OpenAI recently added a premium, higher-cost seat tier to ChatGPT Business, the latest reminder that most AI vendors price primarily on seats, a fixed cost per person regardless of how much or how little that person actually uses the tool. For an AU business comparing licensing costs, that pricing shape matters as much as the headline number, because it decides whether your bill scales with your headcount or with the actual work being done.
Seat-based versus usage-based, and why it matters
A pure per-seat model charges the same whether an employee uses the tool for five minutes a day or five hours. That's simple to budget but can badly misprice a team where usage varies wildly, a few power users running agents constantly, most staff using it occasionally. Claude's plan structure spans this range, from per-seat Team and Enterprise plans through to usage-based agent and Cowork work, which matters specifically for AU businesses where a small number of automated workflows might do more actual work than a large number of light individual users.
What actually drives cost at AU SMB scale
Seat count: straightforward to estimate, scales linearly with headcount regardless of usage intensity
Usage-based agent and Cowork work: scales with what's actually automated, a handful of scheduled tasks running daily can outweigh dozens of light individual chat users
Tier features: higher tiers often unlock connector limits, admin controls and governance features that matter more for AU compliance requirements than raw usage volume
Committed-spend or annual pricing: worth checking against a business's actual expected usage pattern, not just the per-seat sticker price
Why we're not quoting exact figures here
Both OpenAI's and Anthropic's pricing pages change more often than most vendor comparisons account for, and a specific dollar figure printed today is a reasonable bet to be stale within a few months. Any pricing comparison worth acting on should be checked directly against the current live pricing pages at the time of the actual purchasing decision, not taken from a blog post, this one included.
A worked illustration
A 10-person AU consulting team weighing a straightforward per-seat licensing model against a smaller number of seats plus dedicated Cowork automation for its highest-volume admin workflow found the mixed model came out meaningfully cheaper in their specific case, an illustrative estimate in the vicinity of $3,000-4,500 a year in total licensing costs against their actual usage pattern, versus a higher figure for licensing every staff member at the top individual tier regardless of how lightly some of them used it. The right answer depends entirely on your team's actual usage spread, not a generic rule.
The real comparison question
Rather than asking which vendor's headline seat price is lower, the more useful AU SMB question is: what does our actual usage pattern look like, a few heavy automated workflows or many light individual users, and which pricing structure fits that shape. That answer is specific to each business and worth working out with real usage data before signing an annual contract either way.
Watch for tier-feature gaps, not just price
A cheaper tier that lacks the admin controls, connector limits or governance features your business actually needs isn't really cheaper once you account for the workarounds or the compliance risk of a gap. Compare tiers on capability fit first, price second, particularly for AU businesses with specific data-handling or audit requirements.
Renewal timing matters as much as the initial price
Annual AI licensing contracts locked in now may look different against pricing twelve months from now, given how frequently these tiers and prices have shifted over the past year. Where possible, negotiate a shorter initial term or a clear renegotiation point rather than locking in a long commitment against a fast-moving pricing landscape.
A quick audit you can run today
Pull your team's actual AI usage logs, if your current platform provides them, and check how evenly usage is spread. A handful of heavy users and a long tail of light ones is a strong signal that a mixed seat-plus-usage model is worth modelling against a pure per-seat one.
What this isn't
This isn't a pricing guide with numbers you should quote back to a vendor, always verify current pricing directly against Anthropic's and OpenAI's live pricing pages before making a purchasing decision, since prices at time of writing may not reflect what's current when you're actually comparing plans.
Getting started
Map your team's actual usage pattern: heavy automated workflows versus light individual chat use
Check current live pricing pages directly rather than relying on any third-party summary, including this one
Compare tier features against your specific compliance and governance needs, not just the headline price
Model total cost against your actual usage pattern, not a generic per-seat assumption
If you're comparing AI licensing costs for your team and want help modelling the actual usage pattern against current pricing, get in touch: https://www.automataai.com.au/contact
