The Tools of Tomorrow: What Lies Ahead with the AI Revolution

Artificial intelligence has broken free from its awkward, glitch-ridden phase and emerged as a revolutionary tool for success. It is no longer the bumbling intern of technology—accidentally endorsing illegal activities or discriminating against certain age groups. AI has sharpened its edge, becoming the crème de la crème of human innovation, capable of rewriting the fabric of reality itself.

The question now isn’t what AI can do—it’s what’s left for humans to bring out of AI. Will it stay limited to chatbots? Not a chance. The future of AI promises to be a journey into the unknown, where these highly capable assistants don’t just take over the mundane—they anticipate our needs and act with precision. We are already at the heart of an AI revolution. Are we ready for what’s to come?

Mint’s most recent All About AI episode delved into this very topic—what the future holds for AI innovation and how Salesforce is contributing to this journey. This exciting session featured highly esteemed industry experts, including Srinivas Tallapragada, President and Chief Engineering Officer, Salesforce, and Sanket Atal, MD – Operations and Technology at Salesforce India. Watch the full episode below,

The Journey So Far and How Agentic AI Is the Future

Imagine a world where machines don’t just follow instructions but also think and create. That is the promise of modern AI. To get to where it is now, at the pinnacle of human technology, AI has taken monumental strides. One of the most significant advancements came in the form of generative AI. Srinivas Tallapragada, in a bold statement, compared the impact of generative AI to the invention of the internet: “ChatGPT is as big as the internet, maybe even bigger. It has expanded the consciousness of everybody and has dared us to dream about what’s next.”

Generative AI models have enabled modern businesses to automate tasks, generate insights, and enhance customer experiences. But the real game-changer lies in the rise of AI agents—intelligent systems that can perform tasks, make decisions, and collaborate with humans seamlessly. Digital platforms such as Agentforce can be the ideal solution, helping businesses build highly customized AI agents that perform specific duties.

The Challenges with Agentic AI: From Concepts to Products

While the potential of AI agents is immense, their adoption comes with various challenges. Many modern businesses have been experimenting with GenAI platforms and copilots, exploring their capabilities and creating strategies to utilise them. However, most of these experiments remain where they are—as experiments. Very few have actually been promoted to becoming fully implemented solutions. But why is this so?

The answer lies in the complexity of deploying AI agents. Srinivas Tallapragada broke it down: “Just like a human, an AI agent needs a role, like a job description. It needs access to information, knowledge retrieval systems, and the ability to take action through the necessary access to APIs and legacy systems. It also needs guardrails to ensure trust, compliance, and safety.” What’s more, AI agents must evolve continuously. “You must be ready to rapidly grow and evolve,” he added.

Salesforce addresses these challenges by leveraging its expertise in understanding the complex roles, such as those of sales or service representatives. This allows Salesforce to effectively train AI according to what needs to be done, using past data and experience. The results speak for themselves: Salesforce’s AI agents already answer 93% of customer questions before routing them to human agents.

How AI Agents Empower Startups

For startups, scalability is a critical challenge. This is where AI agents come into the picture, serving as the ideal solution to this obstacle. Sanket Atal explained, “When you launch a product, you hope lots of people use it. But that can be an issue. How do you handle the questions and concerns from your growing customers?” In this scenario, AI agents enable startups to not just scale, but scale economically. “Agentforce gives you the ability to conceptualise, create, and deploy agents. But it also comes prepackaged with each of our agents as well, so you’re able to get things out of the box and customise them to your data needs,” he added.

Safeguards and Trust: Tackling Hallucinations and Bias

While modern LLMs are pushing the boundaries of creativity, this superpower can turn into their downfall, resulting in risks like hallucinations, biases, and toxicity. To overcome these challenges and ensure confidence, a trust layer is essential. Just like a professor, the system should not only give you an answer but also provide citations and sources. It should filter for bias and toxicity, and be tailored to language, country, nuance, and industry. This trust layer is what makes AI agents viable for real-world applications.

Scale, Talent, and Innovation: India’s Role in the AI Revolution

When it comes to AI, India is not just a participant—it’s a powerhouse. With its massive population, rapid digital adoption, and thriving tech ecosystem, India is uniquely positioned to lead the charge in the AI era. Sanket Atal captures this sentiment perfectly: “When I think about India, I think of scale. Seeing AI in action so that the greater population is addressed would be super exciting to me. That’s the promise of India. Plus, our adoption rate of digital technologies has been phenomenal. Likewise, with AI and the agentic era, I think we’re going to play a dominant role in its adoption.”

Srinivas Tallapragada provided a unique perspective, focusing on India’s incredible AI talent pool and the potential to drive the agentic AI sector forward. He explained, “We do cutting-edge work and collaborate across the world. We have a lot of AI and cloud teams here, with amazing talent.” From developing advanced solutions to solving real-world problems, Indian teams are at the forefront of innovation.

The Next Frontier

As the session drew to a close, the next wave of AI innovation stood out as the prominent topic of discussion. Srinivas Tallapragada painted a vivid picture of what’s coming: “Currently, LLMs are running on text and video. Next, we have world models. Most of us learn through sensing—that’s our primary source of information, not just text and videos.”

This shift toward sensing models represents a leap forward in AI capabilities. Imagine a personal robot that doesn’t just process data but can also see, hear, touch, and interact with the physical world. Such advancements have the potential to revolutionise industries like healthcare, manufacturing, and even everyday household tasks. And that’s where we’re going. As AI rapidly evolves and India places itself at the forefront of its innovation, these are no longer distant dreams but tangible possibilities.

Disclaimer: This is a Mint editorial initiative, sponsored by Salesforce.