When companies look at the R&D tax credit, they often start with expenses: engineer salaries, prototype materials, cloud computing costs, or outside contractors. But this starts the analysis in the wrong place.

The R&D credit begins with activities, not expenses. First, you identify the work that qualifies under Section 41. Then you determine which costs connect to that work and meet the separate qualified research expense (QRE) rules. The IRS itself treats these as distinct questions when examining research credit claims.

Think of it as activities vs. expenses: “What did your team do?” vs. “What did you spend to do it?” This guide stays on the first question.

If this introduction has left you with more questions than answers, first read What is R&D in Business? Conversely, if you've already identified your qualifying work, continue to R&D Tax Credit Qualified Expenses for the cost side of the calculation.

Read on for answers to the following questions:

  • What are R&D tax credit qualified activities?
  • How does the IRS define a qualifying activity?
  • What are examples of qualifying R&D activities by industry?
  • What are some activities that don’t qualify?
  • How do you document qualifying activities for R&D tax credit?
  • How do you turn qualified activities into credit?

What Are Qualified Research Activities for the R&D Tax Credit?

Qualified research activities are activities that satisfy Section 41's requirements for qualified research and avoid its specific exclusions. Broadly, the work must relate to developing or improving a business component, use technological principles, and use experimentation to resolve uncertainty about function, performance, reliability, quality, capability, method, or design.

The most useful place to begin is the business component.

Section 41 defines a business component as a product, process, computer software, technique, formula, or invention that a company plans to sell, lease, license, or use in its own business.

That means a company does not simply say, “We are innovative, so our engineering department qualifies.”

Instead, you might analyze:

  • A software company's new recommendation engine.
  • A manufacturer's redesigned production process.
  • A food company's reformulated shelf-stable product.
  • An engineering firm's new structural solution for a project.

Then you look at what the team actually did to develop or improve that component.

The IRS also reminds us of two important points:

  1. Job titles do not decide the answer. A founder can perform qualifying research. So can a machinist, developer, formulation scientist, or engineer. Conversely, putting “R&D” in someone's title does not make all of their work eligible.
  • Success is not required either. A failed prototype or abandoned technical approach can still involve qualifying research if the underlying activities satisfy the Section 41 rules. The IRS focuses on the uncertainty identified, the alternatives considered, and the process used to evaluate them.

How Does the IRS Four-Part Test Apply to R&D Activities?

The IRS four-part test asks whether the research connects to eligible research expenditures, seeks technological information for a new or improved business component, serves a permitted purpose, and includes a process of experimentation. You apply the test to each business component, and the activity must also avoid Section 41's statutory exclusions.

For a more thorough explanation of each prong, see The R&D Tax Credit Four-Part Test. The important point here is that the test applies to activities, not simply projects or departments. Here, we’ll just give brief examples to illustrate what each means:

Qualification
Definition
Example

Permitted purpose

The research aims to develop or improve a business component's function, performance, reliability, or quality

An engineering team redesigns a production process to improve its reliability

Technical uncertainty

You are unsure how and if the desired result can be achieved

A software team knows the performance target it needs to hit but doesn't know which architecture can achieve it

Process of experimentation

Your team identifies the uncertainty, considers one or more alternatives, and evaluates those alternatives

Engineers build and test alternative prototypes to determine which design meets the required specifications

Technological in nature

The experimentation must fundamentally rely on principles of the physical or biological sciences, engineering, or computer science

A manufacturer applies established engineering principles while testing different materials for a new component

 

Note here that the One Big Beautiful Bill Act (OBBBA) also changed the surrounding research-expense law. The new Section 174A allows deductions for domestic research or experimental expenditures paid or incurred in tax years beginning after 2024. Section 41's current qualified-research definition now references domestic research or experimental expenditures under Section 174A.

What Are Examples of Qualifying R&D Activities?

Qualifying R&D activities can occur in software, manufacturing, engineering, life sciences, consumer products, agriculture, and other industries. The industry itself does not create the credit. The activity needs to satisfy the Section 41 requirements, such as resolving technological uncertainty through experimentation, and it cannot fall within one of the statutory exclusions.

Here are some common activities worth evaluating. Several of these industries have their own examples page that you can consult for further guidance:

Industry
Activities that
may qualify
Activities to
separate out

Software & tech

Designing algorithms, testing architectures, improving scalability, resolving integration or performance uncertainty

Routine maintenance, ordinary bug fixes with a known solution, certain internal-use software

Manufacturing

Developing production techniques, testing materials, prototyping tooling, improving yield through technical experimentation

Routine inspections, ordinary quality control, copying an existing process

Engineering & A&E

Evaluating structural alternatives, developing technical systems, modeling solutions to engineering constraints

Purely aesthetic work, standard drafting, adapting an existing design without technical experimentation

Life sciences & medical devices

Formulation work, prototype development, testing device performance, evaluating technical alternatives

Routine post-production testing or excluded funded work

Food & CPG

Developing formulas, testing ingredients for stability or performance, improving manufacturing methods

Taste-only changes, market testing, seasonal packaging changes

Breweries & beverage companies

Testing fermentation methods, shelf stability, production equipment, or technically uncertain formulas

Consumer preference surveys or routine batch quality checks

Agriculture & agtech

Developing equipment, irrigation technology, biological processes, or technically uncertain production methods

Ordinary production activities without experimentation

 

These examples are starting points, not automatic qualifications. For more examples of R&D tax credits by industry, peruse this overview of R&D Tax Credit Examples for Top Industries.

Or simply select your industry from the drop-down menu below.

What R&D Activities Don't Qualify for the Credit?

Section 41 specifically excludes several activities even when the work feels like R&D. IRC Section 41(d)(4) excludes research after commercial production, customer-specific adaptation, duplication, certain surveys and management studies, routine data collection and quality control, certain internal-use software, foreign research, social-science research, and research funded by another party.

This is where companies often need to draw much sharper lines:

Potentially Qualifies
Doesn’t Qualify

Developing a new product or process: Designing and testing a new product, software feature, formula, or production process when technical uncertainty exists

Research after commercial production:

Routine work performed after the business component has reached commercial production

Improving an existing business component:

Experimenting to improve a product or process's function, performance, reliability, or quality

Customer-specific adaptation:

Adapting an existing business component to meet a particular customer's requirement or need

Testing alternative designs:

Modeling, simulating, prototyping, or systematically testing alternatives to resolve technical uncertainty

Duplication:

Reproducing an existing business component from plans, specifications, physical examination, or publicly available information

Collecting data as part of experimentation:

Gathering and analyzing data when it forms part of the process used to evaluate technical alternatives

Routine data collection or quality control:

Collecting data or testing products against established standards without resolving technical uncertainty

Solving technical problems:

Using engineering, computer science, or physical or biological science to evaluate potential solutions to an uncertain technical problem

Surveys and certain studies:

Efficiency surveys, management studies, market research, advertising, promotions, and similar activities

Conducting qualifying research in the U.S.:

Performing otherwise qualified research within the United States, Puerto Rico, or a U.S. possession

Foreign research:

Research conducted outside the United States, Puerto Rico, or a U.S. possession

Research grounded in eligible sciences:

Experimentation fundamentally relying on engineering, computer science, or physical or biological sciences

Social sciences, arts, or humanities research:

Research in these fields falls within Section 41's statutory exclusions

Contract research where the taxpayer satisfies the applicable rules:

Depending on the contractual arrangement, research performed for another party can warrant a Section 41 analysis

Funded research:

Research can be excluded when another party funds it under the Section 41 rules. The contract and facts determine the result

 The most common trip ups

1. Research after commercial production

Section 41 excludes research conducted after the beginning of commercial production of the business component. That does not mean a product can never generate qualifying research again once you sell version one. A later project may involve a genuinely new or improved business component or qualifying improvement. The point is that routine work after commercial production does not become eligible merely because engineers continue touching the product.

Companies should therefore mark where development work transitions into ordinary production, maintenance, or support.

2. Funded research

The funded-research exclusion deserves special attention for contractors, engineering firms, manufacturers, and businesses doing development work for customers. Section 41 excludes research to the extent another person or government entity funds it. However, the actual analysis depends heavily on the contract.

IRS guidance focuses on issues such as whether payment depends on successful research and whether the party performing the research retains substantial rights in the results. So “the customer paid us” does not answer the funded-research question by itself. You need to read the agreement.

That makes statements of work, IP provisions, payment terms, and risk allocation important records for an R&D credit review.

How Should You Document Qualified R&D Activities?

To claim qualifying activities, document them at the business-component level with contemporaneous records: project notes, design iterations, test results, and time tracking that tie the technical work to the four-part test. Good documentation connects each qualified activity to its qualified research expenses and is your best defense if the IRS examines the claim.

The key word is connection. You must be able to demonstrate where your activities connect with qualifications. Unfortunately, a $2 million engineering payroll number does not prove a $2 million pool of qualified research expenses. The IRS has specifically warned about claims that fail to establish a nexus between expenses and qualified research activities.

A better record tells the story at the business-component level. Consider this example list as a method for drawing up the necessary items for filling:

  • Business component: New inventory optimization engine
  • Technical uncertainty: Whether the proposed architecture could process required volumes within the target response time
  • Activities: Benchmarking, modeling, load testing, alternative architecture tests, iterative redesign
  • Result: Final architecture selected after two approaches failed performance targets
  • Related costs: Qualified wages and other potentially eligible expenses tied to those activities

Then your QRE analysis answers the separate expense question.

For more on that side, see R&D Tax Credit Qualified Expenses.

Form 6765 makes business-component records more important in 2026

For tax years beginning after 2025, the IRS's current Form 6765 instructions require Section G for many filers. Subject to the stated exceptions, taxpayers must report business-component information covering at least 80% of total QREs, with no more than the top 50 business components individually reported.

The IRS also tells taxpayers to use business-component names or identifiers that align with the books and records used to substantiate the activities and associated QREs.

There are exceptions. For example, the current instructions exempt certain original-return filers from Section G when total controlled-group QREs do not exceed $1.5 million and average annual gross receipts for the prior three tax years do not exceed $50 million. Qualified small businesses making the specified reduced payroll-tax-credit election also receive an exception under the current instructions.

You do not need to turn every engineering meeting into a tax memo. You do want records created in the normal course of business that let someone reconstruct what technical question the team was solving, what alternatives it considered, and how the claimed expenses relate to that work.

For the filing mechanics, continue to our IRS Form 6765 Ultimate Guide.

Turning Qualified Activities Into an R&D Tax Credit

For many SMBs, the opportunity to capture credits sits in work they already perform but do not call “R&D.” A software company may call it architecture work. A manufacturer may call it process improvement. A food business may call it formulation. The label matters far less than the underlying activity.

The takeaway: reframing the way your company thinks of qualifying activities reduces the risk of leaving money on the table.

Arvo helps companies make that analysis practical. Our R&D tax credit work starts with what your team actually did, then connects those activities to the financial records needed to calculate and support the claim. If you're ready to start sketching out a solid filing plan, start with Arvo's R&D Tax Credit Calculator. Beyond this, learn more about R&D Tax Credit Consulting Services.

[The R&D tax credit] is a really great approach to actually incentivize people to do things that are creative and innovative…It made obvious sense to work with Arvo because it meant that we could actually get our work done quicker.”

–Jason Marks, CEO of Telo Trucks