For decades, the equation was simple: you paid people for their time, and time translated into output. Hourly rates, day rates, monthly salaries—the entire machinery of employment and contracting runs on this assumption.
A provocative new paper circulating in policy and economics circles, titled “Who Prices Cognitive Labor in the Age of Agents?”, argues that this model is about to break. When an AI agent can draft a contract, analyze a dataset, or write marketing copy in seconds, what exactly are you paying a human professional for?
The Core Argument: Anchor Wages to Compute
The paper’s central thesis is uncomfortable but hard to dismiss. It suggests that as AI agents become the primary engines of knowledge work, compensation models should be anchored to compute costs rather than human hours.
Think about it this way: if a legal associate and an AI agent both produce a due diligence report, but the agent does it in three minutes for a few rupees of compute cost, the market will eventually refuse to pay associate rates for the same output. The paper argues this isn’t a distant future—it’s already happening in pockets of the economy.
The authors propose that hybrid work—where humans supervise, edit, or augment AI output—needs new pricing frameworks. A “compute-anchored wage” would factor in what portion of the work was done by AI, what compute resources were consumed, and what unique human judgment was added.
What This Means for Indian Outsourcing
India’s $250 billion IT and business services industry was built on labor arbitrage—skilled workers at lower costs than Western markets. This paper suggests that arbitrage window is closing, not because Indian wages are rising, but because AI compute is falling.
When a US company can run an AI agent for pennies per task, the cost advantage of an offshore team billing by the hour diminishes. The question for Indian service providers becomes: are you selling human hours, or are you selling outcomes?
Companies like Infosys, TCS, and Wipro have been talking about AI-augmented delivery for years. But their billing models still largely depend on headcount and effort. The paper suggests that clients will increasingly demand outcome-based pricing—and will want transparency on how much of the work was done by humans versus AI.
HR and Procurement Feel the Pressure
For founders and CXOs, this shift creates immediate practical questions. How do you structure contracts with vendors when you’re not sure if you’re paying for human expertise or compute capacity? How do you compensate employees who use AI tools to 10x their output?
HR technology vendors like Darwinbox, Keka, and Zoho People will need to build new frameworks. If an employee uses company-provided AI tools to deliver more, does their compensation reflect that productivity? Or does the company capture all the gains?
Procurement teams face similar puzzles. When hiring a marketing agency or a consulting firm, the RFP process may need to specify: what AI tools will be used, who pays for the compute, and how is pricing adjusted based on AI contribution?
Gig platforms like Upwork and Fiverr are already seeing this play out. Freelancers who use AI tools can deliver faster—but clients are starting to question why they should pay the same rate for AI-assisted work.
The Measurement Problem
The paper acknowledges a significant hurdle: we don’t yet have good ways to measure AI contribution to cognitive work. Unlike manufacturing, where you can count widgets, knowledge work output is fuzzy.
Some startups are working on this. Tools that track AI usage in workflows, platforms that separate human and AI contributions in code repositories, systems that log prompt-and-response cycles. But these are early days.
The risk for companies that ignore this shift is being caught on the wrong side of a pricing correction. If your competitors figure out how to bill for outcomes at compute-anchored rates while you’re still selling hours, your margins will disappear.
What This Means for You
If you run a services business, start experimenting with outcome-based pricing models now—before your clients demand them. Understand what percentage of your deliverables involve AI and build that transparency into your proposals.
If you’re a buyer of services, ask vendors directly: what AI tools are you using, and how does that affect your pricing? Don’t pay 2019 rates for 2025 delivery methods.
If you lead HR, begin thinking about compensation frameworks for AI-augmented roles. The employee who produces twice the output using AI tools is valuable—but so is the AI investment. Deciding who captures that value is a conversation you need to have now, not after it becomes a retention problem.
The billing-by-the-hour era isn’t ending tomorrow. But the paper makes a convincing case that the clock is ticking.