By Sandeep Alur, CTO, Microsoft Innovation Hub -India
As we approach the two-year milestone of ChatGPT’s debut, generative AI and AI assistants have significantly impacted various fields. From multilingual tools like Bhashini, which have simplified villagers’ access to government services to copilots that are redefining productivity and efficiency standards in organizations, their influence has been substantial.
Both Large Language Models (LLMs) and Small Language Models (SLMs) have also significantly advanced in their reasoning capabilities, becoming more responsible, empathetic and aligned to people.
AI itself is evolving and while ChatGPT, copilots and other AI assistants rely on human triggered prompts and do not have autonomy to operate on their own, AI assistants are now progressing to operate as agents. Agents are more of task-solvers, equipped with skills to plan, reason and execute.
Agentic reasoning is no longer science fiction
Imagine you need to board a flight but still have some unfinished tasks to complete. You wish you could stay connected and complete your work, but you need to switch your device to airplane mode! Now, wouldn’t it be incredible if you could only instruct your AI Agent before take-off to handle a few tasks, and loop back once you land at your destination? This is no longer science fiction but a reality in the making.
Tasks that are currently performed by humans using natural language can be handed off to AI agents to execute. For example, a financial institution can have a KYC Agent, specialized in this process, that can be called by any individual or a process. Similarly, in a BPO organization, an AI agent can take over a few tasks, such as customer conversations and complete them, basis human approval.
So how do agents work? Agents are designed to act as human proxies, emulating the way we plan and execute tasks. Consider, you need to create a two-page report. You would plan how to structure the content, conduct the necessary research, collate data, draft the report, reason over the content for appropriateness, perform final checks, and complete the task.
Agents mimic this process, and when there is a need for a human to be in the loop, AI agents notify the user, seeking intervention. However, these models operate under boundary conditions, aligning to the guardrails provided by the instructions. Hence their autonomy to plan and execute tasks is within the set boundaries.
This transition to functioning as task solvers opens a new paradigm in how we view knowledge assistants, complementing the need for automation. Agents can help wherever automation is a key lever as part of business process optimization. From an adoption perspective, lot of downstream business processes stand to benefit significantly from agentic workflows.
Redefining the future of productivity with agents
The future of productivity lies at the intersection of technology and human ingenuity, and AI agents are poised to be the catalysts of this transformation.
In India, the enthusiasm for AI adoption is palpable. According to the Microsoft and LinkedIn 2024 Work Trend Index, an astounding 92 per cent of knowledge workers in India already utilize AI in their daily tasks[1]. This reflects a broader trend of employees seeking to leverage AI to save time, boost creativity, and focus on more intellectually stimulating work.
To put this into perspective, EY’s AIdea of India report estimates that generative AI has the potential to contribute between $359 billion and $438 billion to India’s GDP in 2029-30, over baseline estimates[2]. This economic impact is mirrored globally, where AI is expected to add trillions of dollars annually. Analogous to the transformative power of electricity, AI can revolutionize industrial output and enhance the quality of life.
The advancement to AI agents promises a future where human capabilities are significantly enhanced, unlocking new thresholds of prosperity and ushering in a new era of creativity and innovation. As sophisticated task solvers, AI agents will become the true productivity multipliers.
The technology building blocks are now in place, and the responsibility lies with us to identify the tasks that can be delegated to these agents. By doing so, we can harness their potential to achieve more, drive economic growth, and foster a more creative and productive society.
Disclaimer: This article was first published by moneycontrol.com in India on October 2, 2024.
[1] https://news.microsoft.com/en-in/92-of-indian-knowledge-workers-use-ai-in-the-workplace-finds-microsoft-and-linkedin-2024-work-trend-index/
[2] https://www.ey.com/en_in/services/ai/generative-ai-india-report/economic-opportunity-of-gen-ai-india
