From Simple Prompts to Complex Workflows
Not long ago, the ability to write a clever one-off prompt for a large language model (LLM) was seen as a distinct advantage. It was about asking the right question to get a specific output—a marketing slogan, a block of code, or a summary of a document.
Today, that is considered table stakes. The real value is shifting towards orchestrating multi-step AI tasks. This means breaking down a complex problem into a sequence of smaller, interconnected prompts where the output of one step becomes the input for the next. Instead of just asking an AI to write an article, a multi-step workflow involves prompting it to first research topics, then generate an outline, draft each section, and finally refine the entire piece for tone and clarity. It is the difference between being a passenger and a pilot.
Why the Shift Is Happening Now in India
Several factors are driving this evolution in the Indian market. Firstly, businesses are moving beyond AI experimentation and are now focused on embedding AI into core operations for genuine productivity gains. They need reliable, scalable, and enterprise-grade solutions, which one-off prompts cannot deliver. Secondly, India’s role as the world's 'back office' and a massive hub for Global Capability Centers (GCCs) puts it at the forefront of this change. Companies are not just looking for cost arbitrage; they are building complex AI-powered systems in India. This has led to a surge in demand for professionals who can manage what is known as the 'orchestration layer'—the system that coordinates multiple AI models and tools to complete a business process. Hiring for roles related to AI orchestration and autonomous workflows has grown significantly as a result.
What a 'Multi-Step Task' Looks Like in Practice
Consider a financial analyst. A one-off prompt might be, “Summarise the latest quarterly report of Company X.” A multi-step task would be far more sophisticated. The analyst would orchestrate an AI workflow to: 1. Ingest quarterly reports from multiple competitors. 2. Extract key financial metrics and analyst commentary from each. 3. Perform a comparative analysis against historical data. 4. Identify trends and anomalies. 5. Draft an initial summary report with charts and visualisations. 6. Finally, ask the AI to play devil's advocate and highlight potential risks in the analysis. This is not prompt engineering; it is AI-augmented business strategy. This approach requires a blend of domain knowledge, logical thinking, and an understanding of how AI systems work together.
The New Skillset for Indian Professionals
This shift demands more than just technical proficiency; it requires a new mindset. Prompt engineering is becoming a foundational skill, not a specialised role. According to industry body NASSCOM, while many professionals are becoming 'AI-proficient,' the goal is to become 'AI-native'—able to use AI for deep engineering and problem-solving. The most in-demand skills now include strategic thinking, process decomposition, and AI governance. Professionals need to be able to map out a business process and identify where different AI tools can be chained together for maximum impact. NASSCOM and other organisations project a significant demand-supply gap for such advanced talent, with India expected to need over a million AI professionals by 2026-2027. This highlights a massive opportunity for those willing to move beyond basic prompting and develop these deeper, more strategic capabilities.
