India Inc. and AI: 9 Indian companies that are doubling down on AI efforts

Earlier this year, Tech Mahindra announced a partnership with Google to launch a cleverly named Gen AI-powered email solution called “amplifAIer”, as part of its suite of artificial intelligence offerings.

OpenAI launched ChatGPT 3.5 almost a year ago - on November 30, 2022. ChatGPT set off widespread adoption of and debates around generative artificial intelligence (Gen AI), how it works, its uses, ethical and safety concerns around it, and whether others around the world could quickly build large models to train AI at a fraction of the cost.

(At the time of publishing on November 18, shy of the one-year anniversary of ChatGPT 3.5, OpenAI has fired its CEO Sam Altman for not being "consistently candid in his communications with the board", and the company's co-founder and president Greg Brockman has resigned).

As we approach day 365 since the launch of ChatGPT, let’s look at what’s been happening closer to home this past year with India Inc. and artificial intelligence.

Fear of too many options, simplified

If you’re of the opinion that the convenience of Swiggy can never be overstated, prepare to double-down on this further. Through generative AI, the food delivery platform is about make search and discovery for you, the user, even more intuitive, as well as bolster its ecosystem of restaurants and delivery partners.

Swiggy’s neural search feature, which was announced in July (whose pilot has been underway since September), is aimed at making decision-making easier. Users will be able to type in queries like “Show me healthy post-workout lunch options” and the search engine will use large language models and give the user a specific recommendation from its catalogue of over 5 crore dishes. The search is expected to be integrated with Instamart (grocery delivery) as well as Dineout (where gen AI will help with specific language parsing, such as ambience, pet-friendliness of restaurants and more).

Also read: Want to invest in AI in India? Here are top stocks with artificial intelligence play

HealthifyMe may have launched a health and nutrition focused chatbot some five years ago, but the technology has evolved significantly since then. Earlier this year, the fitness app raised US$30 million in funding to accelerate its AI capabilities. “Ria” is now a full-fledged, talkative AI companion prioritising your personal nutrition and training goals. The company is also crafting AI-enabled diet and fitness “Smart Plans” which will be available to users via a subscription model.

Content generation and process optimisation

In early September, Amazon launched a gen AI-enabled system which made it easier for sellers to create more engaging product listings. The new capabilities use LLMs to generate text to build more comprehensive product descriptions. Sellers can just give the system a very basic input and this will populate all the required listing fields and categories. Form-filling is often the most dreaded activity in human-company interaction, and this is indeed a welcome addition. What’s more, these comprehensive listings should make it easier for buyers to find products on the Amazon marketplace.

Processing, parsing, and using large volumes of data to optimise processes is made easier with AI. For instance, Razorpay’s AI-powered payments routing system, the company claims, will protect businesses from annual losses of over ₹7,000 crore due to transaction failures. The system is designed to handle over 5,000 online transactions per second.

Logistics solutions provider, Ecom Express, on the other hand, uses AI for routing, predictive demand forecasting as well as uses an address engine that can accurately pinpoint delivery location (solving the problem of inconsistent/non-standard formats). While Tata Technologies has managed to reduce power consumption on its shop floor by 18 percent using ML and AI for anomaly detection and tooling failure rates.

Services and beyond

Earlier this year, Tech Mahindra announced a partnership with Google to launch a cleverly named Gen AI-powered email solution called “amplifAIer”, as part of its suite of artificial intelligence offerings. The solution’s multi-fold goal is to first serve as a single solution to manage email comms of customer service contact centres, and additionally, provide end-to-end solutions for email automation with applications in customer relationship management and finance. While Google Cloud’s Gen AI capabilities are advanced, this solution is designed on the “human in loop” principle, which will allow organic intervention in reviewing AI-generated responses before they reach their final recipients.

IT-firm Happiest Minds is incorporating Gen AI through its new purpose-specific AI business unit. It appears to be taking a more foundational approach to the use of artificial intelligence, by actively upskilling its employees in Gen AI concepts. Through a partnership with Microsoft, the company will leverage large language models (LLMs) to aid in digital transformation initiatives, having identified over 100 use cases for the technology.

LTI Mindtree’s Canvas.ai is an enterprise-ready generative AI platform which allows clients to build, manage and consume Gen AI solutions. Alongside several pre-built business solutions, the company appears to be focused on tool-building, through the imminent launch of Canvas Lite (targeted at developers for productivity use cases).

Training humans is important, but training AI models is arguably even more important. TESA is Tata Exlsi’s smart annotation platform, which itself is AI-powered, automating the process of annotating data for machine learning models. This is done by creating simulations of real-world scenarios, allowing for accelerated training of AI models for specific tasks.

Looking back

Let us remember November 30, 2022, and the weeks that followed.  It was a significant period in the history of human beings on the internet. Where, for a moment, masses went from the darkness of social media doom-scrolling, to becoming enlightened artificial intelligence experts. Empowered by virtual keyboards and emboldened by divine purpose, humans became one with an intelligence that seemed to hold a conversation better than most other humans, while promising to make their life easier by drafting boring emails, serving complex information in digestible chunks (occasionally with reasonable accuracy), as well as theorising on the meaning of life itself.

It might have been a newfound enchantment for us, the common people, but we should be aware that boffins have been at this stuff for over half a century. In 1957, Frank Rosenblatt’s Perceptron was capable of pattern identification in data, while Joseph Weizenbaum’s ELIZA could engage in simple conversations with our kind, way back when England wiggled its way to a World Cup, in 1966.

Let us also not forget that capitalists have been looking to leverage technology to streamline process, reduce expenses, and create better and more cost-effective products since time immemorial. Machine learning and the use of AI models in pursuit of these ends has persisted, and their efforts have, unsurprisingly, increased exponentially over the past year internationally.