barriers to capturing value from AI is no longer the capability of the technology itself. It is the way people interact with it.
technology designed to increase productivity can end up adding another layer of complexity to the working day.
It is increasingly common to find AI-proficient professionals moving among multiple applications, using several AI models, transferring information between tools, rewriting instructions, searching for documents, constantly switching context and using different assistants for different tasks. The result is a paradox: technology designed to increase productivity can end up adding another layer of complexity to the working day.
At the same time, operating cost is becoming increasingly important.
Recent analysis by Grey Journal highlights how rapidly AI token expenditure can escalate as organizations expand their use of generative AI. Regardless of the quality of the output, the economics of this way of working can eventually become a material business consideration.
As organizations scale their use of AI, token consumption and the indiscriminate use of frontier models can become a significant component of total operating cost, potentially reaching hundreds or even thousands of dollars per user per month in intensive-use scenarios.
This is creating renewed interest in hybrid approaches: combining high-performance cloud models with more efficient cloud-based alternatives and locally deployed models, while reserving the most expensive computational resources for those activities where their additional capabilities genuinely create value.
All of this points toward an important conclusion.
The next major transformation will not simply come from having access to better artificial intelligence models. It will come from redesigning the way people experience and perform their work.
For decades, organizations have invested heavily in ERP, CRM, BPM and automation technologies to optimize business processes. Yet knowledge work has remained surprisingly manual.
Executives, analysts and managers still spend a substantial portion of their working day searching for information, transferring data between applications, preparing repetitive documents and coordinating activities that could increasingly be supported by artificial intelligence.
Within a few years, asking which AI model an organization uses may become almost as strategically irrelevant as asking which Internet browser its employees use to access corporate applications.
The opportunity, therefore, is no longer simply to add AI to existing activities.
It is to rethink how people work.
What will matter is an organization’s ability to integrate multiple models, multiple AI agents and multiple sources of information into a coherent, efficient and human-centered work experience.
Artificial intelligence will continue to evolve at extraordinary speed. But the next source of competitive advantage will not reside solely in the intelligence available to the organization.
It will reside in the organization’s ability to orchestrate that intelligence and turn it into a significantly better way of working.
Across the organizations where I focus my work: EXYGE Consulting, e-Consulting Global Solutions in its work supporting other consulting firms, and the Red Iberoamericana de Consultores Foundation initiative, we believe this will become one of the most important strategic management conversations of the coming years.
Because digital transformation has never been an end in itself. It has always served a larger purpose: Helping organizations and their people work better.
From our archive. Originally published on LinkedIn on 10 August 2026. Read the original. A Spanish version first appeared on our sister company’s blog.