Why voice, not keyboards, may define AI's next chapter

For years, the artificial intelligence (AI) race has revolved around one question: who can build the biggest and smartest model? Companies have poured billions of dollars into developing more powerful systems, while researchers have competed on benchmark scores that measure an AI model's capabilities. But the next phase of AI may not be decided by computing power alone. Instead, it could be shaped by something much simpler, how naturally people can interact with technology.

According to Dheemanth Reddy Bhumireddy SingaReddy , Co-founder and CEO of Maya Research, the industry's focus is gradually moving away from technical achievements towards making AI useful for everyday users. In his view, the real breakthrough will come when people can simply speak to technology instead of typing commands or navigating complicated apps.

The shift from powerful AI to practical AI

The first wave of AI innovation largely centred on creating increasingly capable models. While these advances remain important, businesses are now asking a different question: How can AI solve real problems for more people?

Dheemanth believes benchmark scores were useful during AI's early years, but they are no longer the ultimate measure of success. "Benchmarks are how you keep score in the early years. They were never the game," Reddy says.

He argues that AI will have achieved its true potential only when it becomes accessible to people regardless of their education, language or digital skills. That includes users who may never have been comfortable typing in English or navigating complex digital interfaces.

Why voice could become AI's natural interface
For decades, keyboards have been the primary way people interact with computers and smartphones. However, Reddy believes that was more a limitation of technology than a natural human behaviour. "Human beings have spoken for tens of thousands of years and written for only a few thousand," he says.

Voice allows users to communicate in the language they think in instead of adapting themselves to machines. Rather than memorising commands or searching through menus, users can simply ask questions, explain what they need or continue a conversation naturally. As AI becomes better at understanding speech, the keyboard may gradually lose its central role for many everyday digital tasks.

Why India matters in the voice AI race
India presents one of the world's most challenging environments for voice technology. People routinely switch between English and regional languages during conversations. The country also has dozens of widely spoken languages and hundreds of dialects, making speech recognition far more complex than in many other markets.

Dheemanth believes this challenge could become India's competitive advantage. "We are the hardest language market on earth. If you can make voice AI work here, it will work almost anywhere else in the emerging world."

Instead of following global technology trends, he believes India has an opportunity to lead innovation in multilingual conversational AI.

Industries that could benefit the most
Voice AI has applications far beyond virtual assistants. Retail is one obvious example. Before online shopping became mainstream, customers often spoke directly to local shopkeepers to compare products, negotiate prices or ask for recommendations. Digital commerce replaced many of those conversations with search boxes and menus.

Conversational AI could bring those interactions back into the online world. Banking and insurance companies, which handle millions of customer conversations every day, could also benefit from AI systems capable of understanding multiple Indian languages. Healthcare and education are other sectors where natural conversations often matter more than forms or buttons.

According to Dheemanth , AI should adapt to the way people communicate instead of forcing people to adapt to technology.

Businesses now want results, not hype
The enthusiasm around AI has matured. Companies are no longer investing simply because AI is the latest technology trend. Instead, they are looking for measurable improvements in customer service, operating costs and productivity.

Dheemanth says organisations respond better when AI is presented as a business solution rather than a technological achievement. For businesses that spend heavily on customer support, AI becomes valuable if it can improve conversations while reducing costs.

Smaller AI startups are finding new opportunities
Building cutting-edge AI models was once considered possible only for technology giants with enormous budgets and computing resources. While large foundation models remain important, Reddy believes innovation is increasingly shifting towards startups that focus on solving specific problems with specialised AI systems.

"A small team that picks a sharp, specific problem and designs its architecture around that problem can now do work that used to demand a war chest," he says.

This changing landscape could create more opportunities for Indian AI startups working on regional language technologies and enterprise solutions.

Trust will decide AI's future
For voice AI, accuracy alone is not enough. Users quickly notice if a digital voice sounds robotic, unnatural or responds with noticeable delays. Such experiences can reduce trust almost immediately.

Dheemanth says conversational AI must feel effortless, reliable and human-like if it is to become part of people's everyday lives. The future envisioned by voice AI developers is one where technology fades into the background. Instead of opening multiple apps or typing search queries, users could simply speak naturally, switch languages mid-conversation and expect AI to understand context without friction.

If that happens, AI's biggest achievement may not be building the most advanced algorithms. It could be making digital services easier to access for billions of people, whether they are shopping, learning, seeking healthcare or managing finances.

As AI moves from laboratories into daily life, the next chapter may be written not with keyboards, but with conversations.