A true up is the clause that lets Anthropic bill you for seats you added during the term, and on the surface it sounds fair. You grew, you used more, you pay for what you used. The problem is that the mechanics inside that clause decide whether the extra seats cost you the rate you negotiated or a much higher one, and whether a temporary spike in headcount locks you into a permanently larger commitment. If you are at the point of signing or renewing a Claude Enterprise agreement, the true up language is one of the few places where a single sentence can move your bill by six figures. This is how the traps work and how we negotiate them out before you sign.
When you license Claude Enterprise you commit to a seat count and a price. A true up is the contractual method for reconciling the difference when your real seat usage rises above that committed count during the term. At an agreed point, often quarterly or annually, Anthropic counts the seats you actually provisioned, compares that to your commitment, and bills the gap. In principle this protects both sides. You are not forced to predict your exact headcount on signing day, and Anthropic is paid for the access you consumed.
The reason true ups deserve scrutiny is that the clause is rarely neutral. The default language tends to favor the vendor in three specific ways: the rate applied to the extra seats, the direction the adjustment is allowed to move, and the way a mid term peak resets your floor for the rest of the term. Each of these is negotiable, and each is routinely left at the unfavorable default because buyers treat the true up as boilerplate rather than as a pricing term.
The first and most common trap is a true up that prices new seats above the rate you negotiated for your committed seats. You spend weeks driving the per seat number down, you sign, and then every seat you add during the term comes in at list or at a smaller discount. Because growth seats are added quietly, one team at a time, the higher rate often goes unnoticed until the reconciliation invoice arrives.
The fix is a single clause that fixes the rate for all true up seats at the same discounted rate as your committed seats, for the full length of the term. This is the most valuable sentence in the whole section, because a right sized rollout is designed to grow, which means most of your seat additions will come through the true up. If those additions are not protected, the discount you negotiated on day one only applies to the smallest version of your deployment.
The second trap is asymmetry. A standard true up reconciles usage above your commitment but offers nothing when your usage falls below it. If adoption stalls, a team is reorganized, or a project winds down, you keep paying for the committed seats while losing the ability to recover anything from the seats you are not using. The clause is a ratchet: it only ever turns in the vendor's favor.
You will rarely win a fully symmetric true up that refunds you for under usage, because Anthropic relies on the committed floor for revenue predictability. But you can win the next best things. You can negotiate the right to reallocate seats across teams without penalty, so a department that shrinks frees seats for one that grows. You can negotiate a true down option at renewal that resets your floor to real usage rather than to your peak. And you can cap how far the true up can swing in a single period, so one unusual quarter does not produce a shock invoice.
The third trap is the most expensive and the least understood. In some agreements, the highest seat count you reach during the term becomes your new minimum for the remainder of the term, and sometimes the baseline that carries into the renewal. So a short lived spike, perhaps a seasonal project or a one time onboarding of contractors, permanently raises the floor you pay against, even after those people are gone.
This is how a company that genuinely needs four hundred seats most of the year ends up paying for six hundred because of a single busy quarter. The defense is to define the committed floor explicitly as your steady state usage, not your peak, and to ensure that any spike is billed only for the period it occurred. Temporary capacity should be priced as temporary capacity, never as a permanent reset of your baseline.
The fourth trap hides in definitions. What counts as a seat for true up purposes? A provisioned account that has never logged in? A user who was active for one day? The measurement date matters too: if the count is taken on a single snapshot day rather than averaged across the period, a one day spike can be billed as if it lasted the whole quarter. Loose definitions hand the vendor the right to interpret the count in the way that produces the largest invoice.
Tighten the language. Define a billable seat as an active, provisioned user, and specify the measurement method, whether that is an average across the period or a count of seats active for a minimum number of days. Ambiguity in a true up is never accidental, and it always resolves against the buyer when the invoice is calculated.
Individually each trap costs money. Together they compound into something much worse. Picture a deal where the true up rate sits above your committed rate, the clause only moves up, a peak resets your floor, and the counting is generous to the vendor. A single busy quarter now triggers extra seats billed at a premium rate, locks that higher count in as your minimum for the rest of the term, and gives you no mechanism to recover when usage falls back. One unfavorable section has quietly rewritten the economics of the entire agreement.
This is why we treat the true up as a pricing negotiation in its own right rather than as legal cleanup at the end. The headline discount gets all the attention in the room, but the true up decides what that discount is actually worth once your deployment starts to move.
When we run a Claude Enterprise negotiation for a buyer, the true up section comes back with a short, specific set of protections. The rate for true up seats is fixed at the committed rate for the full term. The committed floor is defined as steady state usage, never the peak. Temporary capacity is billed only for the period it is used. Seats can be reallocated across teams without penalty. The counting method and the billable seat definition are spelled out precisely. And a renewal true down resets the floor to real usage so a single good year does not inflate the next contract.
None of these asks is exotic. They are the ordinary protections any disciplined buyer would want, and Anthropic grants them far more readily before signature than after. The leverage you have on an unsigned deal is the leverage that makes these clauses cheap to win, which is exactly why the true up has to be addressed during the negotiation rather than discovered at the first reconciliation.
A true up clause is not boilerplate. It is the term that decides whether your negotiated rate survives contact with real growth, whether a temporary spike becomes a permanent cost, and whether the contract can ever shrink when your usage does. Read it as a pricing term, negotiate the rate, the direction, the floor, and the counting, and write the protections in before you sign. The work takes an afternoon and it routinely protects more value than the headline discount everyone fought over. If you have a Claude Enterprise agreement on the table, send us the true up language and we will tell you exactly where it bends against you.
Send us your Claude Enterprise contract and we will price the true up risk before it becomes an invoice.
Get a QuoteWeekly intelligence on Anthropic pricing moves and the buyer side counters that work.