Deloitte Study Suggests AI Agents More Effective Than GenAI for Enterprise Productivity

New Delhi: A new study by Deloitte, a British professional services firm, suggested that Artificial intelligence (AI) agents could be a more effective tool compared to large language models (LLMs) or GenAI applications. These AI agents can help open up new possibilities to drive enterprise productivity and program delivery through business process automation.

AI agents are autonomous systems that use AI to interact with their environment, gather data, and complete tasks without needing human help. The study explains that tasks once considered too complex for GenAI can now be handled securely and effectively with AI agents.

The study also highlights the difference between GenAI and AI agents. While LLM-powered chatbots can only follow the "input-output" pattern of traditional apps, they often struggle with tasks that need multiple steps. 

"They (LLM or GenAI) conform to the 'input-output' paradigm of traditional applications and can get confused when presented with a request that must be deconstructed into multiple smaller tasks. They also struggle to reason over sequences, such as compositional 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 smaller volumes of data, typically sacrifice depth of knowledge and/or quality of outputs in favour of improved computational cost and speed," the study said.

GenAI is typically used in simple applications like creating personalized ads based on user search history or reviewing contracts. In contrast, AI agents can handle more complex tasks and use specialized digital tools for the job.

"AI agents equipped with long-term memory can remember customer and constituent interactions--including emails, chat sessions and phone calls--across digital channels, continuously learning and adjusting personalised recommendations. This contrasts with typical LLMs and SLMs, which are often limited to session-specific information," the study added.

The study also points out that businesses will need multiple AI agents working together, as a single AI agent has its limitations. However, AI agents come with new risks, such as bias in decision-making and vulnerability to data breaches or cyberattacks, which could compromise sensitive information.

"A significant risk is a potential bias in AI. Algorithms and training data, which can lead to inequitable decisions. Additionally, AI agents can be vulnerable to data breaches and cyberattacks, compromising sensitive information and data integrity," the study concluded.

(With agency inputs)