Agentic artificial intelligence (AI) is the next disruptive form of the technology and will be key to monetising language models, according to a new report.
The fifth edition of GlobalData’s Artificial Intelligence Briefing outlines that progress in large language models (LLMs) is slowing down, and it suggests that their long-term use cases remain unclear. It notes that small language models (SLMs) trained for specific verticals are cheaper and faster to deploy, and it suggests they are likely to see greater adoption by enterprises as a result.
Smarter leaders trust GlobalData
Moreover, SLMs can be better suited to agentic AI due to the greater levels of accuracy that can be achieved with them compared to LLMs, greater operational efficiency through requiring less computing power and greater propensity for integration across ecosystems due to their smaller size and resource demands.
GlobalData’s report suggests that agentic AI agents can be considered “an evolution of AI assistants”. They are capable of carrying out tasks autonomously by making decisions and taking actions within the context of a given environment, whereas generative AI agents require human inputs to produce outputs. In short, they have the capability for agency rather than only generation.
In practice, this means agentic AI could respond to a customer query by accessing their account in a CRM system and issuing a refund, while generative AI could merely produce an email response to the query. Agentic AI could schedule a hospital patient for relevant tests and alert doctors about urgent symptoms, while generative AI could simply provide a report summarising symptoms and potential conditions.
Rena Bhattacharyya, practice lead for enterprise technology and services at GlobalData, commented: “Many organisations have been experimenting with generative AI with varying degrees of success. What was an opportunistic stance to generative AI in the past, will evolve to a more structured and strategic approach that incorporates analysis of where the technology can make the greatest business impact.”
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Thank you!
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below form
By GlobalDataGlobalData cites the key features of agentic AI as being autonomy, goal-oriented behaviour, adaptability, decision-making, independent operation, problem-solving and creativity, flexible applications, self-optimisation, context awareness and collaborative scalability.
“Non-AI automated agents have existed for several decades across manufacturing, logistics, electricity grids, private and public transportation and mining,” its report states. “Non-AI multi-agent systems were later introduced to enhance quality and productivity.
“In the LLM era, these agents have become AI-enabled, using machine learning and LLM technologies to transmit data and control machines in real time. Due to the complexity of operational technologies, multiple AI agents will be deployed, each with their own knowledge base, to work together on separate but related tasks.
“It is anticipated that operational technologies that necessitate low latency and high levels of security and data ownership will implement private cloud to run their agentic AI programs. The public cloud may still be used for backup and archiving processes.”
Among the agentic AI startups currently operating are the likes of Ada, which provides customer service and claims near-100% customer query resolution; Maisa, which can produce software code; and AIQ, which is aimed at resolving challenges in drilling performance, reservoir management and quality control in the oil and gas sector.
