Unlock AI Savings: R&D Tax Credits for Builders

Hello and welcome to the Runpoint Podcast. I'm your host, Matthew Hall, and this is my co-host and partner, Sam Gaddis. Sam, how you doing today?

I'm doing great, Matthew. Ready to get into it.

Perfect. Perfect. This is a podcast about building cool stuff with AI primarily, and that's what Sam and I do at Runpoint Partners. We build cool stuff for AI for internal teams. And we've got a very special guest today, Ari Salafia, to talk to us about a really important concept. Which is R&D tax credits, how to build cool stuff with AI while also saving money. It's an important time of the year. We're, we're really close to year end, November 2025. A lot has changed this year in terms of tax credits with the one big beautiful bill and how businesses can capitalize their expenses and what they can do going forward in terms of getting the most out of their AI initiatives. Ari, welcome to the podcast. How you doing?

Hey guys, it's great to be here. Thanks so much for having me.

So for our listeners, Ari is CEO and founder of a company called TaxTaker in Austin, Texas, and she and her company have helped businesses return $100 million through R&D tax credits and other tax credits. They're not R&D tax credits specialists only, but clean energy as well. Ari, why don't you give us a brief overview of yourself and TaxTaker?

Yeah, thanks so much. And very impressed that you got my last name correct right off the bat. That's a feat, a Sicilian feat in and of itself. Yeah, so, uh, TaxPeaker is here really to extend runway for companies that are also building cool stuff. And a lot of those now are building cool stuff with AI. So, I'm really excited about this conversation. I got into this space almost 15 years ago and was hooked. One, because I get to meet really innovative, passionate, creative founders every single day. And so, I would expect much like the both of you, getting exposure to all different types of of technology is, is really exciting. And then who doesn't want free money, right? So oftentimes we don't get into business thinking about our accounting, let alone our taxes, especially when we're at it for the first few years. Maybe, maybe you were bootstrapped or maybe you have VC funding. You're thinking taxes, who cares? But the fact of the matter is the, the government has an enormous amount of incentives and grants throughout the country. And the R&D tax credit has been around actually since since the '80s, but it's become more favorable in the last decade or so. So TaxTaker was built to help companies better claim these incentives without, you know, a ton of business disruption. And it's been a heck of a ride and a lot of fun.

That's perfect.

That's perfect.

So why— go ahead, Sam.

I was just going to say, I think it's really cool what you do. And I think just to provide a little context on how we actually got interested in this, taxing taxes in general is not something that I usually jump to. To talk about or think about, but we had a client that was, kept pushing us to track our time. And that's not something we typically do. We work on retainers. We don't think about tracking time. And I found it to be pretty annoying, frankly. And after pushing back and not doing what we were supposed to be doing enough, he finally sat us down and said, hey, the reason that we're doing this is because if you track your time in a certain way, and if we're doing the right things, we can actually claim tax credits that effectively reduce our fees, RunPoint's fees, significantly so that the contract that we have with them is much more palatable, meaning they will extend it, we can do more, we can go further together. And suddenly this became very exciting to us. So that's, that was kind of the genesis of us reaching out to you. And I think that's really cool.

It's super helpful context and another example of why context matters. And I hope we'll get into this at some point today, but definitely want to share even some tactical tips for folks inside of organizations to be able to bring this information kind of upwards of, you know, why do we do things or why something might matter because there might be actually monetary value or extension with it.

That's perfect. So that's—

I'm sorry you had to get uncomfortable tracking, because I know it's a common objection.

You know, we, we actually found a way to automate it with AI. So we're good.

Yep.

Well, we'll have to hear about it.

And it was one of those things of us, us learning more about the ins and outs of it, I think makes us better partners because if we understand the types of projects that qualify and don't qualify, we're better at pitching new features. We're better at everything we do about kind of understanding the, the spirit behind the legislation. Taking more risks, right? It's like getting your partners to take more risks is a hard thing to do, you know, from the outside. But so if there's financial leverage to do so, uh, we want to be able to exploit that. So that leads me to my first question for you, Ari. Why should businesses care right now in particular about R&D tax credit?

You know, they should really care because they're likely already spending the money and doing the work. So why not claim the dollars available to them? That's what people find most surprising. Most companies will say, if they don't have any experience with this, will say, oh, what do I have to do to get it? And I reframe that up around what are, just walk me through kind of what your company is building. And oftentimes. A light bulb goes off and they're like, wow, I just didn't realize everything I basically do qualifies, right? And that's not always the case. We know that we do routine things and, and kind of to kind of your earlier point, right? We don't necessarily want to steer where a company is doing certain projects and not others, right? For the sake of a tax credit alone, but we often find that overwhelmingly, especially, you know, technology companies, they're already spending the money, so they should be getting this on the backside.

What do they stand to gain?

Yeah, so The R&D tax credit is really a wage-driven credit. So we're looking at a few key buckets of expenses, namely the direct engineers or software developers or designers, if we're talking about kind of a tech company, their involvement in building new and improved features or projects. And then we're also looking at those that supervise them or support them. So you might have a C-level team that might not be you know, engineers, or they might be engineers, but they're not necessarily in the code in the daily. And then folks that are supporting those roles. The next bucket of expenses we're looking at are outside contractors. So that could be, you know, AI consultants like Runpoint or one-man kind of developers. And then the next bucket of expenses we're looking at is around cloud. It's, it's not as much, but once a company burns through their AWS or Google Cloud, we can look at a portion of their testing and staging environments prior to commercialization. And for our physical companies that have tangible products, we are looking at supplies consumed in the R&D process.

Yeah, theoretically a business who the project they did this year is 100% qualified in those, in those buckets. It's innovative. We'll talk about what that means in a second, what it means to be qualified, but let's just say it costs $1 million between internal resources, external resources, hardware, whatever else. What percentage, how would they actually go about qualifying for the R&D tax credits and what percentage of those dollars do they see back? How does, how does the, how does the mechanism work?

Yeah, so activities absolutely matter. So I'd love to go into that, but, uh, the, to get just right ahead of it, why should they be looking at it? It's usually close to 10% of their spend. So on a million dollars, you're looking at close to $100,000 on average, uh, which That's huge. 10%, that's effectively a yes to another project. That's potentially another hire or that's immediate kind of reduction in either tax or refunds for your payroll taxes. So it really can increase your cash flow and your bottom line really fast. I think people like to get ahead of the dollars right away, right? And we do have to take a step back just to make sure that we are solid on the activity standpoint. It is an activities-first incentive. So there are a few tests that we have to overcome. And now with the big beautiful bill, we are so excited because it's not so much what's changed, it's kind of going back to how it was before the TCGA. So the big beautiful bill basically restored the immediate deduction of R&D expenses. In the last few years, The, the legislation lapsed and basically companies had to amortize their expenses over a number of years, 5 years domestic and 15 over foreign. And that threw people into some uncomfortable positions. There were some companies that were insulated if they're pre-revenue, but others all of a sudden were thrown with big tax bills. So everyone's cheering because now it's a lot more simple. The tax credit's what you get. It's a dollar for dollar offset against your income or your payroll taxes, and it's fully 100% expense deductible. In the year in which it occurs. So do you want to dive into kind of how you qualify? Would that be—

I would.

Yeah.

And maybe this is, this is my own misunderstanding. Tell me the difference between the dollar-for-dollar expense versus the 10% that you mentioned in terms of the tax credit. Do those, are those things working in concert? Are those the same thing? What's, what's—

Yeah, no. So I just mean that a tax credit is a dollar for a dollar. So once we actually get to the number of tax credits available, that's a dollar-for-dollar offset against any income taxes or payroll taxes owed. And it's also immediately getting expensed on your balance sheet.

I see.

Versus having to carry that expensing over a number of years.

So, but the percentages are applied, the percentages are applied differentially to the buckets. So like the expenses would be 20% and cloud is some other percentage. Is that right? That's how you derive the amount?

Uh, no, not necessarily. So we're looking at all buckets of expenses. And we're summing them. Now, not each bucket carries the same weight. So a wage earner, because the spirit of the program is really to create job growth and boost innovation here in the United States. So there's more weight inside of the calculation in the bucket for a W-2 wage earner than that of a 1099 contractor. So for simple math, let's say you have a $100,000 software engineer, they're spending a majority of their time on new and improved technology. They can, you can get up to $100,000 of their salary. So that'd be $10,000 in R&D tax credits. On that same $100,000 in spend with an outside consultant that's US-based, even if 100% qualifies, you're going to get 65 cents on the dollar. So you'd get $6,500 off that $100,000. Um, so you're sticking a 35%. Percent haircut, and then the last expenses are, you know, just relative to the spend, they're smaller, but that's sort of just the weighted system.

Term I see in this a lot is qualified research expenses, QREs. Can you explain to me what those, what?

Yeah, so those are those buckets of expenses. Those are the employees, the consultants, potentially the cloud, and potentially supplies if there are consumables.

Perfect. Yeah, I think that's a great segue in terms of what kinds of projects qualify and what don't. What are the, the guardrails here and what's the system set up for?

Absolutely. So whenever you're talking to a CPA about this or an advisor there, or you just maybe, uh, are looking this up online, you're going to see something called this 4-part test. And that's how we qualify activities. And then we, I'd love to get into the real-world examples because that's where things tend to click. But essentially, uh, a company needs to meet these tests and it's first is, is the company doing something that's technological in nature? Are there activities based in any of the hard sciences? Or computer science? If so, we move on to the next bucket. This is actually going to be the one that's most looked at, especially going forward. The IRS has made statements about this. This is where folks can get really tripped up under an audit, and that's the business component. A company has to clearly define what's being built or enhanced, what product it is that they're working on. Surprisingly, a lot of companies can get flimsy on this. We need to be able to also document, you know, how it's intended to function, just why, why the work really matters from the business's perspective as a business component. Sometimes that's just one component over the course of a year. Sometimes that's several components. And then we move on to uncertainty. We have to be able to demonstrate that there is some type of risk for the taxpayer. You know, a lot of us like to say like, oh, we're going to build this and it's going to be great and it's We're going to nail it right away. But the fact of the matter is like we fail all the time or, you know, talk to enough business owners and they'll talk more about the more failures you have is usually a leading indicator of how successful they, they are, right? So also not discounting that for purposes of documenting your time and the credit. Something doesn't have to be yet commercially available yet or highly successful or monetized in the market to qualify. You could have a number of failed projects over the course of of the tax year. So being able to demonstrate that, the IRS loves risk and failure. And then the last bucket and where your time tracking probably came really to be helpful is this iterative process of experimentation. So just being able to demonstrate that you've, you know, been able to test different functions, different kind of options or steps along the way. You might have a good idea of how it's going to be built, but you're testing along the way and iterating along the way to make sure that that business component is functioning as intended, or maybe it changes. So those are the 4 things that we need to, to take a look at. We could get into this, but if a company is building software for internal use, that it's not going to be sold to an end user, there are a few extra steps that you also have to overcome as well.

Great. Let's, let's talk about those now. What are those for, for internal use?

Yeah. So essentially internal use software and why I think it's actually very important to talk about it now with the advent of AI and AI projects is that not only does it have to be innovative, but it has to be a substantial improvement upon what the company is already doing, using, building, and it needs to carry significant economic risk. Those are the top two. So it's not to say, oh, you know, I only spent like $2,000 on this, so therefore, or 2 hours of time, therefore this is a qualified project. Like that's not going to be material beneficial. They really need to be able to demonstrate why it really kind of goes back to that business component, why it matters and what they stand to lose if unsuccessful. Will they have a piece of their business break if this is, if it isn't deployed, if it isn't mission critical? And then the last bucket is that it's not commercially available. But we do see sometimes it working so well for a company that they ended up using that for their industry. And so then they'll go out and license and sell that out, and then it doesn't even have to pass those extra test because then it just defaults to the 4-part test.

Got it. Got it. Yeah. I've, um, and there's also some, there's some, there's some stipulation about like customer-facing at all, right? So, it could be an internal tool, but if there's parts of it that are customer-facing, then that might get you into the, the other bucket. Is that right?

Bingo. Yes. Yes. So, if there's some type of communication interface of any sort, you know, plugin, any type of, you know, back and forth between a customer, usually that's enough to just go back to the 4-part.

How do you, Ari, how do you define if something is innovative? How do you, I mean, I can, I can just see myself trying to justify this both ways. How do you, how do you say whether it's important or not?

Yeah, so I mean, it's going to be closely paired with the initiative at hand and the economic risk. So what I mean by innovative is mostly that it's not just minor upgrades. Right. It's not just, oh, I, you know, had to spend some time custom configuring, um, an off-the-shelf product, uh, to fit a workflow that my—

that everybody already does—

team use. Yes.

Yeah. Got it. Okay. I think this will be, I think this will probably become a lot more clear when we dive into the specific examples.

Do you want to jump into that now and then come back? I think that, I think that's, I think we're all intrigued.

Yeah, let's weave it in.

Let's do it.

I have so, I, yeah, I have a bunch that I want to ask about.

Yeah, yeah, I want to throw a couple at you and you tell me how you go about, how you go about breaking it down.

Let's do it.

First is, let's talk about, um, something we see a lot, this trend, which is because, you know, SaaS has now been around for 20 years and businesses have these longstanding relationships with major SaaS providers. Let's just use CRM as, as an example. You know, Salesforce is the giant, but there's vertical CRMs for pretty much every business. And we see more appetite now for building companies, building their own CRMs, right? It's with AI coding assistants, it's gotten easier and cheaper to do that. But then they also want to, they want to customize it specific to their business. They've been frustrated with the existing workflows out there. So they want it to be exactly 100% of what they want, not 60% of what they want. And they want to start from scratch so that they can take advantage of AI capabilities throughout the, the workflow. And maybe combine not just CRM, but some ERP functionality or some, you know, invoicing and some customer-facing stuff, but basically make a stack of their own that they, that they're using some open-source stuff for, right? That's not like they're starting literally everything from scratch, but they are building a suite of solutions for CRM for themselves, you know, in a specific industry. How would you go about consulting that, that client and, and what parts of that qualify and what don't?

Sure. So right away, right? That's a huge business component. I mean, that's, several functions of their organization's kind of critical needs throughout their business's lifespan, right? So you named one thing. I mean, CRM, CRM alone is, you know, a worthy enough business component. I saw this actually maybe 7 or 8 years ago where we worked with a business in the aerospace industry that quite literally was not allowed to use Slack just because of potential security vulnerabilities. So they had to create something that was Slack-like for internal communications. And then they said, okay, you know, should we bang our heads against trying to mitigate kind of APIs with other tools that we use, or should we just keep going? And the answer was we got to keep going and just build our own, build our own tools, which was probably a lot more expensive then than it would be now. And so that, I mean, that passed even then. So now it's, it's just a little bit nicer that it's a little bit easier, right? So no, that would be it. That could be a very highly qualifying project. I would say, you know, my only pause where I've seen folks sort of make a stretch is that they are just layering on kind of little improvements upon existing off-the-shelf SaaS solutions. Being that this is completely scratch, they want to get this right from the get-go and not have to, you know, systematically pair. As needed. That should be something that passes the smell test right away.

But if you fork an open source CRM and you're changing fields, you're adding custom layers, maybe doing AI analysis on field service workflows or something like that, do you need to delineate, we spent half the time configuring the CRM and half the time adding on new stuff?

Yes. Yeah. We want to parse out what's custom configuration versus kind of new novel development.

Cool. Got it.

Essentially, there's a reason why is because the IRS doesn't want two taxpayers to claim the same dollar. So any kind of off-the-shelf suite of project or open project that could be claimed cannot be claimed by every person that touches it.

Oh, right. It's already been built. Yeah. Yeah. Yeah. Um, that does make sense. And so for that type of business, Is it better to, you know, set up your projects with this in mind? So you are, I don't know, you might be even labeling tickets as new and novel or off-the-shelf config or something like that so that it's easy to, to, to go back and audit later. Or is this something that you can do retroactively and look at the project once it's done and slice it up afterwards?

Yeah, I mean, I would obviously say yes, but that's also very unrealistic, right? Yeah. Again, go back to kind of the earlier message around like, we're not building for tax credit optimization. We're building to build, you know, to build amazing products that function working well for teams and customers, right? And then with the benefit of like, now let's go actually document and tag things appropriately so that we get the best of both worlds. People laugh when I say this, but the IRS actually can be reasonable. And I mean reasonable by they do still allow for reasonable estimations of time. Meaning while time tracking is the gold standard, it's, it can be inherently flawed by the organizations that use it effectively. You could have two developers. Day to day looking like they do the same things, logging their time very differently. So that also takes, you know, another person to, you know, to keep those standards kind of accountable and up to date and, you know, autonomous. But we are allowed to go back and make appropriate estimations of personnel's time. But the, the key there is we have to clearly be able to see, you know, what projects they were, were touching. And, and back up that information of their involvement. And then as well, we still sometimes have to do that because if you just took, you know, developers and their, you know, GitHub or like Jira commits, you could be missing those folks that were supervisory or support that are not in the code, right? So sometimes it's, it's also a mix of time tracking plus percentage estimations.

All right. Hypothetically, if a business was contracting with an AI consultancy and implementation shop, I don't know, maybe rhymed with gunpoint or something like that, what should that business be thinking about in terms of setting up their contract terms with those contractors to take advantage of the R&D tax credits?

Yeah, so coming right off of that, what's US versus non-US, we want to be able to decipher who's involved and This is where terms really matter when it comes to payment terms. So it should always have, right, the client retaining the IP. That should be explicit in the contract. That's number one. Secondly, how the payment terms are structured matter. The IRS really wants to come back to who's bearing the most economic risk, especially if this project is unsuccessful. So generally speaking, a time and materials agreement is going to be a favorable kind of checkbox for, for the customer, because there's the ability to continue to accumulate hours and involvement, and they're holding the bag at the end of the day, versus a fixed fee arrangement. And now we see a lot of folks that are offering flat fees to kind of become more competitive and create maybe like manage expectations for the customer. But that being said, the IRS views those as the contractor really assuming the financial risk or because at the end of the day, there's a great ability for scope to increase and not, I mean, you can always kind of add on for that. So that's important, right? If there's scope creep, but basically the contractor, if they're set at a flat fee, there's an ability for you to go over, but only be paid for that agreed upon price. So if you do have a flat fee, one sort of, it's, it's not a perfect workaround, but a short-term solution would be let's heavily document the time and materials by that contract, the time by that contractor, much like your client kind of asked of you, but it would just make more sense to restructure your contracts, at least going forward, so that there's a time and materials, and then you just have, you know, more regular milestone check-ins so that you can control, right, cost, but time and materials are a green light for a few if you've got consultants fixed fee, pause, and assess.

Ari, one quick question on that. What about the other alternative, which is a monthly retainer where you— Fixed capacity. Yeah, it's fixed capacity. So you're getting, you know, paid X amount a month, but you don't specify when the project is gonna be complete.

Yeah. So I mean, perpetuity, you're— that customer's paying for, for risk, right?

Yeah. That's what it seems like.

Yeah, absolutely. I would still say let's continue to just always be documenting involvement and, and tagging types of project, right? What's kind of R&D versus what's maintenance or quality control.

And so just to put a bow on that, location, so where is the work getting done? IP retention, so making sure that the client maintains the IP after it's built. And then financial risk is the last one, right? Where's this?

Got it.

Here's another one I see all the time. People hate their BI tools, their business intelligence tools. Like they're overly reliant on analysts to create dashboards for executives, and then the executive changes what they want in the dashboard later on. And so a thing we see all the time is it's just now so much, you know, faster to enable those executives to create their own dashboards with AI or a system by which they can use AI to do data analysis and do kind of really data science, but with the AI doing some of the lifting. And obviously That's a, that's, that's a big bucket of, of need, right? And I think a lot of it, my hunch is a lot of that would not qualify, but I want to know what might make it qualify, right? It's like building internal dashboards, internal tools, and that whole suite. What are the ins and outs?

Yeah. So again, we're going to be analyzing, right? Is, is it new or are we just kind of customizing off of a dashboard that we like? If new, then, which it sounds like that would be the case here, you know, who was involved? Did they hire, you know, a firm like Runpoint and then Let's also be reasonable on those executives' time. How much time was dedicated to inform, you know, test, provide feedback to your team on this project? Is it, you know, 5%, 10%? We need to look at their year. It would be a big red flag for me if, you know, a highly compensated exec was saying, oh, 100% of my time was spent on my dashboard. Unlikely, right? They need this dashboard so that they can do all the things that they do over the course of the year. That's where I see folks reach. Truly on projects is where they're effectively saying where their time is and associated costs. And that's going to be something that would get scrutinized. So, I would say, yeah, that could be one of many kind of initiatives for a business over the course of the year, but we need to, one, look at who's involved, and then is there really material benefit in pursuing it?

Great. All right.

Okay. I've got one. I'm going to jump in, Mathew. So—

This is fun. I like these scenarios.

Company has an internal process where they have to let's say, source the right person for the job, resource allocation. This is something that they've been doing forever, but it's been a laborious process. And so they hire a consultant to build a system that aggregates from all of their existing internal systems, bringing different data together, and maybe even scrape some external information to create a new interface that still does the same thing. It just enables somebody to do resource assignment better than they used to. And, you know, yeah, that's it.

I would, I would put it similarly into the dashboard bucket. That would be sort of another one-off project that one, we would need to know the scope of what was involved by the consultant, you know, what was kind of true development versus what's kind of just known kind of guidelines.

So it needs to be broken out by project.

Yes, definitely. So it is, we actually need a document. Unless you get to a size where we can, that we just should pull a sample, we need to be documenting each project at a project level.

Interesting.

Okay, cool.

I've got, I've got a different one that I think might be more closer to the not 100%, but more, more applicable.

Um, I'm not sure to say everything qualifies to, you'd be spending a ton. I mean, we're the firm too where we're like, no, and they're like, but, and we're like, no.

But thematically, we see businesses who see AI as an opportunity to develop new service offerings or extend an existing service offering in a meaningful way. Maybe they've been selling a consulting service of some kind. Maybe it's an audit, maybe it's a technical transformation service. It could be anything. It could be wealth management. It could be all sorts of things, right? But they see AI as this way of extending what those experts can do. So to go, maybe go down market for a much cheaper offering that where a lot of it's automated, where the AI is performing things behind the task, maybe they're training, fine-tuning their own models as part of that process. Right. But, um, that to me speaks to more of the spirit of the legislation of this is a business who is transforming their go-to-market and their actual, their actual product with, with AI. Does anything change about the, the, the equation there or how they would set it up or think about how that qualifies.

I think so. I mean, that's really an— they're creating kind of a SaaS offering, right, for their business. They're, they're enabling themselves to be one-to-many versus, you know, a consultant having to just one-by-one with a customer, right? So that they can come downstream. Similarly, years ago, we, we drank our own Kool-Aid. We said there had to be a way to automate, you know, parts of this study for customers so that we could gather information faster so that they can spend less time on this and more time continuing to build cool products and get the money they deserve. So we had to translate our tacit knowledge and decades of experience in, in these tax codes to be able to filter it through almost kind of like a TurboTax experience, if you will, for some of our smallest clients so that we could come downstream. So in some kind of a similar way, we've done that. We've qualified for our own R&D tax credits and, and benefited along the way. So I think anytime that you're, you know, being able to extract kind of knowledge and truly build a tech product around it, that should be something that's absolutely evaluated.

Amazing. Let's, let's pivot to what advice you would give to that per— that, that persona we talked about earlier, which is an enterprising employee at a business who wants to, you know, wants to create something new, wants to transform the business, is classically trying to get a business case authored and approved and sees potentially a way to offset some of the costs with this. What advice do you give them either in how they set the project up, how they might contract with, with a business like ours, or anything else that they should be aware of?

Yeah, well, so timely. I hope they see your podcast during year-end planning when companies are lining up budgets, right? And budgets have gotten really tight. You know, we've seen that the last couple of years. So for an engineering manager, I mean, if they know about this or they hear about this, this should be very empowering for them because they can make a case to say, hey, I think that this project is going to involve X amount of personnel, whether that's internal or by leveraging an outside firm. This is what we're hoping to accomplish. This is the business case, but then also here is what we expect it to cost, but then being able to kind of right away line out, but we should also be able to, it's actually on, it's on discount. It's not just girl math. It's real. I can spend to save, but we should expect that, you know, around 10% of this budget should come back to, to help the business. And at the business level, depending on the size of the company, if a company is less than 5 years old with less than $5 million in revenue, they can use this against their payroll taxes, and it's not just the payroll taxes of the technical personnel. So it's generated from that department, but it applies to the entire organization. So that can be very powerful and very compelling case for a manager to bring to a business ops or finance person. If a company is profitable, paying tax, or older than 5 years, it's an income tax offset at the federal level. There are also 38 states that offer sizable R&D benefits as well. Some closely mirror the federal program, others are kind of on an application basis, but that could open up, you know, more than 10% savings back when you compare the federal and state. So I would encourage one kind of to get a health score on, on the project itself and the expenses, and then be looking at federal and state opportunities. It's funny because a lot of times when we talk to companies, it's like, oh yeah, I know about the R&D credit. And then you say, well, do you leverage it? I don't know. Or I did it my last job. You know, it's not, I'm not in accounting. So like, why do I care? But you should, right? And then, and then a lot of folks are like, no, I didn't hear about it. So they can actually be a huge hero in their organization. I mean, we see companies getting to tens of millions of revenue forgetting to do this. I just met a company last week that missed out. A Bay Area, like, tech firm that you would know the logo of that missed out of 12 years of R&D. Wow. That's painful.

Wow.

That is so, so painful. Now they can go back to claim, you know, the open years in statute as a C corporation. C corp actually can go back all the way if you've never made a dollar, but that's a huge headache and you will be reviewed.

Yeah.

Yeah. But for these startup companies, especially in their early days, being able to leverage the payroll offset is massive. And if you miss payroll offset, you can't go back to claim payroll offset. You can only amend to get credits for income tax offset. So if you don't have any liability on the backside, you can still go back, but you have to drag those forward and carry them forward until they can be utilized.

I read somewhere, and I won't attribute this to you because it was somebody else's study, but that two-thirds of businesses who are eligible for R&D tax credits are not applying, or not at all, that only one-third of eligible companies are actually leveraging this. Does that directionally feel right to you, or do you think that's completely made up?

Yeah, I mean, gosh, uh, I only have stats back to 2022 because that's how slow the IRS is in releasing information. Uh, but it does not feel off. Uh, typically kind of the, the message in the industry is that oftentimes 90% of startup companies miss out and about half of companies in general miss out on the incentive. And when you're talking about, you know, the last stats in 2022, there was close to $15 billion claimed, and we're talking there should be about $30 billion, right? So that's a huge pot of money that's unclaimed on an annual basis. And the nice thing about this program too is it's not, it's not a, there's only a certain amount and then it's gone. It's completely relative right now to your spend and your business. So it's not contingent upon, you know, what anyone else is getting. It's what you're getting. Now, states, they do differ. Sometimes there is a cap. Of how many dollars can be claimed. And, you know, they'll sell out just as fast as the Taylor Swift ticket. But in any event, you know, this could be a huge win for a manager that's trying to get buy-in, but also keeping budget attainable.

When it comes to deploying these solutions, especially when working with partners like us, is there anything else these businesses should know about once they, once they go live?

Yeah. So one area that folks folks get tripped up on is specifically around the outside contractors and where those folks are located. I always want to, you know, preface this that we know talent is global. Like, what a time in the world to be working with all different types of people across the US when it's, you know, more challenging than ever to recruit the best people for organizations, especially when you see these crazy taglines, right, where basically, Data engineers are getting stolen for like, for, you know, $100 million jersey. Like it's, they're making more than an MBA just so that the other guy doesn't have them, right? Nuts. Good for them. But all that is to say is we do see folks that they say, oh, well, like I pay a US agency, therefore, like, who knows where, where these folks are located? That's fine. That's allowed for your business. That might be great for your business., but you cannot claim the US R&D tax credit on those expenses. So it's very important if you're working with an outside provider that you know the lay of the land and, and who makes up the team. And if some of that's mixed, if it's all US, that makes it easy. If it's mixed, we really need to, to decipher. And I always get pushback on that from folks, you know, they're like, oh, I pay an agency in DC, but you know, that's where the money is. And we're like, no, it's where you're located. It'd be the same thing if you had an employee, you know, go work overseas.

Yeah, that makes sense.

Um, you pay taxes where the work gets done, right?

Not where the— Exactly. And, and yeah, you just can't claim the credit on them. You're getting, you're getting probably a better deal for them anyways. So, you know, make up the difference in other areas.

Well, fantastic. Ari, thank you so much for your time. I think for the listeners out there, I'd love to kind of consolidate a lot of the lessons you imparted here into a, into a checklist people can use if they're about to try to get their business case approved. Would love to get your eyes on that if you're open open to it.

Very happy to.

Yeah. Great. Great. So we'll add that in the show notes for everybody listening. And Ari, where else can people find you? What kind of companies should come to you for, for tax taker advice?

Yeah, I mean, any company that's working with Runpoint, right? Or, or thinking about AI. And I mean, honestly, we work with companies of all different sizes and industries. Makes our job really fun. So that's, you know, software businesses, that's life sciences, CPG companies, anyone that's robotics. Our day is never dull. So really, if you're building a product or a technology, we We would love to see if we can help you. And it's really easy to find us. It's just TaxTaker.com.

Perfect. Amazing. Thanks so much. Have a great day. Sam, anything, any last words?

Thank you, Ari. No, that was really, really helpful and wonderful. We appreciate your time.

Appreciate both of you for getting this message out there. So many companies are going to benefit because of it.

Thanks.

All right.

See you next time.