Skip to content

Documentation Builder

The strongest documentation is created while the work is happening. The builder captures it as a by-product of how your teams already operate.

What's included

  • Project narratives drafted during the year from contemporaneous records and mapped to the four-part test
  • Time and wage allocation records tied to named individuals
  • Version-controlled evidence library with a change log
  • Quarterly documentation health report

Who this service is for

Engineering, software and product organisations with continuous development cycles.

Discuss this service (opens in a new tab)

How it works

  1. Step 1: Connect your systems

    We connect to the systems your engineers already use — issue trackers, version control, design review tools, lab notebooks, time systems — and translate that activity into structured project records aligned to the four-part test.

  2. Step 2: Draft the narratives

    Drafted narratives describe the technical uncertainty faced, the alternatives evaluated, the experiments run and the outcomes observed. Your technical leads review, correct and approve them, so the record reflects their own account of the work. AI-drafted narratives are only as reliable as the source records and the reviewer's approval.

  3. Step 3: Assemble the claim

    Because records accumulate through the year, the annual claim becomes an assembly exercise rather than an archaeology project, and staff turnover becomes much less of an evidentiary risk.

Common questions

Read all questions
Does my company need a laboratory to qualify?

No. Qualification turns on a four-part test, not on job titles or facilities. Software teams, manufacturers, engineering firms, food and beverage producers, agricultural technology businesses and many others can qualify when they resolve technical uncertainty through a process of experimentation and the activities are not excluded. Software developed primarily for internal use must also meet the high threshold of innovation test.

How does creditARMOR use AI in a claim?

In the mock audit, AI (Google Gemini) reads the documents you upload, matches them to the examiner-style questions, rates the evidence, drafts first-pass responses and flags inconsistencies. In Swanson Reed's separate services, automation also parses ledgers and payroll data and groups activity into candidate projects. If requested, a person at Swanson Reed can review your mock audit results and re-perform the mock audit for assurance. The taxpayer remains responsible for the positions on its return.

What are the four parts of the qualification test?

The activity must have a permitted purpose: the costs must be eligible research or experimental expenditures (IRC sections 174 and 174A), and the research must relate to a new or improved business component (a product, process, computer software, technique, formula or invention) and aim to improve its function, performance, reliability or quality. It must be technological in nature, meaning the process of experimentation fundamentally relies on principles of the physical or biological sciences, engineering or computer science. It must be intended to discover information that would eliminate uncertainty about the capability or method for developing or improving the business component, or about its appropriate design. And substantially all (at least 80%) of the activities for the business component must be elements of a process of experimentation, such as modelling, simulation or systematic trial and error, used to evaluate one or more alternatives.

Run a free R&D tax credit mock audit.

You can also book a free thirty-minute scoping call with a specialist. Upload your files and see how your evidence would hold up to the questions an IRS examiner is likely to ask. A higher score means a more audit-ready claim.

creditARMOR is a free service. Passing this mock audit does not guarantee that your claim will pass an IRS examination. Book a call or email hello@creditarmor.ai for a login.