SR&ED is a massive opportunity for Canadian startups, yet many founders still underclaim. What core problem is Beaverly solving that others have missed?
The underclaiming problem has three layers, and they compound:
- The first is awareness. Every year, thousands of new startups are founded by people tinkering, building, and solving hard technical problems, often without knowing SR&ED exists. And even when they hear about it, the instinct is usually “that probably doesn’t apply to us.” Figuring out whether it does requires engaging a consultant, going through the program, and spending time they don’t have. At the early stage, every hour a founder spends on tax administration is an hour not spent talking to customers or building. So many eligible companies simply opt out before they ever get started.
- The second layer is recall. For founders who do file, the traditional process asks them to reconstruct months of technical decision-making from memory at year end, long after the work happened. Engineers don’t remember why they went down a particular path. They don’t remember the dead ends, the pivots, the moments of genuine uncertainty. Memory is not evidence.
- The third layer is evidence quality. Even when teams do document their work in tickets or commits, that documentation was never written with SR&ED in mind. The intent is there. The work is there. But the evidence is thin, inconsistent, or missing the specific signals CRA looks for, and by the time the consultant flags it, it’s too late to go back and fix it.
All three of these compound into the same outcome: underclaiming what was legitimately earned, overclaiming through inaccurate reconstruction, or spending enormous time and effort on back-and-forth that still produces an incomplete picture.
Beaverly was built to solve all three at once. We connect to engineering environments from day one — GitHub, Linear, the tools where real R&D happens — and capture evidence continuously as the work unfolds. Founders don’t need to know SR&ED terminology to benefit. Engineers don’t need to document anything differently.
And when claim season arrives, the evidence is already there: contemporaneous, specific, and grounded in how the work actually happened. Canadian software companies doing legitimate R&D deserve claims that reflect the full scope of what they built. Beaverly makes that possible.
How does Beaverly integrate into developer workflows like GitHub without slowing teams down — and what has adoption taught you about engineering culture?
The design principle behind Beaverly is simple: engineers should never be burdened by it. We connect to existing engineering systems — GitHub, Linear, your dev tooling — and work in the background. No new tools to learn, no tickets to fill out, no meetings to attend. The integration takes minutes and the ongoing experience for the engineering team is as close to zero friction as we can make it.
What adoption has taught us is that engineers aren’t resistant to SR&ED, they’re resistant to being interrupted and second-guessed. The traditional process treats engineers as a resource to be interviewed rather than a source of evidence to be observed. When you remove the burden entirely, the resistance disappears. Engineers are actually proud of the technical work they do. They want it recognized. They just don’t want to spend a Friday afternoon trying to remember what they were solving in Q2.
Modern software development has changed dramatically — AI-assisted coding, “vibe coding,” prompt iteration. How does Beaverly handle capturing R&D evidence when the work itself looks nothing like what CRA originally designed SR&ED around?
This is one of the most important questions in the SR&ED space right now, and most of the industry isn’t talking about it yet.
CRA’s framework for SR&ED was designed around a model of R&D that looks like a lab — hypothesis, experiment, result. Modern software development doesn’t look like that. It looks like a developer iterating through forty commit variations at 11pm, using an AI coding assistant to explore three different architectural approaches simultaneously, reverting, branching, trying again. The technical uncertainty is real. The systematic investigation is happening. It just doesn’t map cleanly to a T661 narrative.
What we can say is that Beaverly was designed from the ground up for how engineers actually work today, not how CRA imagined they would work twenty years ago. The signals that indicate genuine R&D are still there in modern development. We capture them. That’s true whether a developer is writing code from scratch or prompting their way through an unsolved technical problem.
If anything, the rise of AI-assisted coding makes continuous automated evidence capture more important, not less, because the work moves faster and the paper trail evaporates even quicker.
You position Beaverly as enabling consultants, not replacing them. How do you see the future of SR&ED evolving between automation and human expertise?
SR&ED consultants are experts in claim strategy, CRA navigation, and audit defense. None of that is going away, and none of it should. The technical and financial complexity of a well-prepared SR&ED claim requires human expertise that software cannot replicate.
What software can do is solve the raw material problem. Right now, even the best SR&ED advisor in Canada is working with incomplete inputs: fragmented engineer recalls, sparse ticket histories, and whatever the founder half-remembers from eight months ago. The constraint was never the consultant’s expertise. It was always the quality of what they had to work with.
Beaverly’s role is to make sure that when an advisor sits down to build a claim, the evidence is already there: structured, timestamped, and grounded in actual engineering activity. Richer source material. Stronger technical substantiation. Better audit defensibility. The consultant’s judgment becomes more valuable, not less, when it’s applied to complete evidence rather than partial memory.
The future of SR&ED is a clear division of strength: automated infrastructure for evidence capture and initial eligibility detection, human expertise for claim strategy, CRA navigation, and audit defense. Beaverly owns the first part. Consultants own everything that requires knowing the client, knowing CRA, and making judgment calls that no software should make alone.
Beaverly.AI
You’ve made Beaverly free for software companies, which is an unusual call. What’s the business model, and how do you think about building a two-sided marketplace between engineering teams and SR&ED consultants?
Making Beaverly free for software companies comes down to one observation: startups are the biggest skeptics in this space, and for good reason. SR&ED has a reputation for being complicated, consultant-driven, and hard to access. Early-stage founders are already stretched thin, and the last thing they’ll do is pay for another tool before they’ve seen it work.
But there’s a more urgent reason. As AI tools become the default for everything, we’re already seeing some founders turn to LLMs to help them identify SR&ED eligible work, draft technical narratives, and piece together claims. The output looks reasonable. The problem is that it’s often brittle, built on assumptions rather than contemporaneous evidence, and optimized to sound eligible rather than to actually be defensible. Founders won’t realize the claim is weak until CRA comes back with audit questions, by which point it’s too late to fix the evidence. The same way someone might ask an AI about a medical symptom and only discover the answer was wrong when there’s already a consequence: the error is invisible until it isn’t.
We’d rather give startups access to something that actually works — evidence grounded in what their engineers really did — than watch them build fragile claims with tools that weren’t designed for this. Making it free removes every barrier to doing it right. Between using an LLM and using Beaverly, the cost is the same: free. The difference is that one was built for this.
The business model works through the advisor relationship. SR&ED consultants and accounting firms are our paying customers. For them the value is straightforward: Beaverly helps them do their work faster, with better source material, and stronger audit defensibility. That’s a clear return on investment for a professional services firm.
The result is a flywheel. Startups adopt Beaverly because it’s free and frictionless. Consultants adopt Beaverly because it makes their practice better. Each side reinforces the other, and the entire SR&ED ecosystem ends up with claims that are more accurate, more defensible, and less painful for everyone involved.
You’re sharing the floor at Web Summit Vancouver with some of the most well-funded startups in the world but Beaverly is pitching a new approach to how software companies capture R&D evidence. How does that story land with an international audience, and do you see a version of what you’re building that travels beyond Canada?
The story lands better than you might expect, because the underlying problem is universal. Every major economy has some version of an R&D tax incentive: HMRC’s R&D relief in the UK, the R&D Tax Incentive in Australia. Every single one of them suffers from the same fundamental flaw: companies have to prove work that already happened, using evidence that was never designed to be captured in the first place. The proof gap is not a Canadian problem. It’s a global one.
Canada is the right place to start. SR&ED has a well-developed ecosystem of consultants and accounting firms, the proof gap problem is acutely felt across thousands of software companies filing every year, and the audit scrutiny is real enough that companies take evidence seriously. The international expansion story follows naturally, but we’re clear-eyed about what it involves.
The upstream problem is the same everywhere: engineering work happens, evidence evaporates, and companies scramble to reconstruct it after the fact. Beaverly solves that upstream problem. The downstream translation, structuring that evidence for HMRC versus CRA versus the Australian Tax Office, requires local expertise and local regulatory knowledge. That’s not something we bypass. It’s where local advisors and consultants remain essential, in every market we enter.
So the international story is really two stories: one about a platform that captures R&D evidence continuously, anywhere engineers work, and one about building the right advisor partnerships in each market to make that evidence count.
What do you hope to achieve at Web Summit Vancouver 2026, and how does it fit into Beaverly’s growth and fundraising strategy?
We started Beaverly because we felt the pain firsthand. We’re Canadian founders who care about pushing Canadian innovation forward, and we watched too many legitimate R&D claims get watered down because the evidence simply wasn’t there. That’s where this started.
Web Summit Vancouver is where our users are. Software founders, CTOs, engineering teams building real things. That’s our community, and showing up there is how we put Beaverly on the map with the people who need it most. We want founders to leave Vancouver knowing Beaverly exists, and knowing their engineers never have to fight to prove what they built.
The longer we’ve worked on this, the more we’ve realized the opportunity is bigger than we originally thought, in two directions. The first is global: the proof gap exists everywhere engineers work and governments offer R&D incentives. The second goes beyond tax credits entirely. The engineering activity Beaverly captures is institutional knowledge. The record of how a company builds, what problems it has solved, where its genuine IP lives. That has value well beyond claim season.
The SR&ED process has worked the same way for decades. It was designed for a different era. We’re not here to disrupt it — we’re here to bring it into the world software teams actually live in today.
As for the name: beavers are ecosystem builders. They work quietly in the background, building infrastructure that makes the whole environment function better. That’s what we want Beaverly to be. The team behind the scenes, making SR&ED work better for everyone who touches it. Quiet, helpful, always building. And very Canadian.
Web Summit is where we plant that flag: connecting with our users, building relationships with SR&ED advisors, and having early conversations with investors who believe the SR&ED process is overdue for change.
Beaverly will be at Web Summit Vancouver, May 11–14, 2026 at the Vancouver Convention Centre. Meet Trang Nguyen in person at booth A6-43 on May 14th.
