PUNE, INDIA—As India prepares for the AI Impact Summit on February 19 and 20, the Indian government’s pitch for wider adoption of artificial intelligence (AI) has centered on the potential for the technology to benefit Indian society. As my colleague Trisha Ray wrote in November, India “appears to be taking a people-centered approach, emphasizing use cases that have the greatest scope for positive impact for the widest swath of the population.”
During my time in Pune, a technology hub in Maharashtra, I spoke with local students, scientists, tech startup workers, and farmers. Though we met to talk about other topics, they consistently brought up AI and how they wished to take advantage of the technology. As I had more of these conversations, I found that to benefit the widest swath of the Indian population, the country should adopt a “sachet” approach to AI as a consumer product. The idea of applying the idea of sachets, or small packets used to package small consumer products, to AI is not new in India. But so far, this model has lacked proofs of concept and investment from the private sector, which has instead attempted to expand access by offering low-tier subscription models.
The sachet approach takes inspiration from the model that sparked India’s consumer goods revolution of the 1980s. Prior to that, Indian consumer products such as shampoo, talcum powder, and hair oil were sold in quantities of 50 grams (g) to 500 g. In the late 1970s, entrepreneur Chinni Krishnan found a niche in selling these products in cheap miniature packets, or sachets, containing as little as a few grams. By the mid-1980s, this more affordable consumer model made a wide variety of products more accessible to broader swaths of the Indian population.
Currently in India, a monthly ChatGPT or Perplexity Pro subscription costs ₹1,999 ($22.17) and a monthly Google AI subscription costs ₹1950 ($21.62). AI companies do recognize the need for less expensive and more accessible options, as the low-tier subscription services ChatGPT Go and Google AI Plus both cost ₹399 ($4.42) per month. Moreover, Google AI Pro and Perplexity Pro are also available for free for a year to eligible college students. And Perplexity AI partnered with telecom giant Airtel to offer a year’s free access to Perplexity Pro to Airtel’s 360 million subscribers. But this still leaves a huge portion of the potential Indian consumer market for AI untapped.
To make AI accessible to the widest possible swath of the population, AI developers should offer not just cheaper monthly subscription models but also sell the equivalent of a sachet of AI. This means offering small-scale uses of AI tools and applications for low fees. For example, one notable approach that has already been adopted is the IndiaAI Compute Pillar, which allows scholars, researchers, and startups to utilize computational power for less than a dollar per hour. To make this a scalable consumer product, however, the private sector would need data from the government on how the Compute Pillar is being used. Such data could make Compute Pillar a proof of concept for the AI sachet model. Under India’s AI Governance Guidelines, metrics for both the scale of adoption and how consistently the service is used could set the bar for whether this proof of concept should spur a larger-scale investment in such services.
India also has ample experience with scaling up such society- and accessibility-driven models. The Aadhaar biometric ID system, the Unified Payments Interface instant payment system, and the country’s digital public infrastructure (DPI) buildout were bottom-up models. For example, from 2011 to 2021, the number of Indian adults (ages fifteen and up) with a bank account rose from 35 percent to 80 percent thanks to this approach.
As an illustration for how this sachet model could be of use, think of the places where shampoo and other kinds of sachets are sold in India—usually small mom-and-pop stores run by one person or family. For such small stores, bookkeeping can be a laborious, time-consuming task. But with a ₹15 AI sachet, a shopkeeper could take photos of that day’s transactions, prompt an AI to parse the handwriting, and calculate revenue and inventory figures. If small business owners were to widely adopt AI sachets for such tasks, it would be a significant step toward demonstrating the scalability of the AI sachet model. This is how shampoos and other consumer goods expanded their footprints using the sachet model.
During my trip to Pune, many of the people I spoke with were curious about how AI can help improve efficiency in areas including business, scholarship, research, management, and farming practices. When it comes to harnessing this demand for wider AI adoption, the government can play a major role in bringing stakeholders such as unions, cooperatives, and trade associations together with private sector AI developers to demonstrate the utility of AI for their respective fields.
