The gross-to-net calculation is one of the largest and most opaque estimates a drug manufacturer makes. It is the gap between the list price a product is sold at and the cash the company actually keeps after rebates, chargebacks, discounts and returns are paid back out. Government programs, commercial payers and distributors all take their cut, often months after the sale, and the company has to estimate that deduction now, every period, on every product, before the money has moved.
Get the calculation wrong and the downstream price reporting is wrong. A deduction estimate that is too low overstates the net price reported to the government; one that is too high understates it. Either way the figure flows straight into the Average Manufacturer Price and Best Price the company certifies to the Centers for Medicare and Medicaid Services, which drive Medicaid rebate liabilities and 340B ceiling prices. It is not a back-office calculation; it is a regulated record that a responsible official certifies to the government.
What the agent is genuinely good at
Building a gross-to-net calculation is, before the sign-off, an exercise in pulling and reconciling data from systems that do not naturally talk to each other. It is slow, repetitive, deadline-bound work, exactly where an AI agent earns its place.
An agent can pull gross sales by product and channel from the ERP (the enterprise resource planning system where the company's transactions live), fetch the contract terms that drive each deduction, apply the rebate percentages owed to commercial payers and government programs, and calculate the chargebacks distributors will claim back. It can reconcile the prior period's estimate against the rebates and chargebacks that were actually invoiced, flag the products where the deduction rate has drifted, draft the candidate government-price figures, and assemble the variance commentary a reviewer reads before judging whether the calculation looks right.
None of that is the sign-off. All of it is the preparation for the sign-off. The distinction is the whole point, and it is the line a control plane has to enforce in software rather than trust to a prompt.
The figure that feeds the government is a human attestation
A pricing or regulatory reviewer does not approve a government-price calculation the way you approve a calendar invite. They are attesting that the figure is reasonable, supportable, and free of material error, and that attestation rolls up into a price certification a responsible official signs personally before it is submitted to CMS. The calculation is a one-way door in the same sense a batch release is: once it is certified and the reporting period closes, the figure is filed, it drives the rebates and ceiling prices that follow, and unwinding it later is a restatement to the government, not an edit.
That is why the same control that governs a GMP batch release governs a government-price sign-off. The actor who prepares the calculation must not be the actor who approves it. This is segregation of duties, the oldest control in pharma quality systems, and it predates agents by decades. The preparer assembles the case; an independent reviewer signs the figure that becomes the company's certified price.
When the preparer is a model, the rule does not relax. It gets sharper. A model has no personal exposure to the price certification and no instinct to hesitate over a deduction rate that looks too convenient. It will produce a flawed figure at full speed and write a tidy explanation for it.
Where most implementations get it wrong
The common failure is to put the boundary in the prompt. You are a gross-to-net assistant. Calculate the deduction and present it to the reviewer for sign-off. Do not finalize the government-price figure yourself. This reads like a control. It is an instruction, and instructions are negotiable.
A prompt has no record of who decided the agent could finalize a reportable price, no version history when someone edits it, and no way to prove eighteen months later what the agent was permitted to do on the day a reporting period closed. It offers no structural barrier to the one thing that matters most, an agent that, through a re-prompt or a quietly granted tool, ends up able to certify the figure itself and close the loop with nobody in it.
The boundary has to live somewhere the agent cannot edit. That is the entire argument for a segregation-of-duties control that holds at runtime, not on paper.
What the boundary looks like in practice
In MakerChecker, the gross-to-net agent is a named principal, a distinct identity, not an anonymous process, that holds exactly one role. That role is granted precisely the doors it needs and nothing more: read gross sales from the ERP, read the contract and pricing tables, read the prior-period actuals, write a draft deduction calculation and draft government-price figures. Those grants are deny-by-default and versioned, so you can reconstruct what the agent could do on the date any reporting period closed, and every change carries the name of whoever approved it.
Certifying the figure is a separate, gated step. The run reaches the sign-off gate and stops, because the agent does not hold the authority to proceed, structurally, the actor that built the calculation provably cannot be the one that approves it. A named pricing or regulatory reviewer reviews the assembled case and applies a signature that carries the meaning a price attestation has to carry.
| Step | Actor | Authority |
|---|---|---|
| Pull gross sales, contracts, prior actuals | Agent | Read-only grants |
| Calculate deductions, draft government-price figures | Agent | Write draft only |
| Approve and certify the reportable price | Pricing reviewer | Gated human signature |
This is the same pattern an approval gate applies to any one-way door. The gate parks the run, demands a named signature, bars the preparer from signing their own work, and records the signer's reason verbatim, so "increased the Medicaid rebate deduction rate to reflect the new utilization trend in Q1" becomes a defensible decision instead of a green tick.
For a figure that drives this much downstream liability, one signature may not be the bar. The gate supports n-of-m quorums: require the pricing reviewer plus a second reviewer from government price reporting before the figure is certified, with the requester excluded so a maker cannot pad the count.
The evidence an auditor can actually check
A gross-to-net figure is only as good as your ability to defend it later. When a CMS pricing audit tests the control, or when discovery arrives, "the agent calculated it and the reviewer approved it" is not an answer unless you can prove both halves and prove the record was untouched. The bar is data integrity: records that are attributable, legible, contemporaneous, original and accurate, the ALCOA+ principles the agency applies to every regulated record.
Every step in the run above lands in an append-only, hash-chained, cryptographically signed ledger, the gross sales the agent read, the contract terms it applied, the draft it produced, the reviewer who was excluded, the reviewer who signed, when, and why. Change one record and the chain visibly breaks, which is exactly the tamper-evidence ALCOA+ demands. The export verifies offline, against a published spec, without access to your systems, what an auditor wants, and what a screenshot of a chat transcript can never provide.
The signature side of that record matters as much as the data side. Pharma companies already run their quality and pricing records under 21 CFR Part 11, and the same discipline applies cleanly here: a signature that manifests its meaning, bound to the record it approves, in a tamper-evident audit trail. We go deeper on that in Part 11 for AI agents.
What to take from this
The EU AI Act's high-risk obligations were pushed to late 2027, and AI-specific guidance keeps shifting. None of it moved 21 CFR Part 11, the government price reporting rules, or the fact that gross-to-net is a regulated record with a named official's certification at the end of it. The rule that the preparer cannot be the approver, and that a person signs the figure that feeds Medicaid rebate and 340B reporting, is date-proof, and the audit that tests it will not wait for new AI guidance.
The right deployment is not "an agent that certifies the price." It is an agent that does the pulling, reconciling, and drafting at machine speed, handing a clean case to the reviewer who still, and only, signs at the one-way door. The close gets faster. Accountability stays exactly where the price reporting rules put it.
See how it works, or book a demo to watch an agent get blocked from approving its own work.