Software major Infosys on May 23 unveiled Topaz, a new offering that combines data analytics, artificial intelligence, and generative AI, signifying the growing importance of this fast-evolving technology.
Infosys, one of the early donors to OpenAI, said it has used its own applied artificial intelligence (AI) framework to develop an AI-first core that potentially has over 12,000 use cases, more than 150 pre-trained models, and over ten platforms.
The company’s efforts in AI also come at a time when technology demand is slowing down in its major markets such as the US and Europe. A new AI-first offering could help the company offer a more compelling pitch for clients, who have turned increasingly cautious about tech spends.
“We are seeing strong interest from our clients for efficiency and productivity-enhancing programs, even as businesses are keen to secure their future growth. Our own business operations have been hugely benefited by Infosys Topaz bringing the power of generative AI platforms and data solutions,” Salil Parekh, CEO & MD, Infosys, said.
Speaking to Moneycontrol, Satish HC, Executive Vice President and Co-Head of Delivery at Infosys, said the company is first experimenting with AI within the organisation, taking experiences from that and then curating solutions for clients. Infosys counts Fortune 500 banks, retail firms, among others, as clients.
“Our clients will hear a lot of advisory on how to go about it, how to scale this, what to do, what not to do, and how do we navigate some of the risks that we see today,” he said.
Competitive intensity is also going up among IT firms seeking to build more in this space.
Cognizant’s CEO Ravi Kumar S had said during the company’s earnings last quarter that breakthroughs in generative AI offer the potential to “fundamentally transform our client's businesses and increase our own productivity”. The company said it working to industrialise solutions to common challenges.
“We are operating pilots that use generative AI to accelerate consulting, design, engineering, and operations with the long-term goal of doubling the productivity of our associates,” he said.
Similarly, Tata Consultancy Services Chief Operating Officer NG Subramaniam said that the company will have a centre of excellence around generative AI. Subramaniam said that generative AI can accelerate the adoption of technology, including those which are at the frontier areas of innovation.
Opportunities and pricing
According to Satish, Infosys is already engaged in generative AI work on many projects with their existing clients, and they are currently involved in over 300 client conversations around this. Satish said that the company decided to get into the space because of the massive opportunity it presents.
Given that Infosys donated to OpenAI in 2015, when it was structured as a non-profit, Satish said that the IT firm does not have a stake in the company. However, he said that many of the things that OpenAI is providing for businesses are things that Infosys was building internally.
However, generative AI varies for enterprise use cases versus publicly available models such as ChatGPT.
“If your engine is going to rely on public signals, generative AI will give you those public signals readily consumable. I don't need to create a brain and train it to consume signals from outside. But if I must build an engine, which will have to consume signals from an enterprise or its ecosystem, I still have to custom build it,” he said.
In addition, Satish said that they can provide their clients with their learnings, experimentation within the enterprise, and valuable insights.
This also means that a suite of services such as this will be priced at a premium — especially given their business model is also that there is better pricing in newer areas. He added that the revenue impact will be big, but it will take time to scale as people are experimenting and watching for clarity.
Use cases and concerns
Satish sees three established use cases. “This has the ability to deliver a new class of solutions and business models, which are ecosystem led, or it has the ability to disrupt your business processes so that they're more intelligent and sentient, or your people become more empowered or productive,” he said.
Satish said that generative AI has led to a 30 percent reduction in the effort required for analytics work.
However, this doesn’t mean that generative AI doesn’t come with its problems, the fact that they hallucinate or that safety and security of information may be a concern, among others.
To tackle this, Satish said, “One, the quality of data must be curated, especially in regulated areas. Second, we have to build explain-ability and traceability and we have to have humans in the loop.”
People and jobs
Satish said that Infosys already has a large data, analytics, and AI team that needs to learn incremental skills. The company is now focusing on prompt engineering.
“We have to help the average or the real user learn to use this technology. They need to understand at least some parts of nuances of prompt engineering — how to leverage the technology, when to use it, when not to use it. We have heard examples of people who have , inadvertently leaked out competitive data. You have to be very clear in classification of data assets within the organization, where you can use public domain large language models, where you use narrow transformer model, or an internally trained transformer model, which is specific to the enterprise,” he said.
Will jobs become redundant?
From the time ChatGPT exploded onto the scene, there has been chatter about what kind of, and how many jobs it can make redundant.
Satish said that he doesn’t see an immediate impact, given the generations of technologies that have come before. With generative AI, he said can solve higher order problems
“We are getting better at asking the right questions, framing the problem. Over a period of time that somebody said, this is the century of right brain people, a lot of us in our jobs will have a higher portion of design and reimagining what we do, which is essentially built in. The consulting industry is going to be democratized with this because knowledge is democratised,” he says.
He maintained that it will not immediately reduce headcount, but over a period, learning will change and help people become more productive.