Token economics explained: what AI agents really cost to roll out

Token economics in under two minutes: base salaries plus token spend plus a one-off enablement layer, why first-year costs go up before they come down, and the 3x productivity choice between more output or fewer people.

A seven-step explainer titled 'Token Economics Explained'. (1) Start with a team of 5 on base salaries. (2) Add AI token spend of $5k to $10k per person per year (non-coding roles), so $25k to $50k for the team. (3) Add the one-off enablement layer (training, workflow redesign, tech integrations, governance and process change) without which token spend won't convert to productivity. (4) With the right setup each person does about 3x the work, giving 5 people the capacity of 15. (5) This gives two choices: Option A, keep the team and increase output; or Option B, keep output the same and reduce headcount. (6) First-year costs go up, not down: base salaries plus token spend plus training plus enablement plus change management, all before full productivity lands. (7) ROI comes after the productivity gains; done well, in 12 months or less. Bottom line: invest first in tokens and enablement, then unlock productivity: either more output with the same team or the same output with fewer people.

Token economics explained in less than two minutes.

When you move from AI assistants (like Copilot or ChatGPT) to AI agents, you quickly run into token economics. Here's the simple version, with an example. Real-world applications are rarely this clean, but the shape holds.

Start with a team of five

Five people, each on their base salary. That's your starting point.

Add token spend per person

For non-coding roles, budget roughly $5,000 to $10,000 per person per year in AI token spend. Across a team of five, that's $25,000 to $50,000 a year, before anyone has saved a single minute.

Add the enablement layer

This is the part most businesses skip, and it's the part that makes everything else work. Training, workflow redesign, tech integrations, governance and process change. It's mostly a one-off, upfront cost.

Without it, your token spend won't convert into productivity. You'll just have expensive people using expensive tools the old way.

With the right setup, each person does more

Get that setup right and each person can do roughly 3x the work. Five AI-enabled people now have something close to the capacity of fifteen.

That gives you two choices

  • Option A: keep the team, increase output. The same five people plus AI spend deliver a 15-person-equivalent output. More scale without adding headcount.
  • Option B: keep output, reduce headcount. The same result from fewer people, at a lower cost over time.

Same 15-person-equivalent output, delivered by five people or by two. Which one you choose is a strategy decision, not a technical one.

First-year costs go up, not down

Here's the part that catches people out. In year one you're paying base salaries, plus token spend, plus training, plus enablement, plus change management, all before the full productivity gains land. Spend goes up first.

ROI comes after the productivity gains

Done well, the return lands within twelve months. The order is what matters: invest first in tokens and enablement, unlock the productivity, then take it as either more output with the same team or the same output with fewer people.

Bottom line: token spend on its own is just a bill. Token spend plus enablement is what turns into ROI.

If you're looking for help applying this to your organisation, let me know.

Keep reading — it's free

Pop in your email to keep reading and join AI Dispatch, our newsletter of practical advice for leaders scaling AI. Unsubscribe anytime.

Related

Why isn't your AI rollout delivering ROI?

If you're rolling out AI and not seeing a return, it's probably because you rolled it out functionally, not organisationally. You're only as fast as your slowest point, and AI productivity dies at the handoffs between departments.