Companies are increasingly focused on gouging out returns from their investments in generative artificial intelligence (gen AI) as they explore complex uses of the rapidly evolving technology.

A joint study by Intel and IDC estimates spending on AI in India could increase to $5.1 billion by 2027 from $1.7 billion in 2023.

In 2022, banking & financial services and manufacturing invested the most in AI.

A global study by McKinsey showed 23 per cent of organizations have assigned numbers on the impact — with at least 5 per cent of their earnings before interest and taxes (EBIT) attributable to AI.

The uses of gen AI are diversifying across sectors: from simple chatbots and AI assistants to underwriting insurance policies, identifying stress in loan accounts and helping retail chains strategise store schematics and personalise product descriptions for customers.

Brokerages are using gen AI to update their algorithms and continually optimise trading strategies based on real-time portfolio performance and evolving market conditions.

In healthcare, hospitals are exploring the use of virtual care assistants with conversational language capabilities to assist with patient queries.

Guru Bala, head of solutions architecture for specialised services at AWS India and South Asia, said enterprises are looking to deploy gen AI and solve multiple business problems after grasping the technology.

“In the past few years, people were trying to understand what the technology was all about. In 2024 we have crossed that bridge and enterprises are now looking at how to deploy gen AI and solve multiple business problems,” Bala said.

“Today most of the organisations that we talk to are looking at integrating gen AI into their existing applications, workflows and processes and augmenting everything they do with gen AI.”

He said some enterprises have a precise return on investment in mind, while others need assistance in identifying the uses.

“The primary business benefits of gen AI lie in productivity gains and better customer experience,” he said.

As an example, Bala said AWS is working with a financial services organization that has around 3,000 call center agents. The company’s board has tasked the finance and IT teams with identifying cost gains and improving operational efficiency.

By leveraging gen AI, large language models (LLMs) can read customer email queries, fetch the loan status, compose a response in the original language of the mail and allow any necessary corrections to be made. This process reduces the time spent in addressing each customer query to around 1-1.5 minutes from 7-8 minutes.

Bala emphasised the need for responsible AI frameworks to mitigate risks. These risks include input data biases and potential legal or safety consequences of generating unethical, irrelevant, incoherent or sensitive data outputs.

“It is always a shared responsibility. At AWS, we attempt to define, measure and mitigate all kinds of risks that come along with AI.

“But customers should have a responsible AI framework within their organisation where they should have clear guidance in terms of how they should be using gen AI tools within their organisations,” he said.