How to successfully scale AI this year
If 2024 was the year for experimentation, then 2025 is the year to break the barriers to scaling AI.
The AI story of 2024 can be summarized as a hyperventilating time of exciting technological advancement and experimentation, albeit tempered by the realization of just how challenging it is to scale AI projects past proof of concepts.
Our quarterly surveys show that executives remain convinced of the value in AI investment, notwithstanding the skepticism that naturally followed the initial wave of enthusiasm for generative AI.
Fewer than 20 percent of companies are scaling their AI experiments.
Most companies have continued to invest in generative AI, increased the number of pilot programmes as well as deployments in production and remained highly satisfied with this powerful new technology.
However, we have also found that fewer than 20 percent of companies are scaling their AI experiments. The divide between those businesses that are scaling AI effectively and those who aren’t is widening.
I see this as an opportunity.
AI is not a tech rollout, it’s a CEO-led transformation
2025 will be the year when companies must prioritize scaling their AI deployments through substantial change management turbocharged by fervent business sponsorship.
Generative AI demands a fundamental business transformation, not just a technology rollout. Today’s AI tools rarely create value without a change in the business.
To be at the top of the AI game is to make wholesale changes. Companies and their leaders must make sure there is a marriage between their core AI plans and growth and their technology teams, employees and stakeholders.
2025 will be the year when companies must prioritize scaling their AI deployments through substantial change management turbocharged by fervent business sponsorship.
That shift is crucial as we find AI deployment is still being led either by the IT departments (in non-tech companies) or the engineering teams (in tech companies).
Other C-suite leaders, on both the business and technical sides, also have essential roles to play. CEOs should take on a mindset of reshaping how they do business, their company culture, ways of working and approach to innovation.
In short, successful AI deployments are enterprise-wide transformations.
Does your tech stack support scaling AI solutions?
The latest advancements in AI models might be generating the most excitement, but clients I talk to realize they need to invest in their technological functions, data management and foundational capabilities in order to scale to ‘AI everywhere’.
In our surveys, we see an increasing number of businesses identifying lack of data readiness as roadblock to their transformation plans.
Integrations and workflows will need to work better with AI models to enable more complex automation. Data memory and storage are other important considerations.
Technology owners in enterprise businesses may need to re-architect operational systems in order to build and maintain the quality, accuracy and safety of data, particularly when it comes to unstructured data. Integrations and workflows will need to work better with AI models to enable more complex automation. Data memory and storage are other important considerations.
2025 is the year when CIOs and other technology leaders, with the buy-in of CEOs, need to make these choices and plan accordingly in a landscape of rapidly evolving generative AI technologies. This requires prioritization and flexibility to adapt.
Agentic AI: fad or the future?
Generative AI is far from its tipping point.
While some trends will undoubtedly be more significant to business than others, one area to watch is what some are calling agentic AI or intelligent agents. Agentic AI in itself is not new and has been a topic of research for years, but we expect to see ‘true’ AI agents – functional and scalable – emerge in 2025.
Where generative AI created outputs based on prompts, agentic AI introduces systems that can act on your behalf, like a digital doer. On an individual level, for example, an AI agent can not only draft an email but send it, schedule follow-ups and track the response timeline. It could order groceries online or curate personalized travel itineraries.
The functions most likely to be the first to deploy true AI agents are those who were early adopters of generative AI, including software, sales, customer service, marketing and administration.
In the workplace, agentic AI could redefine business boundaries by automating and streamlining workflows, helping businesses to run leaner, faster and smarter. Governed by human supervision, agentic AI workflows are capable of interpreting defining goals, taking actions and making decisions to achieve them.
Agentic AI is requiring companies to reexamine and reimagine processes with AI-enabled automation, in contrast to automating individual components of an existing workflow.
The functions most likely to be the first to deploy true AI agents are those who were early adopters of generative AI, including software, sales, customer service, marketing and administration. And with that, I arrive at my final point.
Winning big requires leading boldly
More than half of the business leaders we surveyed said AI has already delivered business value to them, and three quarters acknowledged that AI poses a risk of disrupting their industries.
I think it’s unlikely that businesses will take their foot off the AI pedal in 2025. But the bad news for companies stuck in the AI experimentation mode is that the pace of innovation shows no signs of slowing.
Organizations that rethink their operations through AI-driven strategies can gain a significant competitive advantage.
If 2024 was the year for experimentation, then 2025 is the year to break the barriers to scaling AI. This is the only way to generate real return on the investment many companies have made in AI and to be ready for its next phases of development, including agentic AI. To make real progress in this space, CEOs must treat scaling AI as a change management challenge built on well-engineered applications.
This is an exceptionally challenging feat, one that few businesses have successfully navigated. However, organizations that rethink their operations through AI-driven strategies can gain a significant competitive advantage.
Success requires leadership from the CEO, meticulous change management and a foundation of high-quality data governance supported by enterprise technology tailored to the organization’s needs.