Let's compare notes.

Tax professionals: challenge the workflow

The best product feedback often starts with a specific preparation task. Where would you expect to find a fact? What should happen after a source correction? Which workpaper helps you review the return, and which question interrupts the work without adding value?

PrepReturns began inside a tax practice, but one practice’s experience is not every practitioner’s experience. We are interested in different review habits, awkward supported cases, and places where guidance or output needs to be clearer.

The S-Corp Engine and its capability boundaries give the conversation a concrete starting point. The public examples are synthetic. This website is not a place to upload client documents or taxpayer information.

Alpha 14 is the generation for human testing and broader operating-business coverage. A public hosted showroom is future work; there is no live tax-engine demo or testing account to create on this site. Contact us to discuss the project and the questions you would want an evaluation to answer.

Engineers: inspect the system questions

There are substantial problems at the boundaries between evidence, professional decisions, deterministic calculations, and saved output. They include source authority, explicit unknowns, dependency tracking, historical consistency, owner-specific packages, and visible PDF rendering.

The technology page explains the architecture, and the development story shows how these concerns emerged through successive Alpha generations. Conversations about data models, tests, interfaces, performance, and maintainability are welcome.

We are also interested in what useful external integrations should preserve. Moving a number is easy to describe. Preserving its period, owner, authority, and preparation context is the more interesting design problem.

The engine is private today. Open source is the intended direction, so discussion about contribution models, independent review, and sustainable maintenance is useful before a public release exists.

Software companies: a substantive product conversation

If you build professional tax software and something here overlaps with what you are thinking about—or challenges it—we would be glad to talk about architecture, product design, integrations, research, and collaboration.

PrepReturns is grounded in a practice’s workflow and built as a bounded preparation system. It offers a concrete case study in source review, deterministic tax logic, official output, and AI-assisted engineering under human direction.

We do not assume that every organization should make the same product choices. Established platforms have breadth, history, and obligations that this project does not share. Those differences can make comparison more useful. A discussion could focus on a narrow interface, a review technique, a workflow question, or the implications of practitioner-led development.

Researchers and educators: study the work

PrepReturns connects several areas worth examining: professional judgment, evidence provenance, domain modeling, human-computer interaction, and coordinated model-assisted development.

The public story includes the difficult work, not just the outcome. Correct calculations are one layer; saved-state consistency, visible forms, and human usability introduce different kinds of verification. The Alpha progression shows how those layers accumulated.

We welcome thoughtful questions about that process. The public materials describe a bounded project and its documented development, not an independently certified study or a measured productivity experiment.

Start with the question that interests you

You do not need a formal proposal. A specific observation, an integration idea, a challenging workflow question, or a different perspective on the architecture is a useful beginning.

The Project Brief provides a concise overview to share with a colleague. For a conversation, use the contact page and its existing Trio Tax contact path. Please describe the topic in general terms and keep taxpayer information out of the initial message.