How Infosys Plans To Become AI-First Company

'We have been following the micro-change philosophy, where every six weeks we implement these AI features into our internal systems, products, platforms and then release it to our employees.'

Infosys is set to achieve its target of becoming an AI-first company.

The IT services giant uses artificial intelligence to "amplify" employees' potential and "reimagine" services, Rafee Tarafdar, chief technology officer and executive vice-president of Infosys, told Business Standard's Avik Das and Shivani Shinde in a video interview.

How does Infosys use AI and where do you see the process going?

When we talked about our aspiration to be an AI-first company, there were three parts to it.

One, we said we will use AI to amplify the potential of our employees.

Second, we will use it to reimagine our services.

And third, we will use it to reimagine the industries in which our clients operate.

On the amplification part, we have provided an AI assistant to almost all employees, depending on their persona, be it learning, development, HR, sales and marketing, shared services, and support.

Almost 90 per cent of the company is AI-aware. We have good numbers of AI builders and AI masters.

About the second part of reimagining services, there are broadly three areas.

One is running the IT and infra and Cloud and operations for most of the companies.

We have embedded AI in all services that we offer as managed services.

Second, we are also using AI in transformation programmes now -- be it in SAP, be it in Salesforce, or data transformation programmes.

We have also set up AI labs with many of our clients which are used to innovate and build first-of-its-kind AI solutions contextual to them.

About the last part, we have developed various industry blueprints on how AI can be used.

We just launched an agent foundry where we have launched a number of agents for enterprise functions as well as industry-specific ones.

This is an area that we are a lot more focused on, working with clients to help them become leaders in their industries.

How many such labs are there? Are these labs in the client ecosystem?

We have two types of living labs.

One is the Infosys living labs, where we showcase all our offerings.

AI living labs are... generally a joint, co-innovation approach between us and our clients, where both of us invest and set teams up.

This is where we do all the first-of-its-kind AI implementation, build proof of value, and sometimes work with partners.

Infosys has delivered 400 GenAI projects. How are you helping clients with AI, GenAI and agentic AI?

Value is being delivered in four broad areas. It could be cost, which is tied to productivity, efficiency and velocity.

It will be an improvement in net promoter score (NPS), which is from (the point of) customer experience and service.

The third will be revenue and growth opportunities, and the fourth is risk protection and compliance.

When it comes to cost, through GenAI and agentic AI, we are able to improve system availability by 90 per cent.

We are able to reduce by half the time it takes to resolve issues or address customer queries.

This means I am able to give better service to customers, which reflects in improved NPS score.

We have already seen a lot of value in projects related to cost, NPS, and risk compliance.

Revenue and growth are evolving because this also requires a bit of process re-engineering, business model re-engineering, and this should happen in the coming months and years.

Can you tell us about GenAI deals?

I think everybody is in implementation mode.

In terms of the nature of these deals, there will be two types of programmes.

Those which are pure play AI projects, which are largely about doing new AI work.

Second is existing work which could be a run programme, a transform programme, a change programme, where AI is getting embedded.

Can you take us through the Agentic Foundry and the need for it?

When we started agentic implementations, we saw that there are a lot of patterns that were repeated across most clients.

This could be either for IT operations, or it could be in core business areas like accounts payable, invoice processing or supply chain.

So why don't we create these agents out of the box so that we can help deliver value faster to clients.

Here are two things. One is where we had our own IP (intellectual property) and platforms that we were using to deliver our services; there we implemented a lot of these agents out of the box.

The second, which is very industry-specific, are reference implementations that our clients can use to jumpstart their agentic projects.

From a cultural point, how do you ensure that AI is embedded in the company?

When we started on this journey to be AI-first, we recognised that we have to be responsible by design.

It means all AI solutions and products being developed are done in an inclusive manner.

When it comes to culture, we have been following the micro-change philosophy, where every six weeks we implement these AI features into our internal systems, products, platforms and then release it to our employees.

They see this working and create a pull effect.

Feature Presentation: Ashish Narsale/Rediff