- At first AI was a tool, now it has become a more grounded and capable technology that has the potential to change the business at the infrastructure level Â
Over the past year, AI has moved away from experiments with generative AI models towards governed, scaled, agentic systems embedded in products, workflows and devices.
Its usage has gone from pockets of enthusiasm to widespread adoption as companies dig into the value it offers and find better ways of extracting this value.
Its narrative has also changed from the two extremes of it being considered magic or dangerous. Now, there’s a more grounded reading of what’s actually happening inside companies that have moved beyond experimentation.
And this move is something that was noticeable to industry leaders long before it became fashionable. As far back as 2024, there was a webinar, which argued that AI was already exiting its novelty phase and becoming something more consequential.
The prediction was that AI’s evolution to embedded infrastructure would be driven by architecture, platforms and human behaviour.
A year later, the analysis stands. As do other predictions made about the potential of the technology and how it is set to challenge the fabric of the organisation over the next year.
Trend 01: The end of AI as a pilot project
Companies were playing around with the idea of AI and what it could do, but they were stuck at the point of data.
Many were concerned about whether or not their systems and data were ready for AI, or if they could benefit from the technology effectively.
By the end of 2024, many South African companies were already moving away from proof-of-concept tools towards how they could use AI as a differentiator.
It marked the moment AI became a part of strategy and planning, and acceleration was driven by familiarity.
AI became normalised and companies stopped marvelling at outputs and instead started to interrogate how those outputs were produced, governed and secured. This reframing has defined the past year of enterprise AI more than any other single model release.
Trend 02: Platforms replaced products
The AI as a tool category has collapsed into consolidation around platforms. This trend was predicted long before it became mainstream as experts rapidly identified how Google, OpenAI and Microsoft had stopped competing at a feature level and instead moved to competition at an ecosystem level. The expectation was that AI would be distributed through platforms people already inhabited rather than sold as a standalone product.
Over the past year, this has played out exactly as AI became more pervasive as a layer inside documents, workflows, collaboration tools, and analytics.
Trend 03: AI empowered hybrid employees
While hybrid working has been on the agenda since 2020, AI was expected to raise the floor but lower the ceiling, giving employees access to data and capabilities that would allow them to become more productive and effective.
Those who were open to relying on AI to delegate work and who became attuned to the nuances of when to use AI, when not to, and how to add value, gained more control and visibility in their roles.
They have also been able to more effectively work from anywhere and connect with intelligence to systems and people without many of the traditional hiccups that come with hybrid working environments.
Trend 04: Trust is a technical problem
While the discussions around AI ethics remain abstract and in flux, one prediction made was more specific – that trust would become an operational challenge. Deepfakes, synthetic audio, and fabricated content have become catalysts for an entirely new enterprise category – verification. AI is not only creating the trust deficit, but it is being used to close it.
Over the past year, this tension has sharpened as companies are looking for ways to authenticate reality at scale. An emphasis on AI-enabled security, watermarking and verification anticipated a market change that is now underway.
The insights gathered by Braintree in their 2024 webinar, that led to this level of accuracy in AI predictions, came from long-term exposure to enterprise architecture and platform economics.
Working inside environments gave the company insight into pressures facing companies on the ground, in real time.
Looking ahead, these predictions are one step on the AI ladder. The technology will continue to create friction and value for companies, and it will continue to introduce entirely new ways of working and thinking.
In 2026, the prediction is that the key word will be convergence as AI becomes more embedded with a greater emphasis on security, governance and ethical use. A trend that’s already taking centre stage across corporate discussions and government regulations.
And as for any other predictions? Wait and see.
