Download the latest issue of Business Today Magazine just for Rs.49
Indian IT Bets On AI

Indian IT Bets On AI

As Indian IT services companies move beyond run-of- the-mill businesses and towards digital transformation, artificial intelligence is taking centre stage.

Illustration by Raj Verma Illustration by Raj Verma

I can understand how the limited perspective of an un-artificial mind would perceive it that way. You'll get used to it": Samantha taunts Theodore Twombly when he calls her just a computer voice.

These are lines from the critically acclaimed 2013 movie Her based on artificial intelligence (AI). Today, a 'Her' is closer to reality than ever. Chat bots, voice-based AI solutions and cognitive robotic process automation (RPA) are now the order of the day. Big services players like Infosys, Wipro, TCS, Cognizant and HCL Tech, which embarked on the AI and automation journey less than five years ago, see upwards of 30 per cent of their total revenues coming from these digital segments.

A look at the revenues of the top three IT companies tells the emerging story. Market leader TCS saw more than 50 per cent year-on-year revenue growth (constant currency terms) in its digital business in the third quarter of FY19, which now contributes 30.1 per cent to the overall revenues. Infosys also registered a 33 per cent revenue growth year-on-year in the same period, with digital business brining in 31.5 per cent of the overall revenues. Wipro saw a 35.4 per cent year-on-year increase with digital business now accounting for 33.2 per cent of the overall revenues.

Though the term 'digital' is yet to be standardised, most companies use it as an encompassing term to include automation and AI. Even though IT services companies are yet to break up services lines and business, signs of automation and AI becoming turnkey elements for both revenue and operating margin have started becoming clearer.

What's on Offer

Over the past few years, homegrown IT companies have been establishing separate units to focus on automation and AI offerings. Infosys has Nia, TCS has Ignio, Wipro has Holmes and HCL Tech has DryiCE. They are following the lead of large global companies like Accenture (myWizard), IBM (Watson) and Microsoft Azure ML.

Work is on to turn partially or fully manual, repetitive and rule-based tasks through RPA and adding an AI layer that can recognize patterns, images, sounds, pictures and analyse data to predict outcomes, forewarn on critical issues, mimicking human intelligence of reasoning and deduction.

Harrick Vin, Global Head, Digitate, and VP, TCS (Photograph by Anirudha Karamarkar)

For instance, a chatbot or bot could be used for a simple RPA solution for employees applying for leave through a chat window. A systems engineer in a telecom company could use an AI solution for a more complex situation such as predictive maintenance of telecom towers and getting the system to fix them.

In 2013, when N. Chandrashekaran was the boss and Rajesh Gopinathan was the CFO, TCS laid the foundation for Digitate, which eventually went on to house Ignio, its AI platform. "The whole concept of creating Ignio or more importantly, Digitate, was that we have to be true to ourselves; we have to stand up as a product," recalls Harrick Vin, Global Head, Digitate. Vin was earlier the Chief Scientist at the Tata Research Development and Design Centre and also the Global Head for Innovation and Transformation for IT Infrastructure Services. Today, Ignio is tripling its revenues annually, and aims to reach a $100 million in revenues by next year.

For Infosys's Nia, it was a change in avatar - from 'Mana' and before that from Products Platform and Solutions business. The fully owned subsidiary, Edgeverve, which houses Nia, has a core engineering team of 280-300 focussing on AI.

Nia adds a cognitive layer to both the RPA platform AssistEdge, and to all the business application offerings of the company by adding bite-sized consumable AI. Atul Soneja, Global Head of Edgeverve, says Nia's approach is comprehensive. "You can pick and choose the different components, as well as have some industry-specific solutions, which are readymade and easy to deploy," says Soneja.

Wipro, which launched Holmes in early-2015, much of the work was a part of its own transformation. Like most peers, this open source platform too adds cognition to its automation and business offerings for improved user experience, delivery efficiencies and unit economics. Bhanumurthy BM, President and COO, Wipro, says, "The advantage is that you have an intelligent tool but you don't necessarily have to leave everything to the tool. You could still guide it to do things in your fashion."

With over 90 bots and 350 customers, automation initiatives have helped Wipro improve revenue per employee by 4.1 per cent on a year-on-year basis.

The latest entrant is HCL Tech with DRYiCE, which sells AI through a productised platform. It has augmented its offerings with acquisitions and intellectual property partnerships. The company has made over eight acquisitions, including the recent $1.8 billion acquisition of IBM products. "We are taking some of the modern, relevant contemporary pieces from the acquisitions and running them like a software product business," says Kalyan Kumar, Global CTO of HCL Technologies.

Distinguishing Features

Since most of these offerings sound similar in function, what distinguishes them? In infrastructure and application maintenance, there is a lot of similarity. "For example, what Holmes does for SAP and what Ignio does for SAP are very similar kinds of task automation," says Tapati Bandopadhaya, Vice President of Research at HFS. Distinction comes in when companies provide end-to-end service delivery. "I would say Wipro and Infosys are relatively better off. They are (more) distinctive than the late starters, but even they have a long way to go before they reach the level of, say, Accenture myWizard," says Bandopadhaya.

Most companies have a similar two-pronged market approach - taking the AI automation standalone services to new as well as traditional services clients, and bundling it with other offerings to build a quicker pipeline. From product licensing to pay-per-use kind of a model, pricing of these AI offerings is as dynamic as the AI ecosystem itself. In terms of adoption of the AI offerings, the banking and financial services sector leads, followed by manufacturing, retail, and energy and utilities.

The share of revenues from the adoption of these newer technologies is much in line with the segment-wise revenue of IT companies and therefore most companies now have dedicated sales force for these units. This is not surprising; according to a September 2018 IDC forecast report, Worldwide Spending on Cognitive and Artificial Intelligence Systems, global spending on AI systems is estimated to reach $77.6 billion by 2022, and everybody wants their share from this.

Making a Mark

If offerings are similar then how do the customers choose? The answer to that lies in a mix of several variables such as talent, investment into newer technologies, business models, and adoption by existing clients.

With AI rapidly changing service delivery models and clients moving away from the traditional time and material model, a cultural change is inevitable. While companies are taking different routes based on their technical abilities, a tangible market leader is yet to emerge. Also, the Indian services companies are yet to reach the levels of IBM or Accenture, which have superior data ecosystems that are enabling AI development.

Bhanumurthy BM, President and COO, Wipro Ltd

"In terms of technology usage, Wipro and Infosys will fare better because they had been investing in this (AI) context. In terms of talent, TCS and Infy (Infosys) have an edge by virtue of having their own product companies," says Bandopadhaya. However, HCL Tech's business model stands out. "HCL, of late, has been offering their chatbots by the unit of work - clients pay as they consume. It is a very simple but very intricate kind of a model," she adds.

No doubt these AI entities hold the key to the future of the IT services companies, but who will lead the pack? Many factors will come into play, two of which would be focus and expertise in a sector and talent pool, says Raja Lahiri, Partner, Grant Thornton.

"Otherwise, you could easily get lost in the quagmire," he says. As skill sets such as data mining analysis and designing are set to become the next Java, most companies are investing heavily in re-skilling their employees. Wipro, for instance, had close to 130,000 digitally trained employees at the end of December 2018. In the past year or so, TCS has re-skilled over 292,000 employees.

On the downside, several IT companies have been trimming their workforce. In 2016, research firm HFS had predicted that India was set to lose nearly 650,000 low-skilled positions to automation by 2021. With 'digital adoption' by clients - both as a means to save costs and increase efficiency - now a norm, for IT services companies it is either push AI or perish.