
Hexaware Technologies Global CTO Satyajith Mundakkal has said most companies are still in the first stage of agentic AI adoption, where AI provides outputs for humans to act on. Speaking to The…
Hexaware Technologies Global CTO Satyajith Mundakkal has said most companies are still in the first stage of agentic AI adoption, where AI provides outputs for humans to act on. Speaking to The Federal, he said some clients have reached stage two, where the agent executes tasks after human approval, while fully autonomous stage three deployment remains rare and limited to security areas.

Mundakkal identified cultural resistance as a major barrier, with employees viewing AI as a job replacement rather than a complement. He also cited poor use-case selection and technological immaturity as other reasons for stalled rollouts. AI has accelerated document processing in insurance and banking by 30 to 40 times, and is speeding up drug approvals in healthcare. Hexaware has partnered with upGrad for systematic employee AI training.
Mundakkal's diagnosis of cultural resistance reflects a pattern seen across Indian IT services firms. Nasscom data from 2024 showed that about 60% of Indian enterprises were running AI proofs-of-concept but less than 15% had scaled them to production. The three-stage agentic AI framework he describes is not an industry standard but Hexaware's own model, which helps clients benchmark progress. The real constraint may be less about willingness and more about data readiness: under the IT Act and proposed Digital Personal Data Protection rules, financial and health data cannot be freely used to train models without explicit consent. The next data point to watch is how many of Hexaware's stage-one clients move to stage two within the next two quarters, as agentic AI tools from Microsoft and Salesforce begin to offer pre-built approval workflows.
Source: thefederal.com
This brief was synthesised by AI from the source linked above.