SambaNova Systems has published the results of its global research on AI adoption within enterprise organizations. It found that business leaders are increasingly deploying AI and progress could be further accelerated by moving beyond a fragmented proliferation of small models. Enterprise leaders are placing AI at the core of a multiyear technology strategy and two-thirds (67%) believe it will be transformational or significantly change how they do business in 12 to 24 months.
THE DRIVER BEHIND THE CHANGE THAT AI WILL DELIVER
When asked about the type of change that AI will deliver, global business leaders cited the top three drivers:
• 80% think that AI will improve the employee or customer experience by streamlining processes and decreasing response time
• 68% think AI will cut costs by automating processes and initiating a better use of head count
• 51% will use AI to increase profit through better use of data or opening new revenue streams
There is another reason behind adoption of AI – it is increasingly becoming a competitive asset. The research found that almost three-quarters (72%) of business leaders believe their competitors are using AI, and of those almost two-thirds (63%) are concerned their competitors will use AI to gain an advantage over their own business.
Marshall Choy, SVP Product at SambaNova commented on the findings: "Enduring enterprises always keep a sharp eye on technology as a way to rise above competitors. Just like railroads, radio, and the internet have done for previous generations, AI is reshaping business as we know it and over the next decade, early investors and adopters stand to yield the greatest benefits."
Businesses need to consolidate models creating an enterprise-wide strategy based on large models
One of the biggest challenges enterprises face is the number of AI models currently deployed in production. Only 18% of organizations utilizing AI are deploying it as a large-scale enterprise-class initiative. The rest - 82% - are introducing it across multiple programs, which can create unexpected hurdles and a less coherent AI strategy.
However, in the era of general purpose large language models, there is now a better way.
Choy elaborated on the benefits of large language models: “You’ll be hard-pressed to find an enterprise that runs more than a handful of relational databases. Most organizations have consolidated their databases, which means they are well understood, maintainable, secure and auditable. This hasn’t happened with AI models yet.”
“Most organizations that have a significant AI footprint have been left with a myriad of hundreds or even thousands of disparate models,” stated Choy. “They are not easily manageable, and certainly not auditable. This is where a single foundation model for language can be the firms’ AI backbone at enterprise scale as the basis for all AI applications and workflows for the next decade.”
