For the past few years, the business conversation around artificial intelligence has focused on adopting the most powerful models available.
In 2026, that conversation is changing.
Businesses are no longer asking only, “What can AI do?” They are also asking: what business outcome will it deliver, and how much will it cost at scale?
Recent industry news demonstrates why this question matters. Reuters reported that Uber used its entire 2026 AI budget in only four months after employees rapidly adopted AI coding tools. The same report found that businesses are increasingly turning to AI-routing platforms and lower-cost models, reserving premium models for work that genuinely requires their advanced capabilities.
Source
Reuters — Cheaper AI is better: Soaring bills are reshaping how businesses choose models
This is not an isolated challenge.
Gartner reported in January 2026 that at least 50% of generative-AI projects had been abandoned after the proof-of-concept stage by the end of 2025. Poor data quality, inadequate controls, escalating costs and unclear business value were among the main reasons.
Source
Gartner — Why 50% of GenAI Projects Fail
Bain & Company’s 2026 research tells a similar story. Nearly 40% of companies that measured their AI cost savings achieved savings below 10%, despite initially targeting reductions of between 11% and 20%. However, 90% of the companies surveyed were still increasing their AI budgets.
Source
Bain & Company — Your AI Budget Is Growing. Your Returns Aren’t
The message is clear: companies are not abandoning AI. They are demanding better economics, stronger governance and measurable returns.
The danger of using one expensive model for everything
Not every e-commerce task requires the largest or most expensive AI model.
A complex decision involving multiple data sources may need an advanced reasoning model. However, a straightforward task — such as categorising a product, extracting standard information, routing a customer query or checking content against a defined rule — may be completed effectively by a smaller and less expensive model.
When every request is automatically sent to a premium model, costs can increase rapidly. This is especially important in e-commerce, where thousands or millions of small daily tasks can turn a minor per-request expense into a significant operational cost.
The solution is not to use less intelligence. It is to use intelligence more intelligently.
The CholaVerse approach
At CholaVerse, we work every day to identify which parts of the e-commerce journey can be automated effectively using AI agents.
Our goal is not to add AI simply because it is fashionable. We want to apply it where it can deliver clear operational value.
That is why we are building an intelligent AI router.
The CholaVerse AI router is designed to select the most cost-efficient suitable AI model for each task. Straightforward work can be routed to efficient, lower-cost models. More advanced models can be reserved for situations in which their additional capabilities are genuinely required.
Cost
Avoiding unnecessary use of expensive models.
Capability
Selecting an AI model appropriate for the complexity of the task.
Speed
Considering how quickly a task must be completed.
Scalability
Making it commercially practical to automate high volumes of e-commerce activity.
Instead of following a one-model-fits-all strategy, CholaVerse is working towards a flexible model in which different AI systems can be selected according to the needs of each workflow.
Learning from the first generation of enterprise AI
Klarna provides a useful example of how enterprise thinking is evolving. After becoming one of Europe’s most prominent early adopters of AI, its chief executive acknowledged that the company had focused too heavily on using the technology to reduce costs. Klarna subsequently adjusted its strategy and resumed hiring people, while continuing to believe that AI could help it improve services and products.
Source
Reuters — Klarna shifts AI focus from cost cuts to growth
The lesson is not that AI does not work. The lesson is that sustainable automation requires the right use case, the right model and the right balance between cost and quality.
The future is smarter AI usage
The next stage of AI adoption will not be defined only by who has access to the biggest model.
It will be defined by who can apply the right intelligence to the right task at the right cost.
That is the future CholaVerse is working towards: practical AI-agent automation for e-commerce, supported by intelligent routing and a clear focus on business value.
