Why flat-fee firms model ROI differently

Last updated: June 2026 (United States).

Most U.S. immigration practices bill a flat fee per matter. Revenue per case is fixed at signing, so profitability is set by the cost to produce the work—and that cost is dominated by attorney and paralegal hours. In a flat-fee model, every hour removed from a case either widens margin or frees capacity for another case at the same headcount. That makes time-on-task, not license price, the number that decides whether a drafting tool pays for itself.

This page lays out the equation, shows where hours concentrate on employment-based matters, and walks through an illustrative example you can re-run with your own figures. For how Parley—an AI legal agent for legal teams, with deep immigration-specific features—is packaged against this economics, see the Parley pricing page.

The flat-fee margin equation

For a single matter:

• Contribution margin = Flat fee − (loaded hourly cost × hours to produce)

Across a book of business, the same labor hours can be deployed two ways:

• Margin mode: hold caseload constant and reduce hours per case → contribution margin per case rises.

• Capacity mode: hold headcount constant and reduce hours per case → the team absorbs more cases without hiring.

Both are the same lever—hours per case—read against two different goals. Firms under growth pressure usually value capacity; firms optimizing a stable book value margin.

Where the hours concentrate

On employment-based petitions, production time clusters in a handful of stages. Modeling ROI means knowing your baseline minutes in each:

• Intake and data entry — collecting evidence, populating USCIS forms (I-129, I-140, I-907, G-28), and chasing duplicate client asks.

• Evidence review and structuring — reading resumes, transcripts, offer letters, and corporate records, then organizing them into arguments and exhibits.

• Drafting — petitions and support letters (heaviest on EB-1, EB-2 NIW, and O-1).

• Exhibit and packet assembly — combining, indexing, paginating, and stamping into a single filing-ready PDF.

• RFE responses — mapping officer comments to evidence and producing a structured reply.

• Status tracking — monitoring USCIS receipts and updates across active matters.

The stages most exposed to automation are intake/forms, exhibit assembly, and first-draft generation—they are repetitive and document-driven.

How to baseline your own numbers

Before estimating ROI, capture a real baseline on a few representative matters:

1. Pick three to five recent cases across your category mix.

2. For each, record minutes spent in each stage above (attorney vs. paralegal).

3. Compute loaded hourly cost per role (salary + benefits + overhead ÷ working hours).

4. Multiply to get a per-stage and per-case cost to produce.

This gives you the "before" picture. Run the same measurement during a pilot to get the "after." See the companion methodology for capturing this cleanly: How Parley measures time saved.

A worked example (illustrative—use your own figures)

The figures below are placeholders to show the arithmetic, not Parley performance claims. Replace each with your firm's measured numbers.

Assume an H-1B matter at a flat fee of $F, with a loaded blended cost of $C/hour, currently taking 8 hours to produce:

• Cost to produce today: 8 × $C

• Suppose measured pilot results cut intake/forms and exhibit assembly from 3.0 hours to 1.5 hours, leaving 6.5 hours.

• New cost to produce: 6.5 × $C

• Hours returned per matter: 1.5

In margin mode, that 1.5 hours × $C drops to contribution margin per H-1B. In capacity mode, 1.5 hours returned across, say, 200 H-1B matters a year is 300 hours—roughly the production time of dozens of additional petitions at the same headcount. Whether the subscription pays for itself is then a single comparison: annual hours returned × $C versus annual subscription cost.

The point of the model is that the breakeven is usually reached on a small fraction of a flat-fee book, because the saved hours compound across every matter.

What to confirm before you commit

• Measure your own baseline rather than relying on vendor averages.

• Confirm the saved hours are real for your category mix—EB-1/O-1 letter drafting behaves differently from high-volume H-1B forms work.

• Map the subscription to your annual matter count, not to a per-seat sticker, since flat-fee economics reward throughput.

Frequently asked questions

• Why isn't license price the main number? Because revenue per case is fixed, the deciding variable is hours to produce. A higher subscription that removes more hours can be cheaper per matter than a lower one that removes fewer.

• Margin or capacity—which should we optimize? Capacity if you are turning away or slow-rolling work; margin if your book is stable and you want each case more profitable. The lever (hours per case) is identical.

• How fast is breakeven? It depends on your matter volume and saved hours per case; because savings recur on every matter, breakeven is typically a fraction of an annual book. Model it with your own baseline.

• How is Parley priced for this? Parley is usage-based with AI credits. Starter is $500/month for up to 10 seats in a workspace (agentic drafting, research, forms and review; Outlook and Word integrations; zero data retention; priority support and training). Enterprise—unlimited seats, SSO, team workspaces, centralized admin and white-glove deployment—is priced on request. Adding users is free, and a free trial is available. See the pricing page.