Artificial Intelligence (AI) Agents can be a more effective tool than large language models (LLMs) or GenAI applications, opening new possibilities to drive business productivity and program delivery through business process automation, said the British professional services company Deloitte in a study.

With the help of an AI agent, cases that were previously deemed too complex for GenAI can now be activated at scale safely and efficiently, the study says.

By definition, the AI ​​agent is an autonomous intelligent system that uses AI techniques to interact with its environment, collect data and perform tasks without human intervention.

• Also read: Adoption of artificial intelligence gaining momentum thanks to private banks: RBI study

Explaining the difference between AI generation and AI agents, the study adds that typical LLM-based chatbots usually have limited ability to understand multi-step prompts.

“They (LLM or Gen AI) conform to the “input-output” paradigm of traditional applications and can be confused when presented with a query that needs to be deconstructed into several smaller tasks. They also have difficulty reasoning about sequences, such as composition tasks that require consideration of temporal and textual contexts. These limitations are even more pronounced when using small language models (SLMs) which, because they are trained on more. small volumes of data, generally sacrifice depth of knowledge and/or quality of results in favor of improved computational cost and speed,” he says.

The study indicates that GenAI use cases are mostly limited to standalone applications such as generating personalized ads based on a customer’s search history and reviewing contracts, among others.

On the other hand, AI agents excel at overcoming limitations while leveraging the capabilities of domain- and task-specific digital tools to efficiently accomplish more complex tasks. “AI agents with long-term memory can remember customer and constituent interactions (including emails, chat sessions, and phone calls) across digital channels, continuously learning and adjusting recommendations personalized. This contrasts with traditional LLM and SLM, which are often limited. to session-specific information,” the study adds.

The study further adds that while individual AI agents can offer valuable enhancements, businesses also need multi-agent AI systems, given the limitations of single AI agents. However, the study notes that AI agents also introduce new risks that require robust security and governance structures.

“A significant risk lies in potential biases in AI algorithms and training data, which can lead to unfair decisions. Additionally, AI agents may be vulnerable to data breaches and cyberattacks, compromising information sensitive and data integrity,” he adds.