How AI Patent Drafting Software Is Priced: Seats, Matters, and Pay-Per-Workflow

AI patent drafting software is sold through a handful of billing structures: access priced by user seats, subscriptions that include a pool of usage credits, packages priced around a patent matter, and pay-per-workflow pricing where each production step is charged when it is used. Some products combine two of these. Compare the billable unit and what it includes, not the plan name.

The useful comparison is not “Which plan costs least?” It is “What will this team pay for the production work one matter needs, and who can take part in reviewing it?”

Start with the billable unit, not the plan label

These are four billing units buyers should identify when evaluating AI patent drafting software pricing. They describe structures a buyer may encounter, not a claim that every vendor uses all four.

Billing unit	What it means	Main question for the buyer
Seat	Access is allocated by named or permitted users	Which participants need paid access?
Subscription with credits	A recurring plan includes a defined pool of usage	What does a complete workflow consume?
Matter	A defined package is priced around a patent matter or deliverable	What exactly does the matter include?
Pay-per-workflow	Individual production steps are charged when used	Which steps will this matter actually need?

A workflow can be charged per use without the whole matter being sold as one bundle. For example, a search, a full draft, and a review can each be separate billable steps.

That distinction matters because the same patent team can produce very different costs under each structure. A team with occasional matters may use only a few workflows in one period. A high-volume team may need to compare the cost of repeated production with the cost of giving many people access.

Seats: separate the user count from the output

When a contract prices access by named users, headcount is the billable unit. The first question is not simply how many people work in the IP function. It is which people need to enter the system, run workflows, review outputs, or approve decisions.

Ask:

• Do reviewers, inventors, patent engineers, administrators, and outside counsel each need paid access?
• Is reviewer or view-only access distinct from authoring access?
• Does the plan also cap output, so that a “user” plan is not a pure seat model?
• If the number of included users is not published, is it a quote-stage question?

A seat model can be easy to budget when the user group is stable. It can be less aligned with production volume when a small team handles many matters, or when many people need occasional access to review a particular invention.

The important comparison is therefore between access and output. A team should record both the number of people who need to participate and the number of matters it expects to run.

Subscriptions with credits: map credits to the whole workflow

Credit-based plans meter individual actions. As a result, the same plan can cover very different amounts of work depending on what a matter needs.

A credit allowance is not a matter allowance. Count the search, claims, specification, figures, revisions, and review actions your team expects to run.

Before choosing a credit plan, ask:

1. What counts as one credit-consuming action?
2. Are revisions and regenerations charged again?
3. Are drawings, searches, and review workflows metered separately?
4. Do unused credits expire, roll over, or stay in a purchased balance?
5. Can credits be pooled across matters and users?
6. Which capabilities are available only in a custom plan?

The practical challenge is translating a plan's abstract usage allowance into the work of a real matter. A full application may involve more than one claim-development pass, a specification revision, figure generation, and review. A prior-art search may be a separate workflow rather than part of the drafting allowance.

That is why buyers should model a complete matter rather than ask only how many credits a plan contains.

Matter and pay-per-workflow pricing: define the boundary

A workflow price answers, “What does this step cost?” A matter price should answer, “Which steps, revisions, outputs, and jurisdictional versions does this matter include?”

Pay-per-workflow pricing can suit teams whose volume varies. The team pays for the workflow it runs rather than buying unused workflow capacity in advance. Buyers should still confirm how revisions, figures, searches, and multi-office versions are charged.

Matter-based pricing may be easier to compare when the deliverable has a clearly defined boundary. The boundary needs to be specific. For example, a buyer should determine whether the matter includes:

• invention intake or interview work
• prior-art search
• claim development
• specification drafting
• abstract and summary
• figures
• revisions
• professional review
• exports
• versions for different jurisdictions

Without that list, “per matter” can conceal the same consumption questions found in a credit plan.

The Matter Economics Worksheet

A useful way to compare models is to price the same basket of work under each one.

Formula

Expected software spend = fixed access fees + plan fees + workflow charges + expected top-ups

Then score each option against five questions:

1. Unit alignment: Is spend driven by users, consumption, matters, or individual workflows?
2. Inclusion boundary: Are search, claims, specification, figures, revisions, exports, and review included?
3. Volume behavior: What changes if the quarterly matter count doubles?
4. Collaboration cost: Can the required reviewers take part without extra access charges?
5. Governance fit: Does the workflow give clear points for patent-professional review and approval?

Hypothetical worked example

Assume a team runs 12 matters in a quarter:

• 6 matters need an invention interview
• 12 matters need a full draft
• 8 matters need a prior-art search

Under pay-per-workflow pricing, the team pays for 26 workflow runs.

Under a seat model, it pays for every user who needs access, whether or not those users work on these matters during the quarter.

Under a credit plan, it pays for the plan and may need top-ups if the 26 workflow runs consume more than the allowance.

This example does not establish which model is cheaper. It shows the correct comparison method: price the same basket of work under each model. Do not compare a per-user plan with a per-draft price without accounting for what each includes.

Review is a pricing question and a governance question

Software pricing should not be separated from the review workflow. A low-friction way to produce text is not useful if the team cannot identify who reviews it, what they review, or when their decision is required.

Questel's 2026 IP Industry Outlook Research surveyed more than 500 IP professionals between January and March 2026. It reported that 83% use AI to save time and costs, while 88% spend up to half their time reviewing work by trainees, AI agents, or external suppliers. The research supports the relevance of review workflow. It does not establish a return on investment for any particular software model. Read Questel's 2026 IP Industry Outlook Research

For US-facing work, the USPTO's April 11, 2024 guidance explains that existing rules and policies apply when AI tools are used in practice before the Office and discusses risks that require human governance and verification. AI-assisted drafting alone should not be treated as a general requirement to disclose AI use. The guidance is a reason to ask where professional review occurs, not to treat review as an optional add-on. Read the USPTO guidance

How Esgenix connects pricing to production work

Esgenix describes its commercial model as Outcome-Based Pricing: searches, analyses, and drafts are priced when they are delivered, across tiers. Esgenix Self-Serve, available from October 1, 2026, lets users create an account, choose the workflow they need, pay online per workflow, and start immediately. See the Esgenix workflow platform or create an Esgenix account.

The structural principle is:

Software cost should follow patent work produced, not seats provisioned.

That principle connects pricing to Esgenix's operating model. Esgenix is a decision-gated agentic patent workflow platform. An agent is an AI component assigned a defined role and able to carry out several steps toward an objective. In Esgenix, specialist agents perform production work between five named professional decision gates:

1. Intake: The professional confirms the invention understanding, resolves technical gaps, and chooses jurisdictions.
2. Strategy: The professional sets claim direction, breadth, and fallback positions.
3. Claims: The professional approves, modifies, or rejects the claim set.
4. Specification & Figures: The professional reviews the specification, provenance, figures, and consistency.
5. Sign-off: The professional completes review and approves the output.

The gates are described in more detail on About Esgenix. They give a buyer a concrete way to ask what the software is charging for. Is the team paying mainly for access licenses, or for production work completed between substantive professional decisions?

Search is another boundary to clarify. Esgenix describes Prior Art Search as a workflow whose findings can inform claim strategy and drafting in the same matter workflow. Review whether prior-art search is included in the priced workflow or charged separately. See Esgenix Prior Art Search.

Questions to ask before choosing a pricing model

Use these questions during a demo, quote review, or self-serve evaluation:

1. What is the exact billable unit?
2. What does a complete matter or workflow include?
3. How are revisions, figures, search, and jurisdictional variants charged?
4. What happens when volume exceeds the allowance?
5. Where does patent-professional review occur before work progresses?

The answers should be mapped to one representative matter and one expected quarter of work. That makes it easier to see whether the commercial model follows the team's actual production pattern.

For smaller firms and solo practitioners, choose a workflow and start today. For in-house and corporate IP teams, bring one active invention disclosure to a demo and map the search, strategy, claims, specification, figures, and review work to a work-based commercial model.