The Complete Guide to SR&ED Claim Preparation: Documentation, Automation and AI

Preparing a Scientific Research and Experimental Development (SR&ED) claim shouldn’t require weeks of chasing engineers, reconstructing project histories and searching for documentation.

Yet for many Canadian businesses and SR&ED consultants, that’s exactly what happens every year. We know it well ourselves. We’ve personally spent several weeks every Spring explaining to a consultant what our own team had built and gathering the data to prove it.

We think engineering teams should be focused on developing products, solving complex problems and experimenting with new technologies, not doing admin work. Engineers should spend less time thinking about how to document their work for a future tax credit claim and more time creating. Unfortunately, what usually happens when claim preparation begins is that engineers and engineering managers start spending time trying to gather information that is scattered across project management systems, source code repositories, spreadsheets, emails and other employees’ memories.

The result is often a time-consuming process that creates unnecessary administrative costs, incomplete documentation and uncertainty about the work and expenditures being claimed. Technically, the result doesn’t meet the CRA’s expectation that the underlying records, especially time allocation, were created contemporaneously through the year and not reconstructed at filing or audit time. This is a big problem if you are audited.

There’s a better approach: build your SR&ED claim from the work your team is already doing.

Modern AI and software integrations are making it possible to identify potentially eligible activities, organize supporting evidence, reconstruct project timelines and assist with technical documentation throughout the year.

A business gets less disruption and more control over its own claim. A consultant gets a way to take on more clients without adding headcount to every engagement.

This guide explains how SR&ED claim preparation works, what documentation you need, where traditional approaches fall short and how AI-powered automation can improve the process.

1. What is SR&ED, and why does claim preparation matter?

Canada’s Scientific Research and Experimental Development program (opens in a new tab) provides tax incentives to businesses undertaking eligible research and development activities in Canada.

The program is administered by the Canada Revenue Agency (CRA) and is intended to encourage businesses to conduct research and development that advances scientific knowledge or achieves technological advancement.

Qualifying activities can include experimental development, applied research and basic research, as well as certain supporting activities.

For example, a software company might undertake eligible experimental development while attempting to overcome technological uncertainties in a new data-processing architecture. A manufacturer might conduct experiments to develop a production process that achieves performance characteristics that existing technology cannot reliably deliver.

However, not every innovative project qualifies for SR&ED. To be eligible, work must meet the program’s requirements, including the pursuit of scientific or technological advancement through systematic investigation or experimentation. Routine engineering, ordinary software development and commercial product improvements do not automatically qualify.

The test for SR&ED eligibility is actually defined in the Canadian Income Tax Act, in subsection 248(1) (dated August 13, 2021). They break it down into the “Why” requirement and the “How” requirement. The “Why” must be to generate or discover knowledge that advances the understanding of science or technology. The “How” must be a systematic investigation or search that is carried out in the field of science or technology by means of experiment or analysis.

What goes into an SR&ED claim?

Preparing a claim generally involves five activities:

1. Identifying potentially eligible work and organizing it into SR&ED projects.
2. Establishing the technological uncertainties, advancement sought and systematic work performed.
3. Gathering supporting technical documentation and financial records.
4. Calculating eligible expenditures, including qualifying employee labour.
5. Preparing and submitting the required tax forms, including Form T661 (opens in a new tab), along with the supporting project descriptions.

The CRA recommends involving “knowledgeable technical and financial personnel in the preparation process”, which basically means that your engineers need to be involved. Businesses are responsible for the accuracy and completeness of their claims, even when an external consultant prepares them, and so it is in the best interests of the business to understand their claims.

The obvious challenge here is that the people with the technical knowledge are often the same people responsible for delivering the company’s most important development projects. Taking them away from that work to reconstruct months of technical activity is expensive and disruptive.

2. Why traditional SR&ED claim preparation is inefficient

The traditional SR&ED preparation process is often retrospective.

At the end of the financial year, a business or its SR&ED consultant begins gathering information about the projects completed during the preceding months. Engineers and technical leaders are interviewed. Project histories are reconstructed (usually by extracting data from systems of record, like Jira). Employee time allocations are estimated. Technical narratives are written, reviewed and revised.

As the creators of ActiveSRED, we’ve personally sat through the “reconstruct six months of engineering work in three weeks” process ourselves. We’ve also sat through numerous SR&ED audits and the countless hours of audit preparation work required to gather timesheets, payroll data, etc. We know the pain all too well and it’s one of the reasons we created ActiveSRED.

Unfortunately, the SR&ED process creates several recurring challenges…challenges we believe we have solved.

Information is collected too late

By the time claim preparation begins, employees may struggle to remember the details of experiments conducted months earlier. Failed approaches, intermediate results and technical decisions can be particularly difficult to reconstruct. Ironically, it’s usually the failed experiments and trial and error notes that are the best evidence that new technical or scientific knowledge was generated.

Evidence is scattered across multiple systems

Project management tools, source code repositories, design documents, test results and employee calendars may all contain relevant information. Gathering and organizing it manually can require considerable effort. We’ve learned that it can be tough to offload this work to someone outside of engineering (i.e. Finance) because they simply don’t have the engineering understanding or context to know what information is relevant.

Technical employees become a bottleneck

If the business has hired consultants, engineers may be asked to attend multiple interviews, review extensive spreadsheets and explain work that is already documented elsewhere. Their involvement is essential, but the administrative burden need not be.

Supporting evidence may be incomplete

Retrospective narratives cannot always be connected to the underlying technical records. This can make it harder to substantiate the work if the CRA selects a claim for review.

The same work is repeated every year

Businesses and consultants often repeat the process of gathering records, organizing projects, conducting interviews and preparing narratives for every claim cycle.

More often than not, the information already exists somewhere. Nobody has organized it for SR&ED purposes yet.

3. What documentation does the CRA expect?

Documentation is central to preparing a well-supported SR&ED claim.

The CRA expects claimants to retain evidence demonstrating what work was performed, who performed it, when it was performed and how the associated expenditures were calculated.

Documentation created while the work was taking place is particularly valuable. The CRA identifies dated, specific contemporaneous records as strong supporting evidence, although other relevant evidence can also be used.

Technical documentation

Depending on the nature of the project, relevant technical evidence may include:

• Project plans, specifications and technical design documents.
• Records of experiments, testing procedures and results.
• Software source code, commit histories and architecture documentation.
• Project management records documenting technical challenges and development activities.
• Engineering calculations, prototypes and laboratory records.
• Meeting notes describing technical decisions, unsuccessful approaches and subsequent investigations.

These records should help establish the technological uncertainty being addressed, the systematic work performed and the knowledge gained.

Simply maintaining a large collection of technical documents does not establish SR&ED eligibility. The content must support the actual work being claimed.

Employee time and financial documentation

Businesses also need records supporting their claimed expenditures. We’ve found that this can be quite difficult because it is tedious to try to tie the work documented in engineering systems to the payroll data in an HR or Finance system.

For employee labour, this means establishing which employees performed eligible work, what they did and how the claimed time or expenditures were determined. Supporting information may include timesheets, project assignments, payroll records, employee interviews and contemporaneous development histories.

Other expenditure categories require their own supporting financial records, such as invoices, contracts and accounting records.

Quick SR&ED documentation checklist

• Technological uncertainty and advancement sought are described.
• Project scope and timelines are documented.
• Experimentation, analysis and test results are retained.
• Technical decisions and outcomes are recorded.
• Employees involved and their time records are supported.
• Payroll and other applicable expenditures can be substantiated.
• Technical narratives and financial calculations have been reviewed.

This is a preparation aid, not an exhaustive CRA eligibility or filing checklist.

4. Your engineering tools may already contain much of the evidence you need

One of the most significant opportunities for improving SR&ED preparation is making better use of information that businesses already generate through their normal course of operations.

For software companies, project management platforms and source code repositories can provide a detailed history of development activity.

Consider a team developing a new data-processing engine. Over several months, its engineers might create project tickets, requirements documents, investigate performance limitations, commit experimental code, run benchmarks, document failed approaches and implement revised architectures.

Much of this activity may be recorded in systems such as Jira and GitHub.

From engineering activity to SR&ED evidence

1. Existing work records: Jira issues, project histories and technical discussions; GitHub commits, pull requests and code changes.
2. AI-assisted evidence organization: Connect related activity, identify potential SR&ED work and organize supporting records.
3. Human review and validation: Engineers confirm the technical facts, eligible activities and their involvement.
4. Supported claim preparation: Organize evidence, review project descriptions and document employee time allocations.

Rather than asking employees to remember every experiment, an automated system can use these records to establish an initial project history.

AI can help connect related information, summarize development activity and highlight potentially relevant technical work for further investigation. An engineer or SR&ED consultant can then review the evidence, confirm the underlying facts and identify which activities meet the program’s eligibility requirements.

Importantly, a code commit or Jira ticket is not automatically eligible SR&ED work. The records are evidence to be interpreted, not a substitute for technical assessment.

For companies already using development management tools, this approach can reduce the amount of information that needs to be reconstructed manually.

5. How AI is changing SR&ED claim preparation

Artificial intelligence introduces opportunities to automate the administrative work involved in preparing SR&ED claims.

Instead of starting with a blank spreadsheet and a series of interviews, businesses and consultants can begin with existing records, use AI to organize the information and direct human attention toward the areas requiring technical judgment.

Here’s where we’ve found it tends to help most in practice.

Identifying potentially eligible activities

Project descriptions, technical discussions and development histories are a lot to review by hand. AI is well-suited to scan them for activities that may warrant further SR&ED investigation.

For example, a development ticket in Jira describing an unresolved performance limitation, several experimental approaches and comparative benchmark results could be flagged for review.

This helps technical staff and consultants focus their attention on projects with potentially relevant evidence, rather than manually reviewing every development activity (of which there could be millions!).

AI-generated suggestions are not eligibility determinations. A qualified reviewer must establish whether the work meets the applicable SR&ED requirements. However, AI can review eligibility and provide a level of confidence for a human reviewer to consider.

Organizing supporting evidence

Related tickets, commits, test results and technical discussions can be grouped into coherent project histories, which a reviewer can then use to follow the sequence of technical challenges, investigations and results.

Ideally, every material statement in an AI-generated summary should be traceable to its underlying source. This makes the information easier to review, correct and substantiate.

Supporting employee time allocation

Determining how much time employees spent on eligible work is often one of the more demanding parts of claim preparation.

AI-assisted systems can analyze development histories, project assignments and other available records to suggest time allocations. Employees can then review and adjust those suggestions using their knowledge of the work performed.

The objective is not to generate precise-looking numbers from incomplete information. It is to establish a reasonable, supportable allocation based on relevant evidence and employee confirmation.

Drafting technical narratives

Once the relevant work has been identified and the evidence organized, AI can assist in preparing technical project descriptions.

A structured workflow can guide employees through questions about the technological uncertainties encountered, the experiments performed, the results obtained and the advancement sought or achieved. Essentially, AI can run employee interviews to reduce the burden on consultants and give employees the flexibility to do the interview at their convenience.

Those responses, combined with the supporting documentation, become a draft narrative for technical review. The employee or consultant remains responsible for ensuring the description accurately represents the work.

Identifying gaps before filing

The same review process can surface missing information, inconsistent project descriptions and activities that need further explanation before filing.

For example, it may identify a project with substantial claimed labour but limited supporting technical records, or an experimental result that has not been adequately explained. Resolving these issues during preparation is generally preferable to discovering them after the claim has been submitted.

AI should assist your SR&ED experts, not replace them.
The goal of automation is to reduce repetitive administration while preserving the judgment of engineers, financial professionals and SR&ED consultants. Human validation remains essential for determining eligibility, confirming expenditures and approving claim documentation.

6. Moving from annual claim preparation to year-round documentation

A more efficient SR&ED process starts well before the filing deadline. Instead of treating documentation as an annual exercise, businesses can integrate it into their existing development processes. This is where we believe an automated solution can help the most.

An example of a year-round SR&ED workflow

Step 1 — Capture development activity continuously. Engineers continue working in their normal development tools. Relevant project records, technical discussions and experimental results are preserved as work progresses.

Step 2 — Review potentially eligible work monthly. Technical leaders or SR&ED consultants review identified activities, clarify technological uncertainties and flag missing documentation while the work is still fresh. AI makes this easy for them by organizing and surfacing relevant claim materials.

Step 3 — Validate employee involvement quarterly. Employees review proposed project participation and time allocations. AI can automatically analyze and reconcile payroll and financial information so that Finance and/or Engineering just has to review and approve it.

Step 4 — Prepare and finalize the annual claim. The business and its advisers review the accumulated evidence, finalize project descriptions, determine eligible expenditures and complete the required filing documentation.

The benefits extend beyond administrative efficiency. Continuous documentation can make it easier to identify missing records, understand the evolution of technical projects and prepare for a potential CRA review. It also creates a repeatable process that can be improved over successive claim cycles.

7. Preparing SR&ED claims in-house versus working with a consultant

Canadian businesses have several options for preparing their SR&ED claims. Some manage the process internally. Others rely on specialized consultants. Many use a combination of internal resources and external expertise.

The appropriate approach depends on the company’s technical complexity, available resources, claim history and appetite for managing the preparation process.

Approach	How it works	Considerations
In-house	Internal engineering and finance teams manage claim preparation.	Greater direct control, but requires internal SR&ED knowledge and preparation capacity.
Consultant-led	An external specialist manages the engagement and works with company personnel.	Access to specialist expertise, with fees and ongoing employee involvement.
Technology-assisted hybrid	The company and its consultant collaborate using shared preparation software.	Combines automated evidence collection with external expertise, subject to the software’s collaboration capabilities.

These approaches need not be mutually exclusive (we’ve tried all of them ourselves over the years).

A business might use software to organize its technical records throughout the year and engage a consultant to assess eligibility, review project descriptions and advise on expenditures.

Another business might have an experienced internal SR&ED team that uses the same technology to manage the entire preparation process.

In either case, the underlying objective is similar: reduce repetitive administrative work while improving the availability and organization of supporting evidence.

8. For SR&ED consultants: scaling your practice with AI

SR&ED consultants face many of the same preparation challenges as their clients, often multiplied across dozens of engagements.

Each new client may have different development tools, documentation practices, accounting systems and internal processes. Consultants must gather technical information, coordinate interviews, organize evidence, review employee time and prepare claim documentation.

Much of this work is necessary, but not all of it requires the consultant’s specialized expertise.

AI-powered preparation software creates an opportunity to standardize and automate repetitive activities while preserving the consultant’s role as a trusted adviser.

Standardize client onboarding and preparation

Use consistent workflows for collecting development records, identifying potential projects, requesting employee input and organizing supporting documentation. A standardized process can reduce variability between engagements.

Improve client collaboration

Give clients a structured way to contribute evidence and validate information rather than relying exclusively on email requests, spreadsheets and periodic interviews. Consultants can focus their attention on unresolved questions and substantive review.

Increase practice capacity

Reducing repetitive evidence collection and initial drafting can create additional capacity within an existing team. Consultants can allocate more time to eligibility assessments, technical discussions, client relationships and complex engagements.

Develop new service offerings

A technology-enabled engagement can extend beyond annual claim preparation. Consultants can offer year-round documentation support, periodic technical reviews and ongoing advice about clients’ SR&ED processes.

White-label and reseller opportunities

SR&ED consultants may also want to provide claim preparation software directly to their clients.

A white-label arrangement can allow a consulting firm to offer a branded platform as part of its existing service. Alternatively, a reseller arrangement can allow the consultant to offer software alongside its advisory services, creating a potential additional revenue stream.

For consultants, the value of these arrangements extends beyond software distribution. A shared platform can help establish a consistent client experience, maintain engagement throughout the year and make the annual preparation process more predictable.

ActiveSRED is intended to support both businesses and the SR&ED consultants who serve them. Consultants interested in white-label, reseller or client-collaboration arrangements can explore partnership opportunities with the ActiveSRED team.

9. How to evaluate AI-powered SR&ED software

Not all SR&ED preparation software offers the same capabilities.

Some products focus primarily on time tracking or document management. Others provide structured questionnaires, AI-assisted drafting or integrations with existing development systems.

Before selecting a platform, businesses and consultants should evaluate how well it fits their existing workflows and documentation requirements.

Evaluation criterion	Questions to ask
Integrations	Can it connect to the project management and development tools your team already uses?
Evidence traceability	Can reviewers trace generated summaries and claims back to original supporting records?
Human validation	Can employees and advisers review, correct and approve AI-generated information?
Time allocation	How does the system establish and substantiate proposed employee time allocations?
Technical narratives	Does it support structured, evidence-based drafting and expert review?
Collaboration	Can internal teams and external consultants work together with appropriate permissions?
Security	How are confidential technical records protected, stored and accessed?
Consultant capabilities	Does it support multiple client engagements, branding and appropriate separation of client data?

AI-generated outputs should be treated as working material requiring review, not automatically accepted as accurate.

For consultants in particular, evidence traceability, consistent workflows and appropriate client-data controls should be central to the evaluation.

10. A practical roadmap for improving your next SR&ED claim

You don’t need to redesign your entire development process to improve SR&ED preparation. Start with the information and workflows you already have.

Five steps to get started

1. Assess your current preparation process. Identify where your team spends the most time: collecting records, conducting interviews, reconstructing employee hours, preparing narratives or coordinating reviews.
2. Identify existing sources of technical evidence. Review the project management systems, code repositories, engineering records and other tools your employees already use.
3. Establish a regular review process. Assign responsibility for identifying potentially eligible work and reviewing supporting records throughout the year.
4. Introduce automation selectively. Start with a limited number of projects. Evaluate whether AI-assisted evidence collection, time allocation and drafting reduce manual effort while preserving accuracy and traceability.
5. Measure the results and refine your workflow. Track preparation hours, employee involvement, documentation gaps and review effort. Use these measurements to improve the process before expanding it across more projects or clients.

For a business preparing its own claim, this could mean piloting an automated documentation process with one engineering team.

For an SR&ED consultant, it could mean introducing a standardized preparation workflow with a small group of repeat clients before rolling it out across the broader practice.

The objective is to build a process that becomes easier to operate and more consistent with each claim cycle.

11. Frequently asked questions about SR&ED claim preparation

Can AI prepare an entire SR&ED claim?

AI can assist with evidence collection, project organization, time allocation and technical narrative drafting. However, employees and qualified advisers still need to verify the technical facts, assess eligibility and confirm the accuracy of claimed expenditures.

Businesses remain responsible for their submitted claims, even when software or external consultants assist with preparation.

Can Jira and GitHub be used as SR&ED supporting documentation?

Yes. Project management records, source code and related development documentation can provide relevant evidence of technical work.

However, the existence of a ticket or code commit does not establish eligibility. The records must help substantiate the technological uncertainties, systematic investigation and advancement associated with the work being claimed.

Do I need an SR&ED consultant to prepare my claim?

Businesses are not required to engage an SR&ED consultant. They can prepare claims internally, use external advisers or adopt a hybrid approach.

Consultants can be particularly useful when eligibility assessments, expenditure calculations or complex technical projects require specialist expertise.

How much employee time should be allocated to SR&ED?

There is no universal percentage. The amount depends on the eligible work each employee actually performed and the applicable expenditure rules.

Businesses should maintain reasonable supporting records and use them to substantiate their claimed labour expenditures.

When should I start documenting SR&ED activities?

Ideally, documentation should begin when the relevant work starts and continue throughout the project.

The CRA emphasizes the value of contemporaneous evidence, particularly records that are dated and specific to the work performed.

What is the deadline for submitting an SR&ED claim?

For most corporations, the SR&ED reporting deadline is 18 months after the end of the tax year in which the expenditures were incurred. Extensions are not granted, and a claim that is filed late or incomplete is denied, not delayed. The CRA recommends submitting the claim with the income tax return rather than waiting until the final reporting deadline.

Can SR&ED consultants use ActiveSRED with their clients?

ActiveSRED is intended to serve both businesses preparing their own claims and consultants assisting their clients.

Consultants interested in white-label, reseller or collaborative preparation arrangements should contact the ActiveSRED team to discuss available capabilities and partnership options.

12. The future of SR&ED preparation is continuous, collaborative and AI-assisted

SR&ED preparation doesn’t have to be a disruptive annual exercise.

Businesses already generate substantial amounts of technical information through their everyday development activities. With the right processes and tools, that information can become the foundation of a more efficient, better-organized claim preparation process.

AI creates opportunities to automate repetitive tasks, organize evidence, assist with technical narratives and help employees and consultants focus on the decisions that require their expertise. A business spends less time reconstructing the past and more time developing new technology; a consultant gets room to take on more clients without a proportional increase in headcount.

AI’s job here is to handle the repetitive parts of preparation — not the judgment calls. Those stay with the people who understand the work.

This guide provides general educational information and is not tax or legal advice. SR&ED eligibility and allowable expenditures depend on the circumstances of each claim. Refer to the CRA’s current guidance and consult a qualified adviser where appropriate.