
Siddharth Dhar, President & Global Head – Digital IT Operations & AI, Hexaware
As Hexaware gears up to launch a new host of AI solutions in 2026, Siddharth Dhar, President & Global Head – Digital IT Operations & AI, Hexaware, speaks to businessline as to how employees should navigate in an Agentic world with pricing models and headcount adjusting for new AI models working on a token-costing format.
Edited excerpts:
What are your expectations with Agentverse Enterprise AI in terms of returns?
Agentverse will have an impact both on our top line and our bottom line. By the end of the year, productivity gains because of the agents deployed, will be reflected in the bottom line. Revenue gains will be reflected in the top line. These solutions have come due to customer demand. All 600+ agents were already deployed either in PoCS or in production with an existing customer but we aggregated that and allowed access to customers through the Agentverse platform.
Is Hexaware thinking of an AI+ human model for its pricing model?
Yes, 100 per cent. A lot of our pricing will move to headcount plus tokens. In fact, all the platforms that we are deploying have an in-built token cost in the platform.
Can you share any colour in terms of how this pricing model works?
All the models in IT are priced into quantity. So, the price was determined by headcount, hourly billing rate. Now, there will be two ‘Q’s and two ‘P’s. Wherever there is a human in the loop, there is a human price and a token cost. So, everytime the agents run, an inference is made, there is a token cost. There is a separate swim lane of work for the agents and for the human worl. Both get costed based on their inherent costing mechanism. For the human, the costing mechanism is the hours that the human spent for the work, and for AI it’s the number of tokens it spent to do the work.
Does the tokenisation outweigh the human effort from a costing perspective?
Right now, it’s quite balanced. Overtime definitely there’ll be more token utilisation, less human intervention because AI learns things and you develop more confidence with it. It will increase in the future.
Is there any particular task or agenda that most of customers use AI for?
There’s software development lifecycle as a business case that all customers are deploying. Similarly, deploying agents in managing and IT operations is something that sees a lot of traction. Legal is a great use case; getting contracts reviewed, built up. We’re also seeing traction for using AI as an enterprise chatbot, replacing the traditional interface with things like service desk.
Have your AI solutions affected workforce rationalisation?
It has and it will continue to do so. As an organisation, our stance always has been that AI is essentially an augmentation of the human capacity, not a full replacement. We don’t know where it’ll go six months from now but it’s a pretty fast-evolving field. So for the moment, I foresee a much slower need to hire and better utilisation of existing folks. Six months from now we may have a different opinion, but for the moment it’s clearly an augmentation and capacity expansion play and not a human replacement play.
Are you worried that this might possibly affect or lead to any sort of revenue cannibalisation?
There will be revenue cannibalisation but we’re not worried about that because of our size. There’s a lot larger market share that we can gain by being ahead of the curve on something like this. That’s the advantage of being the size that we are – there’s a lot more to gain than to lose.
What are your plans going ahead?
We have a whole calendar of agentic announcements laid out for the rest of the year. Everything will be done by industry or customer. This is the year for Agent AI for Hexaware.
Published on March 23, 2026
Source link
#tokenisation #pricing #outweigh #human #effort #future #Hexawares #Siddharth #Dhar


