AI's next leap is about getting work done
From 'ask' the AI to 'task' the AI: how agentic AI changes how businesses measure value, with examples from Cars24, SouthArc Digital and Harness.
Until recently, we have mostly used AI to ask questions, generate content, write code and summarise information. Useful, yes. But the work still usually stopped with us.
That is starting to change. AI can now take a task, split it into steps, use software and tools, check its own work, fix mistakes, and keep going until the job is complete.
As Jaspreet Bindra, Co-founder of AI&Beyond, puts it: "We are moving from 'ask' the AI to 'task' the AI."
That changes the way businesses should think about value. "How good is the answer?" still matters. So does a harder question: "How much of the task can the AI complete in the real world?"
Where it is showing up
- Software development: AI can inspect an error, look through the relevant files, run tests, and come back with a working fix.
- Business operations: AI can reconcile figures, prepare reports from raw data, conduct market research, and move information between applications.
- Automation: SouthArc Digital used AI to simplify a fragile software layer connecting different systems.
- Enterprise use: Cars24 is testing AI in software development and internal tools, especially where requirements are incomplete and several systems have to work together.
Assistant or agent?
An assistant helps with a task. An agent gets a goal, access to tools, and some room to act. It can plan, take action, check what happened, and adapt without someone directing every step.
For businesses, the unit of productivity is beginning to shift: from how many people a job takes to how much work a single person can manage and coordinate.
People still matter
Domain expertise, judgment, context, and the ability to break a messy problem into workable parts become even more valuable when AI handles more of the execution. People define the problem, set the destination and decide what good looks like.
The advantage will come from turning AI into repeatable workflows that people can trust, monitor, and improve over time.
Source: Times Techies, The Times of India (Akhil George).
- Agentic AI
- AI Agents
- Enterprise AI
