The AI conductor era
The most productive workers in 2026 aren't the ones typing the fastest. They're the ones orchestrating the most AI agents. And a new metric is emerging to measure it: token burn, the volume of AI compute a worker consumes in the course of doing their job.
Token burn as a productivity signal
Kevin Roose reported in the New York Times this month that companies including Meta, OpenAI and Shopify now track employee token consumption on internal leaderboards. One OpenAI engineer processed 210 billion tokens in a single week. A single Claude Code user at Anthropic ran up a $150,000 bill in a month. These aren't cautionary tales. They're the new performance benchmarks.
Jensen Huang made the framing explicit at Nvidia's GTC conference last week: engineers should receive roughly half their base salary again in tokens. His logic is simple. Every engineer with access to more compute will be more productive. Tokens aren't a perk. They're infrastructure.
This tracks with what Shopify's Tobi Lütke laid out a year ago when he told employees that AI usage is a baseline expectation, baked into performance reviews. Teams must now demonstrate why AI can't do the work before asking for headcount. The direction is clear: output per person is the metric that matters, and AI agents are how you multiply it.
The role shift is already happening
The transition from individual contributor to AI orchestrator isn't theoretical. It's playing out across functions right now.
Developer. The engineer's job has shifted from writing code to designing the environment where agents write it. As I wrote in my last post on harness engineering, the best teams are building AGENTS.md files, enforced architectural constraints and garbage collection agents. The developer who averages 3.5 merged PRs per day isn't typing faster. They're running more agents in parallel and reviewing the output.
Customer service rep. Contact centers are becoming AI-first experience hubs. Gartner projects that 80% of common service issues will be handled by agentic AI by 2029. The human agent's job is shifting from answering calls to supervising AI chatbots, handling escalations and coaching the systems that handle the routine. Forrester predicts daily agent workloads will drop by an average of one hour as AI automates narrow tasks like FAQ creation and call summaries.
Sales rep. The SDR role is being rebuilt around AI pipeline management. Tools like Salesforce's Agentforce and Amplemarket's Duo now handle prospect research, lead qualification, personalized outreach and meeting scheduling autonomously. The human rep focuses on relationship-building and closing. Research from McKinsey found that 66% of organizations using generative AI in sales reported revenue increases.
Financial analyst. Wolters Kluwer reports that 44% of finance teams will use agentic AI in 2026, a 600% increase from the prior year. Agents are matching invoices to purchase orders, flagging discrepancies and predicting collection outcomes. The analyst's job is becoming less about building spreadsheets and more about directing the agents that build them.
Project manager. Gartner predicts 80% of project management tasks will be handled by AI by 2030. Already, AI agents are automating sprint planning, ticket triage and resource allocation. The PM's role is evolving from task tracker to orchestrator of human and AI workflows.
Why token burn isn't just another vanity metric
The skepticism is fair. Knowledge work has a long history of bad productivity proxies: hours logged, messages sent, lines of code written. Token burn could easily become the same kind of performative busywork. The phenomenon even has a name now: tokenmaxxing, where employees compete on leaderboards without tying consumption to shipped outcomes.
But token burn has one structural advantage over previous metrics. It's tied to actual work being delegated to AI systems. An employee burning tokens is, at minimum, asking agents to do something. That's a fundamentally different signal than hours in a seat or Slack messages sent. The risk isn't the metric itself. It's leadership failing to connect token consumption to outcomes. A team burning $50,000 a month in tokens while shipping nothing is a management problem, not a measurement problem. The organizations getting this right are correlating token usage with output metrics: PRs merged, deals closed, tickets resolved, reports generated.
The conductor's podium
The shift happening in 2026 is not about any single tool or role. It's about a new operating model for knowledge work. The most effective workers are becoming conductors: people who understand the capabilities of their AI agents, know how to compose them into workflows and can evaluate whether the output meets the standard.
This is the AI Conductor Era. Your value isn't in how much you can personally produce. It's in how much you can orchestrate. The organizations that recognize this, that invest in token budgets, agent infrastructure and outcome-based measurement, will outperform those still counting keystrokes. The baton is there. Pick it up.
Join the discussion on LinkedIn.