EXPO - NASDAQ

Good day, and welcome to the Exponent, Inc. Fourth Quarter and Fiscal Year 2025 Earnings Conference Call. [Operator Instructions] Please note that this event is being recorded. I would now like to turn the conference over to Joni Konstantelos, Managing Director at Riveron. Please go ahead.

Thank you, operator. Good afternoon, ladies and gentlemen. Thank you for joining us on Exponent's Fourth Quarter and Fiscal Year 2025 Financial Results Conference Call. Please note that this call will be simultaneously webcast on the Investor Relations section of the company's corporate website at www.exponent.com. This conference call is the property of Exponent, and any taping or other reproduction is expressly prohibited without prior written consent. Joining me on the call today are Dr. Catherine Corrigan, President and Chief Executive Officer; and Rich Schlenker, Executive Vice President and Chief Financial Officer. Before we start, I would like to remind you that the following discussion contains forward-looking statements, including, but not limited to, Exponent's market opportunities and future financial results that involve risks and uncertainties that may cause actual results to differ materially from those discussed here. Additional information that could cause actual results to differ from forward-looking statements can be found in Exponent's periodic SEC filings, including those factors discussed under the caption Risk Factors in Exponent's most recent Form 10-Q. The forward-looking statements and risks in this conference call are based on current expectations as of today, and Exponent assumes no obligation to update or revise them, whether as a result of new developments or otherwise. And now I will turn the call over to Dr. Catherine Corrigan, Chief Executive Officer. Catherine?

Thank you, Joni, and thank you, everyone, for joining us today. I will start off by reviewing our fourth quarter and fiscal year 2025 business performance. Rich will then provide a more detailed review of our financial results and outlook for 2026, and we will then open the call for questions. We delivered a strong finish to 2025, reflecting the strength, diversification and resilience of our portfolio. During the fourth quarter, we saw growth in proactive engagements, driven by increased demand for user research in consumer electronics, along with continued expansion of our risk management work in the utility sector. Growth in our reactive services was driven by failure analysis and dispute-related engagements across a broad range of industries, including energy, construction, transportation and life sciences. Turning to our engagements in more detail. Growth in proactive engagements in the fourth quarter reflected continued diversification across a broader mix of clients and an expanding range of products and technologies. In consumer electronics, we saw increased demand for user research engagements, driven by the need to evaluate product performance and user interaction as artificial intelligence becomes increasingly embedded in both every day and novel devices. We also saw continued growth in risk management and asset integrity services for utilities, supported by rising energy demand and increased focus on grid reliability. In Life Sciences, engagements increased across regulatory compliance, product performance and safety consulting for medical devices as these safety critical technologies continue to become more complex. Turning to our reactive engagements. Demand for Exponent's failure analysis and dispute-related services drove growth in the fourth quarter, reflecting the essential role our engineers and scientists play when systems do not perform as expected. In transportation, we saw increased failure analysis work tied to electrification and battery systems in commercial vehicles as customers address performance, safety and reliability challenges. We expanded our failure investigation work in data center infrastructure, for example, addressing board-level cooling and thermal management issues where multidisciplinary teams are required to determine the root cause of failure. Across the energy sector, we continue to see robust demand in dispute-related engagements spanning hydroelectric facilities, wildfire-related losses, battery energy storage systems and wind and solar projects. Exponent continues to benefit from powerful long-term market drivers. As artificial intelligence and other complex technologies are increasingly incorporated into novel products, infrastructure and safety-critical systems, demand is growing for our science and engineering expertise to support and enhance algorithm performance. Sensor-based systems that demand the highest level of trust are also frequently found in the most challenging, disrupted or intermittent connectivity environments, creating settings where security and safety are inseparable. While AI delivers value by learning and predicting based on historical data, many of the most consequential challenges arise in physical systems where edge cases, novel conditions and complex interactions fall outside of prior experience. Exponent thrives at the edge where AI meets the laws of physics in high-stakes environments where reliability, performance and security cannot be compromised. These dynamics underpin sustained long-term demand for Exponent's multidisciplinary expertise. Our teams apply deep capabilities in engineering, physics, biology, chemistry, material science, cybersecurity, human behavior and more to help clients validate and enhance system performance, identify risk, ensure security at the asset level and apply scientific judgment where complexity and uncertainty exceed the limits of algorithms alone. As AI-enabled systems are deployed more broadly, failures, whether at the algorithm or the physical system level, are becoming more complex, more difficult to diagnose and more consequential. Determining the root cause of these failures demands rigorous investigation that integrates physical sciences, engineering, data science and human factors to reconstruct real-world conditions and system behavior. Exponent's long-standing failure analysis expertise uniquely positions us to support clients as they navigate these situations, delivering independent science-based insight that informs remediation, accountability and innovation. As artificial intelligence and other complex technologies increasingly intersect with performance and safety critical applications, this capability remains a core and differentiating component of the long-term value that we provide. At the same time, we are leveraging artificial intelligence within our operations to add value and support our teams as demand for our expertise continues to grow. These tools enable our experts to work more effectively, focused on the highest value aspects of their work and deploy their capabilities where they matter most. Looking ahead, Exponent continues to benefit from powerful long-term market drivers, including increasing complexity, rapid technological innovation and rising expectations around safety, health and the environment. As artificial intelligence and other advanced technologies become more deeply embedded in novel products and critical systems, clients are facing an expanding set of complex high stakes challenges. This environment is driving increasing demand for independent multidisciplinary expertise and is supporting continued diversifications across technologies, products and clients as reflected in our results. Together, these dynamics position Exponent to deliver rigorous science-based insights across the full product life cycle and support long-term growth. I'll now turn the call over to Rich to provide more detail on our fourth quarter and fiscal year 2025 results as well as discuss our outlook for the first quarter and the full year 2026.

Thank you, Catherine, and good afternoon, everyone. Let me start by saying all comparisons will be on a year-over-year basis unless otherwise noted. I would like to remind everyone that we returned to a 13-week fourth quarter and a 52-week fiscal year in 2025 compared to a fiscal year 2024, which included an extra week that occurs every fifth or sixth year. The extra week poses a headwind to revenues of approximately 7% in the fourth quarter and 1.3% to the year. For the fourth quarter 2025, total revenues increased 8% to $147.4 million and revenues before reimbursements or net revenues, as I will refer to them from here on, increased 5% to $129.4 million as compared to the same period in 2024. So if you adjust for the 1 week -- 1 less week, net revenues would have grown in the low double digits. Net income for the fourth quarter was $24.8 million or $0.49 per diluted share as compared to $23.6 million or $0.46 per diluted share in the prior year period. The realized tax benefit associated with accounting for share-based awards in the fourth quarter was $99,000 as compared to $591,000 in the fourth quarter of 2024. Inclusive of the tax benefit for share-based awards, Exponent's consolidated tax rate was 27.4% in the fourth quarter as compared to 24.7% for the same period in 2024. EBITDA for the quarter was $34.7 million, producing a margin of 26.8% of net revenues as compared to $31.2 million or 25.2% of net revenues in the same period of 2024. Billable hours in the fourth quarter were approximately 357,000, a decrease of 1% year-over-year. If you adjust for the 1 less week, billable hours would have been up approximately 6%. The average number of technical full-time equivalent employees in the fourth quarter was 992, which is an increase of 5% as compared to 1 year ago. This increase was due to our recruiting and retention efforts. Utilization in the fourth quarter was 69%, up from 68% in the same period of 2024. The realized rate increase was approximately 5% for the fourth quarter as compared to the same period a year ago. This is a result of our premium position in the marketplace, unparalleled talent and differentiated interdisciplinary expertise. In the fourth quarter, compensation expense after adjusting for gains and losses and deferred compensation was approximately flat. Included in total compensation expense is a gain in deferred compensation of $2.7 million as compared to a gain of $629,000 in the same period of 2024. As a reminder, gains and losses and deferred compensation are offset in miscellaneous income and have no impact on the bottom line. Stock-based compensation expense in the fourth quarter was $5 million as compared to $4.9 million in the prior year period. Other operating expenses in the fourth quarter were up 1% to $12.6 million. Included in other operating expenses is depreciation and amortization expense of $2.5 million. G&A expenses increased 17% to $6.7 million for the fourth quarter due to an increase in travel and meals associated with business development, professional development and increased recruiting activity. Interest income decreased to $1.9 million for the fourth quarter, driven by a decrease in cash and lower interest rates. Miscellaneous income, excluding deferred compensation gain was approximately $296,000 for the fourth quarter. During the quarter, capital expenditures were $2.7 million. We distributed $14.9 million to shareholders through dividend payments and repurchased $25.1 million of common stock at an average price of $70.57. Turning to the full year results. Total revenues increased -- total revenues and net revenues grew 4% to $582 million and $536.8 million, respectively, as compared to 2024. Net income for the year decreased 3% to $106 million or $2.07 per diluted share as compared to $109 million or $2.11 per diluted share in 2024. During the year, we realized a negative tax impact associated with accounting for share-based awards of $255,000 as compared to a tax benefit of $2.8 million in 2024. Inclusive of the tax benefit for share-based awards, Exponent's consolidated tax rate was 28% for the full year as compared to 26% in 2024. For the year, EBITDA increased to $148.1 million as compared to $147.1 million during the prior year producing a margin of 27.6% of net revenues, which is a decrease of 80 basis points as compared to 2024. This year-over-year decrease in margins was expected primarily due to the costs associated with our managers meeting during 2025 and the renewal of our Phoenix land lease in June of 2024. Billable hours for 2025 were approximately $1,468,000, a 2% decrease year-over-year. Utilization for the full year was 72.5%, down from 72.9% in the same period of 2024. Average technical full-time equivalent employees for the year were 973, an increase of 1% as compared to 2024. The realized rate increase was approximately 5% for the year. Compensation expense after adjusting for gains and losses in deferred compensation increased 3%. Included in total compensation expense is a gain in deferred compensation of $17.4 million as compared to a gain of $14.9 million during 2024. Stock-based compensation expense in 2025 was $23.8 million as compared to $23.2 million in the prior year. Other operating expenses were up 7% to $49.5 million, driven primarily by an increased noncash expense of our Phoenix lease renewal. Included in other operating expenses is depreciation and amortization expense of $10.1 million. G&A were up 12% to $25.5 million in 2025. The increase in G&A expenses was primarily due to an increase in travel and meals related to our in-person managers meeting in September, which was postponed in 2024. Interest income decreased approximately $694,000 to $9.3 million for the full year. Lower interest income was driven by a decrease in cash and lower interest rates. Miscellaneous income, excluding the deferred compensation, was approximately $840,000 in 2025. Moving to our cash flows. During 2025, we generated $131.7 million from operations and capital expenditures were $9.4 million. For the full year, we distributed $61.5 million to shareholders through dividend payments and repurchased $97.8 million of common stock at an average price of $72.22. As of year-end, the company had $221.9 million in cash and cash equivalents. Turning to our segments. Exponent's Engineering and other scientific segment represented 85% of net revenues during the fourth quarter and 84% for the year 2025. Net revenues in this segment increased 7% for the fourth quarter and 4% for the full year, driven by proactive services, including risk management work for the utility industry as clients addressed energy infrastructure challenges stemming from rising power demands and extreme weather events, regulatory support services for medical device clients and user research services for clients in the consumer electronics industry. Growth during the quarter was also driven by disputes-related services for the construction, energy and transportation industries as clients rely on Exponent in critical high-stake situations. Exponent's environmental & health segment represented 15% of net revenues during the fourth quarter and 16% of net revenues during fiscal year 2025. Revenues before reimbursements in this segment decreased 5% for the fourth quarter and were approximately flat for the full year. The decline during the fourth quarter was primarily due to having 1 less week during the fourth quarter of fiscal year 2025 as compared to 2024. Turning to the outlook for the first quarter and full year 2026. We expect net revenues for the first quarter and full year 2026 to grow in the high single digits as compared to the same periods in 2025. For the first quarter of 2026, we expect EBITDA margin to be 27.5% to 28.5% of net revenues as compared to 27.3% in the first quarter of 2025. For fiscal year 2026, we expect EBITDA margin to be 27.6% to 28.1% of net revenues as compared to 27.6% in 2025. We expect increased demand and corresponding recruiting to result in our average technical full-time equivalent employees increasing approximately 4% year-over-year in the first quarter of 2026 and 4% to 5% for the full year 2026 as compared to 2025. We expect utilization in the first quarter to be 75% to 76% as compared to 75% in the same quarter in the prior year. And we expect the full year utilization to be 72.5% to 73% as compared to 72.5% in 2025. We still believe our long-term target of sustained mid-70s utilization is achievable as we continue to strategically manage headcount and balance utilization with market demand. We expect the realized rate increase for the first quarter to be 3.5% to 4% and for the full year to be 3% to 3.5%. The lower rate realization for the year is based on a historical trend as hiring rates increase. For the first quarter, we expect stock-based compensation to be $8.6 million to $9 million and each of the remaining quarters to be $5.5 million to $6.3 million. For the full year 2026, we expect stock-based compensation to be $26 million to $26.5 million. We continue to believe that our stock-based compensation program effectively attracts, motivates and retains our top talent. For the first quarter, we expect other operating expenses to be $12.7 million to $13.2 million. For the full year, we expect other operating expenses to be $53.5 million to $54 million. For the first quarter, we expect G&A expenses to be $5.4 million to $5.8 million. For the full year 2026, we expect G&A expenses to be $27.1 million to $28.1 million. We expect interest income to be $1.7 million to $1.9 million per quarter in 2026. In addition, we anticipate miscellaneous income to be approximately $300,000 per quarter in 2026 or $1.2 million for the full year as compared to $840,000 in 2025. We expect our first quarter 2026 tax rate to be approximately 30.4% as compared to 29.4% in the same quarter a year ago. For the full year 2026, the tax rate is expected to be 28.5% as compared to 27.9% in 2025. Capital expenditures for the full year 2026 are expected to be $12 million to $14 million. We remain encouraged by the opportunities across our markets and believe we are well positioned to drive improved growth in 2026 while executing against our long-term financial objectives of high single-digit to low double-digit organic growth and margin expansion. I will now turn the call back to Catherine for closing remarks.

Thank you, Rich. Looking ahead, we remain encouraged by the enduring market drivers that support Exponent's long-term opportunities. As the pace of innovation continues to accelerate and systems become more complex, expectations for safety, reliability and performance will only continue to rise. With a differentiated multidisciplinary platform and a proven ability to support clients across both proactive and reactive engagements, Exponent is well positioned to navigate these trends and deliver sustainable growth and long-term value for our shareholders. Operator, we are now ready for questions.

[Operator Instructions] The first question comes from Andrew Nicholas with William Blair.

I guess, first, I was hoping you could hone in a little bit more on the consumer electronics piece of your proactive business. That was something that has been a little bit more challenged the past couple of years. I know last quarter, you spoke to some early signs of improvement there. So any additional commentary on how that business performed in the quarter and maybe what the near-term outlook looks like for that business in particular?

Yes. Thanks, Andrew. That particular part of the business is really primarily two-pronged. We've got a kind of a hardware product development consulting piece of that. And then we've got a user research-oriented piece of that, where we do work around human subject, human interaction with novel devices. And so one of the things we're really seeing is an uptick, particularly on the user research side. A number of these applications and engagements relate to health-related products, for example. They also relate to products that are very novel where artificial intelligence is being delivered via novel form factor. So -- you can think of traditional screen-oriented devices, or you can think of things like glasses or headsets or even things that use primarily audio instead of having a screen or using a visual input. So both the health side as well as the kind of consumer product side is a lot of what was driving that. There's diversification in the product base, and there's also diversification across the client base as more and more -- there are more and more entrants into this arena of trying to deliver artificial intelligence via these novel hardware platforms.

Very helpful. And then maybe a question for Rich on the guidance specifically. I think this quarter, second straight quarter of effectively like double-digit growth if you adjust for the extra week, last year, it looks like your outlook for utilization in the first quarter is as high as it's been, I think, in some time. So just curious on overall visibility and the achievability of guidance, how you think about some conservatism in there to the extent there is any and maybe areas of upside or downside to the outlook?

Yes. So our business, I think what we have good visibility into is these broader market demands and trends that Catherine has talked about in her comment. And I think we are actually seeing real work come in that are related to AI and novel technologies and continuing to see that the complexity of these issues is increasing. As we've said before, I mean, we go out to our business units all the way down to the individuals. And as we're getting forecasts, I think our people have good visibility out over 6, 8 weeks, a little bit lighter after that. But the trends of what we're seeing are positive. As we enter 2026, the reason that we've landed on our guidance that we have here of high single-digit growth is really as we entered last year, we had good headcount growth. We had 2% sequential in the first quarter of last year, which is very strong. It came down a little bit in the second quarter, and then we closed out the year strong. But we're feeling good about really where we can be in the headcount. We're feeling that, that demand is there. That's why we said the utilization will be slightly better than it was a year ago. But all those things combined landed us into that range that we have. Is there opportunity for upside? Yes, I think the demand environment is strong out there. But we -- at this time, this is the best estimate that we have, and we're delivering that with good growth and margin improvement, and we'll take it from there.

The next question comes from Tomo Sano with JPMorgan.

From management perspective, how would you characterize 2026 compared to 2025? And especially, what do you see as the most significant changes or drivers for revenue growth and margin improvement internally and externally, please?

Yes. Thanks for that, Tom. Clearly, we are seeing on a year-over-year basis, some acceleration of growth and of the demand environment that we have across a broad swath of the business. I think that the consumer electronics arena is an important one to call out in this regard. We saw strength in the fourth quarter. We do have a good -- pretty good outlook into Q1 that is helping to drive that. It gets a little less clear after that. But again, with the diversification across the products, across clients and form factors and things, we do expect in that electronics arena, both for user research as well as the hardware side to be part of those drivers, especially as AI is being delivered and making decisions in safety-critical applications like health-related wearables, regulated medical devices and things like that. We also see the energy side as a really important driver for 2026. And this was happening in 2025 to some extent, but we believe can continue to strengthen. This is around utility-related work. We mentioned the risk management work that we're doing. That continues to grow and diversify across clients. The regulatory environment continues to grow. The bar continues to go up in terms of that with relationship to grid resilience to extreme weather and things of that nature. We're seeing it on the reactive side in energy, too, as the demand for power is driving the need for new technologies to be utilized in a lot of these capital projects, whether that's wind, whether that's solar, whether that's fuel cells. You've got data center operators building their own gas-powered plants. You have multiyear long waiting list for gas turbines. And so the risk issues and the disputes that arise in the building out of those energy systems are a piece of this as well. And the data center piece, we're doing more and more failure analysis type work, whether it's around the cooling systems, whether it's around the backup battery supply systems, the performance-critical nature of those data centers really means that they need some powerful multidisciplinary expertise to diagnose some of those issues. You can go over to the chemical side of the business, things like PFAS and its effects on human health and the environment are another area where we expect to continue to increase -- increasing demand as the year goes on. So that's a few examples. I think electrification and automation and transportation maybe kind of round out that collection of things that we see in 2026.

And follow-up on AI. You already touched in the prepared remarks, but I wanted to get your thoughts, especially potential risks of commoditization in certain litigation support or investigation services due to automations. But also you talked, I think, is the significant opportunity to leverage AI for new value-added offerings and margin improvement. Could you talk about that more specifically about the litigation support or investigation services, the space, please?

Yes, absolutely. So there are a number of tools clearly with large language models that we have been incorporating into our operations that are allowing our teams to engage with larger and larger data sets in an even more efficient sort of manner. Being able to have an AI application pull the data out of a police report, let's say, if you're reconstructing a vehicle accident, these are the types of things that can be further automated, and we're seeing more efficiency in that regard and really welcome that. But what we're also seeing is, as you alluded to in your question, the higher value coming out at the other end. The ability to put a large language model application against an increasingly large data set of complex material, which is what we've seen happen over time. When I first started doing litigation work a couple of decades ago, you could fit everything in a black 3-ring binder that was a couple of inches thick. And now 20 or 25 years later, you've got gigabytes and terabytes of data. If you think about that vehicle that's in that accident, the data coming off of all of those sensors create a very complex and large data environment that needs to be analyzed, right? So while we're gaining efficiencies at that sort of lower level, we're also unlocking the ability to differentiate ourselves even further because of the complexity and our ability with our PhD level talent to be able to break that down and understand in a hypothetical situation, if the design were changed, would the product have performed better. So, so far, our reactive business continues to grow. The litigation support piece of the business continues to grow. Automotive is the place where we're seeing the most directly AI relevant work in our reactive business and the complexity there with the testing and those sorts of systems is continuing to grow. So -- and with our population of PhD entry-level talent, this is different than all of our competitors. Many of our competitors have lower-level talent. They've invested perhaps in those lower-level commodity tasks as an important part of their value proposition. That hasn't been the direction that Exponent has taken. That's why we hire PhDs as our entry-level folks, people who know how to solve that unstructured problem, that edge case. So I really do believe that the use of these sorts of tools will make us more efficient, and it will unlock even greater value.

The next question comes from Tobey Sommer with Truist.

What are your expectations for net headcount growth in '26? And could you maybe highlight the areas where you expect to add the most and any areas that you may expect to have fewer heads throughout the year?

Yes. So our expectation is that in line with that guidance, we would expect that the headcount growth would be somewhere in the net 40 to 50 growth in what we're doing. You're going to acquire those over the year that we'd be in that range. It could get up as high as 60, but it's somewhere in that range. Look, the areas of focus and the areas that are getting -- seeing the most net growth are really in these growth areas that Catherine highlighted earlier. Every one of our practices is actually recruiting and bringing people in just as part of our natural part of a consulting firm, we do have turnover that occurs. And as such, we're always looking to bring in new talent, those PhDs that have just done their once never solved before issue that they did their PhD thesis in and integrating them into every single one of our practices every year. But the areas that we'll see the growth are in that higher growth will be in that transportation area, the energy area, battery storage, automation, cybersecurity and actually into that chemicals area that Catherine mentioned around PFAS.

And associated with that headcount, that pace of headcount growth, is it so much so as to have accompanying margin -- negative margin implications? Or since you revived growth in the not-too-distant past, is that behind us and not necessarily reflective of any requisite margin compression?

Yes. Our expectation is that we are going to have margins be flat or up, and that is because we expect to be able to do this level of hiring into the organization based on demand while seeing our utilization be maintained or improved in 2026.

Appreciate that. If I could, I appreciated your prepared remarks, Catherine, on AI with the discussion there. So clearly, it's topical. I want to just ask another simple question. Near term and recent actual results, do you think AI is a net benefit or drag to the total company's growth?

Yes. I think it is a net benefit. If you think about the failure analysis work around advanced driver assistance technologies and automated vehicles, that's directly driven by artificial intelligence making safety critical decisions. The work that I highlighted early in the Q&A around the user research in the electronics industry, this is all about the data collection and benchmarking and validation for devices that are utilizing AI algorithms to make some kind of decision or have some kind of signal, whether it's to tell you your heart isn't beating properly or lots of other -- what your blood pressure is or so forth. And the same on the hardware side and the data center side, right? AI is directly driving those increases in energy demand, which we believe is part of what's driving our energy sector in its growth and especially on the dispute side, but also on the proactive side with the risk management work in utilities. So it's not directly in the project perhaps, but it's a fundamental driver that need for energy in the setting of a crumbling infrastructure.

I mean just on that area, again, we're seeing it drive through if it is thermal management at the board level or it's heavy -- the change in demands that the amount of infrastructure on these racks have, they're all things that fit into Exponent's expertise that will -- that we are getting business related to. So we're not seeing a change in demand for the amount of time or something that we're putting into, again, processing data or doing it. The datasets are growing so much that the clients just want to do more and learn more from it, and it's harder to understand why something made a decision that they are chasing. So that's what we've been seeing over 2025 and continued into the fourth quarter.

And I have one follow-up based on that. At what point in recent history do you think AI started becoming a net contributor to growth? And I might be asking an impossible second part of this. But of the low double-digit year-over-year growth in the quarter ex the extra week, is there a way to get a sense for how impactful AI factors are in that year-over-year growth?

Well, I think it's important to recognize that Exponent has been talking and working on systems that were leveraging the early parts of AI and machine learning and what we've been doing over the last decade or so when we were working with the automotive industry and its early days of steering control or braking or whatever may be coming into it to where we are today, we're all seeing the actual robotaxis on the road in doing it. So it's definitely growing. It is driving growth in our transportation area, and we expect that to continue going forward. The same goes around user research. Back in the day, we had clients that were really trying to understand how to make sure that they were developing inclusive products around facial recognition and technology like that and then driving that into other technologies that is where we got into doing user research and such and then much more into its performance in other health or other applications. But again, been at that for nearly 10 years that we've been doing it, but it's been growing. Same goes around what I would say in the utilities industry, is we've been developing in the risk model area, what's actually playing into why is that somewhat AI related now of our risk is actually that our clients have choices. And some of them have chosen a, let's say, at times, a less expensive AI type model that isn't giving them a refined or accurate enough answer for them to rely upon or justify the actions they took or did not take in situations and are now asking our help in refining those and bringing physics and bringing that higher level of engineering so that we can move AI models to a level of reliability that can be in safety-critical environment. So all those things are going on today, I think we've got a long way, a lot of upward ramp to go. Probably today, somewhere in the mid-teens as a percentage of our business is related to AI, probably either directly or one step removed, not saying all energy stuff or any of that, but really things that we can target in that close -- immediate or near vicinity that relates to that.

The next question comes from Josh Chan with UBS.

I guess following up on Rich's comment just now, I guess, have you seen any evidence of clients potentially trying to use AI themselves to solve problems? I know in some situations, it's completely impossible, but have you seen any evidence of that kind of occurring at your customer base?

Certainly, our -- look, I mean, our customers are absolutely looking to incorporate AI into their operations. There are situations, I mean, Rich just mentioned one where they've -- on the utility side, incorporated those into risk models and have found that AI alone simply is not good enough and so -- but we see that in the medical device environment with sort of software as a medical device. There are -- we get pulled in on the regulatory side of that, right, where they're trying to get their -- develop their plan of attack for getting through the FDA in terms of approval on that. So, yes, I mean, our electronics clients are putting artificial intelligence into all kinds of different form factor devices, and they're asking us for our help in benchmarking and user research. So it's really everywhere we turn. There are different levels of confidence that clients have in it. Some are more sort of skeptical. Others want to dive right in, but they're coming to us for advice and sort of reality checking, if you will, in a lot of these applications.

Okay. And then maybe just a quick follow-up on next year on 2026. Is there anything different about how free cash flow will work in '26 than it worked in '25? Anything to kind of flag there? Or is that a pretty normal conversion?

Our expectation is that we may be able to improve upon that conversion. The area -- we had sort of a heavy amount of reimbursables there at the end of the year tied in with the studies, which made DSOs a little bit higher than the average of where we would want to end the year. So I would expect that, that would naturally and through some efforts we're making come into more balance, so they might be able to bring down DSOs by a few days, and that will help us move to a steady state and improve -- area to improve cash flows going forward.