Riskified Ltd. (RSKD) Q4 FY2025 Earnings Call Transcript & Summary

Good day, and thank you for standing by. Welcome to the Riskified Fourth Quarter 2025 Earnings Call. [Operator Instructions] Please be advised that today's conference is being recorded. I would now like to turn the call over to your speaker today, Chett Mandel, Head of Investor Relations. Please go ahead.

Good morning, and thank you for joining us today. My name is Chett Mandel, Riskified's Head of Investor Relations. We released our results and are hosting today's call to discuss Riskified's financial results for the fourth quarter and full year 2025. Our earnings materials, including a replay of today's webcast will be available on our Investor Relations website at ir.riskified.com. Participating on today's call are Eido Gal, Riskified's Co-Founder and Chief Executive Officer; and Agi Dotcheva, Riskified's Chief Financial Officer. Certain statements made on the call today will be forward-looking statements related to without limitation, our operating performance, business and financial goals outlook as to revenues, gross profit margin, adjusted EBITDA profitability, adjusted EBITDA margins and expectations as to positive cash flows, which reflect management's best judgment based on currently available information and are not guarantees of future performance. We intend all forward-looking statements to be covered by the safe harbor provisions contained in the Private Securities Litigation Reform Act of 1995. These forward-looking statements reflect our expectations as of the date of this call, and except as required by law, we undertake no obligation to revise this information as a result of new developments that may occur after the time of this call. These forward-looking statements involve risks, uncertainties and other factors, some of which are beyond our control that could cause actual results to differ materially from our expectations. You should not put undue reliance on any forward-looking statement. Please refer to our annual report on Form 20-F for the year ended December 31, 2024, and subsequent reports we filed or furnished with the SEC for more information on the specific factors that could cause actual results to differ materially from our expectations. Additionally, we will discuss certain non-GAAP financial measures and key performance indicators on the call. Reconciliations to the most directly comparable GAAP financial measures are available in our earnings release issued earlier today and also furnished with the SEC on Form 6-K and in the appendix of our Investor Relations presentation, all of which are posted on our Investor Relations website. I will now turn the call over to Eido.

Thanks, Chett, and hello, everyone. We ended the year strong, and this momentum positions us for continued success in 2026. Our fourth quarter non-GAAP gross profit of $57.3 million represented strong year-over-year growth of 16% and our adjusted EBITDA of $17.7 million translated to a margin of 18%, demonstrating the scale and strength of the business. This quarterly amount alone exceeded our full year adjusted EBITDA of $17.2 million in '24. Our fourth quarter revenues of nearly $100 million were a record since inception and contributed to our first ever quarter of GAAP profitability. These results are the culmination of consistent high-quality execution across the year. In '25, both annual dollar retention, or ADR, and net dollar retention, or NDR, improved year-over-year. ADR reached approximately 100%, up from 96%, and NDR significantly improved to 105% from 96% in '24. Our go-to-market team had another successful year with particularly strong results in the fourth quarter of '25. During the quarter, we won the highest quarterly amount of new business since our IPO, which represented approximately 55% of the total new business won for the year and was driven by high competitive win rates of over 75%. This year, we won and onboarded several leaders across industries and geographies, including Aerolíneas Star Perú, Abounds, Adastria, ACE Hardware, Banxa, Kassam, David's Bridal, NetEase, Nintendo, Temu, TripAdvisor and XTO. In addition, merchants such as Iberia Airlines, Meta, Fast Retailing, Viva Aerobus, Vivid Seats and Zepz were all upsold in '25 after landing on the platform over the past few years. We believe this demonstrates the power and ROI that our platform delivers to our merchants once onboarded onto the Riskified network. We have processed approximately $750 billion in GMV and have over 1 billion unique customer interactions in our network since inception. I believe that this data moat has created a structural competitive advantage and that we are well positioned to capture even more of the large opportunity in front of us. That is why we are focusing our efforts on deepening our geographic presence and growing faster in our newer verticals while identifying additional verticals to penetrate for continued market share gains. From a geographic standpoint, our non-U.S. regions collectively grew 22% year-over-year, driving faster, more diversified growth. Notably, APAC and Lat Am were key regions of outperformance. We plan to expand further in these regions by developing localized products and features to boost pipeline generation. We have scaled our presence in the payments and money transfer category as evidenced by 66% growth in '24 and 90% growth in '25. Based on the current pipeline and the annualization of new business won in '25 in this vertical, I expect another strong year of activity in '26. And as we capture more data and payment types, leading to more refined models and bespoke features targeted to this vertical, I believe we are positioned to continue penetrating the significant white space. According to recent industry studies, there was a 27% year-over-year increase in fraud losses related to online transactions. The total losses attributed to fraud are expected to more than double over the next 5 years, well outpacing the expected growth of e-commerce. In addition, over 2/3 of U.S. companies experienced an increase in AI-related fraud attempts in '25. We are witnessing escalating complexity of fraud schemes, which now target every touch point across the customer journey from account creation and stored value credentials, all the way through the return, customer service and dispute portals. Every part of the transaction process is at risk. Fraud risk varies across payment types, including ACH, credit cards, digital wallets, crypto and stablecoins, agentic checkout and other methods. We need to be prepared to support and manage risk across the full payment landscape. We are leveraging the capabilities of our AI ecosystem that has been continuously advanced for over a decade. This increase in fraud has further elevated Riskified's role, solidifying our positioning as a key partner for our merchants. I believe that the combination of a more pronounced and complicated fraud landscape, enhanced platform features and functionality and a deliberate effort to expand the top of our fund deal funnel has contributed to an increase in our new business lead generation of approximately 50% year-over-year. Furthermore, in line with our expectations at the beginning of the year, I am pleased that we generated nearly $10 million in aggregate annual revenues from Policy Protect, Account Secure and Dispute Resolve in '25. And we plan to continue to grow our revenues outside of our core fraud services in '26. As we have expanded our offering, the benefits of having a robust platform are becoming even more pronounced. First, we saw an approximately 50% increase in the number of merchants who are now using more than one product during the year. This multiproduct approach has made us stickier. Second, each transaction processed across our suite of products strengthens our flywheel by expanding the breadth and depth of our data sets. This integrated data set compounds across the network, enhancing our identity engine and enabling us to develop dynamic components that can be utilized across the platform. Third, these cross-platform synergies lead to better performance for our merchants. The strong performance with differentiated capabilities allowed us to regularly outperform our competition. And fourth, merchants utilizing more than one product generally leads to higher contribution profit for those merchants. This is part of the reason why in '26, we are focused on driving gross profit growth versus optimizing primarily for revenue growth. As Agi will discuss shortly, we expect non-GAAP gross profit growth to accelerate to double digits at the midpoint in '26, demonstrating the continued leverage in our model. Now on to a very topical theme, artificial intelligence. Allow me to discuss how we are observing AI impacting the market and how Riskified's product platform and internal operations are positioned for success in this environment. There are 2 main dynamics that we are seeing. First is the increased utilization of agentic commerce through general purpose LLMs, but still primarily only for discovery purposes and not checkout. The second is the rise of merchant native LLMs, which are advanced agents within a merchant's ecosystem designed to handle the full shopping journey from answering queries to completing the purchase, closing the loop completely within their ecosystem. Both flows present unique transaction risks that our platform aims to solve. In the first agentic flow, when customers do use general purpose LLMs for checkout, we have seen instances where fraudsters utilize AI to throw agentic traffic off-script by generating synthetic IPs to bypass LLM verification. Our internal estimates indicate that approximately 30% to 40% of essential model features are lost when consumers transact through general purpose LLMs, increasing risk and escalating the prevalence of fraud like this. To combat this, we strive to help merchants by providing clear visibility into agentic traffic and emerging fraud MOs that they don't otherwise have on their own, proactively adjusting models based on low signal environments, segmenting order flows and rapidly developing features to identify emerging agentic fraud MOs. In the second flow, merchants are building out their AI shopping assistance to offer deep personalization and loyalty programs based on customer preferences. Riskified provides a critical risk intelligence layer that helps make these transactions both smart and secure. This is especially critical when those interactions have financial implications. An example of this is providing merchant native AI agents with real-time risk signals while they are in the conversation with customers to offer instant refund or exchange decisions based on that individual customer's risk and eligibility. Because Riskified analyzes the complete purchase history of the end customer across an expansive global network of e-commerce brands, including exact product list, SKUs and cross-merchant behaviors, we can provide highly differentiated data that merchants cannot otherwise access on their own. We are able to provide a decision platform for their agents to make important and accurate financial decisions. We are excited about the continuous expansion and enhancement of our agentic commerce offering. Merchants are actively preparing and ready to support agentic commerce across its various forms and flows. Our ability to not only service the dynamic needs of an evolving market, but also to innovate in real-time is generating an increase in merchant dialogue. I believe that this strategic engagement is a driver for our future business pipeline and growth. Internally, we continue to adopt AI to automate and scale complex business workflows across departments. This is intended to help drive operational efficiency and productivity, lower costs, improve response times and enhance service delivery. For our engineering teams, AI has become a force multiplier. Our developers have moved from basic coding assistance to agentic systems that span the entire development life cycle from discovery and requirements assessment to automated root cause analysis for production alerts. In addition, by using agentic flows for code review and observability, we are reducing technical debt while increasing release velocity. The impact on productivity is measurable. Between Q2 and Q4 of '25, many of our engineers saw more than 2x increase in tickets completed. This enables us to focus on developing new product enhancements and features and to test, train and deploy them more efficiently, strengthening our relationships with the hundreds of enterprise merchants in our network. We are seeing similar functional leverage across the other business units. In finance and analytics, we have moved several initiatives into production to automate processes that reduce human error and manual labor. And the go-to-market team has found success utilizing LLMs to drive merchant inbounds and high-intent queries. We've also developed agents that automate time-consuming cost/benefit analysis of merchant prospecting, minimizing manual work to drive quicker and more accurate outreach. And while we are getting leverage from general purpose LLMs in our own business, I don't believe that those same LLMs pose a true threat to our decision engine. In our view, LLMs lack calibration and the precise probability intervals required for fraud engines. Additionally, LLMs are optimized for text and image, while traditional AI fraud models like ours are much better at analyzing structured data inputs. The data we collect includes browsing behavior, account activity, checkout data and post-fulfillment signals for every transaction. Our models learn from over 5 billion historical nonpublic merchant network transactions that have been labeled intact. With this data, we create, update and continuously deploy features to be used by our models that solve the increasing complexities of fraud. To that end, as we announced yesterday, we have recently developed features to address this problem. Within our Policy Protect Decision Studio, merchants are able to identify and apply business rules to manage the risk of order volume coming from their native AI shopping agents. This control will allow merchants to confidently deploy their branded conversational AI agents without exposing themselves to programmatic refund claim abuse, reseller arbitrage or promotional abuse. We also expanded our AI agent identity signals, allowing a merchant's AI shopping agent to directly query Riskified's identity graph to retrieve associated risk indicators and resolve an identity programmatically. The breadth and sophistication of our platform allows us to train, test and deploy merchant or payment-specific models. We also use this platform to retrain models with updated data, new features and segment calibrations to protect from emerging fraud patterns across our network. All this helps us drive optimized merchant performance, which, at the end of the day, is the key driver of merchant satisfaction. Our ability to rapidly adapt in the face of a shifting landscape does more than just protect our merchants. I believe it serves as the foundation for our sustained financial strength and disciplined execution. Over the past 2 years, we have repurchased shares, representing approximately 2/3 of our current enterprise value. Based on our current expectations of improved free cash flow of approximately $40 million in '26, we anticipate generating a free cash flow yield of approximately 10% relative to our current enterprise value. Looking ahead, I believe that our momentum remains strong. As a reflection of our confidence in Riskified's long-term trajectory, I am pleased to announce that our Board has authorized an additional $75 million share repurchase program. This decision reflects our conviction in the fundamentals of the business supported by strong free cash flow, a debt-free balance sheet and a disciplined capital allocation strategy that we believe will prove beneficial for our shareholders. I want to thank our team again for their focus and strong execution against our '25 financial plan. Our results reflected the top of our revenue and adjusted EBITDA guidance ranges, and we entered '26 in a position to accelerate our performance even further. Now over to Agi.

Thank you, Eido, team and everyone for joining today's call. Unless otherwise noted, this discussion will reference non-GAAP financial measures. We have provided a reconciliation of GAAP to non-GAAP financial measures in our earnings release. We achieved fourth quarter revenue of $99.3 million and full year revenue of $344.6 million, up 6% and 5% year-over-year, respectively. And while we don't plan on reporting our billings going forward, our fourth quarter billings of $103.3 million grew 9% year-over-year. Our fourth quarter GMV of $46.7 billion was the highest quarter of volume reviewed in our history and represented growth of 18% as compared to the prior year period. For the full year of 2025, our GMV grew by 10% to $155.1 billion. During the fourth quarter, revenue growth was partially driven by strong performance in our travel subvertical, reflecting continued momentum from the third quarter. These gains were partially offset by softness in our tickets and live events subvertical, which declined year-over-year, primarily due to tougher second half comparable periods versus 2024's record level of activity and larger live events. Overall, the total tickets and travel vertical was slightly positive in the period. Our money transfer and payments category grew 75% year-over-year, driven by new business wins and upsell activity. Our fashion, cosmetics and luxury vertical grew 8% year-over-year. This was primarily driven by new business and upsell activity and 11% growth during the Black Friday through Cyber Monday period. This growth was partially offset by continued same-store sales pressure in our high-end and sneaker subverticals, similar to the first 9 months of the year. That being said, for the second quarter in a row, we did see year-over-year improvements in some of our largest merchants in this category. Lastly, I'm encouraged that we reverted to a year-over-year growth in the home category as we have now fully lapped the dynamic that impacted the first 9 months of 2025. For the year, our money transfer and payments, fashion and luxury, tickets and travel categories were the largest contributors to our annual revenue growth. The combination of these verticals represented nearly 80% of total billings and are each expected to drive continued growth in 2026. For the full year, revenue in the United States declined 6% year-over-year, primarily as a result of the contraction in our home category. Encouragingly, we continue to grow across all of our non-U.S. regions with accelerated year-over-year growth as compared to 2024. During 2025, APAC grew approximately 53% year-over-year, while other Americas, which represents Canada and Latin America, grew approximately 13% year-over-year, primarily driven by the momentum in new business and upsell activity with particular strength in the travel subvertical. EMEA grew approximately 18% year-over-year with the strongest performance concentrated in our money transfer and payments, tickets and travel and fashion and luxury verticals, supported by both new business and upsell momentum. Our revenue derived from merchants headquartered outside of the U.S. was 46% in 2025, up from 39% in 2024. We believe that our continued international growth reflects ongoing progress in capturing global market share. During the fourth quarter, we achieved record quarterly gross profit of $57.3 million, up 16% from the prior year and $180.3 million for the full year, representing a year-over-year growth of 4%. The full year gross profit growth of 4% was driven by meaningful improvements in our core machine learning models with great performance in our money transfer and payments category and within our 2024 cohort, which delivered the most pronounced year-over-year improvement across cohorts. Our increased revenue from new products further contributed to our growth. This improvement was partially offset by the ramping of merchants in newer geographies, such as Latin America and weaker performance in our 2022 cohort, which while still maturing, has yet to reach the performance levels of the broader portfolio. As a reminder, I encourage you to continue analyzing our gross profit on an annual basis given individual quarters can vary due to the various factors, including the ramping of new merchants and the risk profiles of transactions approved. As it relates to 2026 for the full year, we're targeting non-GAAP gross profit growth of 7% to 12%, with each quarter at or near 10% growth at the midpoint. In addition, we estimate that each quarter in 2026 will approximate the same percentage of the total as they did in 2025. Moving to our operating expenses. We continue to manage the business in a focused and disciplined manner. Total operating expenses were $39.6 million of the fourth quarter and $153.6 million for the full year, representing a decline of 2% from 2024. Our operating expenses as a percentage of revenue declined from 48% in 2024 to 45% in 2025, reflecting leverage in the business model. We ended 2025 with 617 global employees, a decline of 3% from the prior year. This was achieved through the increased utilization of artificial intelligence tools to maximize output and increase efficiency and by strategically reducing head count in areas that were less critical to our product development and growth strategy. Despite this nominal decline, we ended the year with an increase in our development capacity, which we believe is critical to advancing platform innovation, outperforming our competition and improving product accuracy and customer service to deepen our merchant relationships. In 2026, we anticipate quarterly expenses to approximate $41 million to $42 million per quarter in the first half of the year and $42 million to $43 million per quarter in the second half. The primary driver of the increase from 2025 relates to FX headwind, mainly from the appreciation of the Israeli shekel compared to the U.S. dollar. The FX headwind is approximately 400 basis points to our annual adjusted EBITDA margin. On a constant currency basis, we anticipate relatively flat expenses year-over-year as we continue to manage the business in a disciplined manner. We achieved adjusted EBITDA of $17.7 million in the fourth quarter, the highest quarterly amount in our history, which translates to an adjusted EBITDA margin of 18%. We believe that this quarter's results demonstrate that the business is positioned for continued adjusted EBITDA margin expansion and can achieve scaled performance like this over time. For the full year, our adjusted EBITDA was $26.7 million, representing a year-over-year increase of over 55%. On a GAAP basis, we achieved net profit of $5.8 million in the fourth quarter of 2025 as compared with negative $4.1 million in the prior year. I'm encouraged about the progress that we have made on achieving profitability on both GAAP and adjusted EBITDA basis. Moving to the balance sheet. We ended the year with approximately $298 million of cash, deposits and investments and continue to carry zero debt. In addition, we continue to maintain a healthy cash flow model. In the fourth quarter, we achieved free cash flows of $10.7 million and $33.1 million for the full year. Looking ahead, I'm encouraged that we expect our free cash flow to increase at least 20% and be approximately $40 million in 2026. During 2025, we repurchased approximately 22 million shares for a total price of $105.9 million, which contributed to a reduction of 8% in shares outstanding. Since the inception of our buyback program in the fourth quarter of 2023, we have repurchased approximately 52 million shares for a total price of $259.5 million, which helped contribute to a 17% reduction in shares outstanding over that time period. As Eido mentioned, I'm excited to announce that our Board of Directors has authorized up to an additional $75 million of share repurchases, subject to the satisfaction of Israeli regulatory requirements. When combined with amounts that remain available under our existing share repurchase authorization, our total outstanding authorization is approximately $84 million. We believe that our strong balance sheet and liquidity position are strategic assets that provide us with the flexibility to navigate a range of operating environments. We intend to remain disciplined and thoughtful in how we deploy capital to create long-term shareholder value. On the topic of share-based compensation and earnings per share. Share-based compensation expense of $51.6 million, declined from $57.8 million in the prior year. As a percentage of revenue, this amount decreased by approximately 300 basis points from 2024 levels. This was on top of a decline of 700 basis points over the prior 2 years. Looking ahead to 2026, we expect absolute share-based compensation dollars and as a percent of revenue to continue declining due to the gradual roll-off expense associated with large grants made in 2021 and 2022 as the awards fully vest throughout 2026. Our total absolute share-based compensation dollars should approximate $40 million for the year. We expect our free cash flow generation to approximate our share-based compensation in the year. Our annual non-GAAP diluted net profit per share of $0.20 represents an increase of 18% in 2025. Now turning to our outlook. As we look forward to 2026, we currently anticipate revenue of between $372 million and $384 million, representing growth of 8% to 11% within $378 million or 10% to the midpoint. Consistent with past years, we anticipate that our growth will continue to be driven primarily by new business activity. And at the midpoint of our guidance, we're forecasting a similar net dollar retention rate as in 2025. We currently expect all of the quarters in 2026 to reflect a similar percentage of the total revenue as they did in 2025, and growth to accelerate sequentially with each quarter throughout the year. The behavior of the macro environment, our success in retaining our merchants and the level of upsell activity relative to new logo wins only impact our net dollar retention rate and ultimately determine where we fall within our revenue range. In addition, we feel confident about the new business activity levels, which is supported by a robust pipeline of new opportunities. Historically, the timing of when new merchants go live during the year can be difficult to predict and may have an impact on our calendar year revenues. As always, we will continue to monitor the performance and health of our merchants, consumer spending and the broader e-commerce landscape and the impact on our results. Now let me discuss our adjusted EBITDA outlook. We currently expect adjusted EBITDA to be between $26 million and $34 million or $30 million to the midpoint, representing a margin of 8%. This is inclusive of an approximate 400 basis point FX headwind to our adjusted EBITDA margin. Overall, I'm encouraged by our AI advantaged market position, and I'm confident that we can continue to execute on the elements within our operational control. We remain focused on identifying and executing on the many opportunities for long-term growth and our ability to deliver value to our shareholders. Operator, we're ready to take the first question, please.

[Operator Instructions] Our first question comes from Terrell Tillman with Truist Securities.

Congrats Eido, Agi and Chett. The first question, and hopefully, you can bear with me because it's so topical around agentic commerce. It's a multi-partner, and then I'll have a follow-up question. As it relates to agentic, I appreciate kind of how you talked about 2 types of kind of agentic use cases. I'm curious if you can quantify any early GMV from those 2 different scenarios? And then also, what would the monetization or take rate look like in transactions in that type of flow? And then how many merchants are you actually actively working with that are just trying this out at this point? And then I had a follow-up.

Sure. Terry, so I'll take that. So maybe taking a step back, right, we feel we're an advantage -- in a great position to talk to over 50 publicly traded companies and really understand what their agentic commerce strategy is. And the way they're laying it out to us is pretty clearly, there are 2 main flows. The first flow, what we call the merchant native AI agents where they continue to own the relationship with the customer, and I think they have a lot of -- shows a lot of promise in their mind, right? And here, you would have an AI agent on their website that can support the entire life cycle from discovery to checkout, to returns and customer support interactions. And this entire experience is happening in their website. So if they're a luxury fashion merchant, they can have the right type of images and product descriptions and kind of flows and recommendations. If they're an OTA, they can have the right type of filters that are appropriate for kind of traveling and routing. So I think they're putting a lot of emphasis on that area. What we're seeing there is because LLMs are really -- it's easy to challenge them and get them to move off script and make financial decisions that you did not intend them to make, we're serving really as this guardrail or intelligence layer that they're querying in real time to understand, hey, should I approve this transaction? What type of refund should I provide to this customer? And we're just really leveraging the entire network that we already have, making it even more unique, the value that we're providing in there. So that's kind of the merchant native AI agent. The second flow is more kind of that general-purpose LLM, where it can either act as a good referral or do the purchasing on behalf of the consumer. There, to be clear, we're seeing predominantly referrals. Actually seeing agent traffic and purchasing is still extremely low and limited. From a take rate perspective, on the general-purpose LLMs, we are seeing higher risk traffic there right now. So again, even if it's very small, whenever you have some of these newer flows, fraud tends to come in because there's more limited data, there's lack of experience, there's less guardrails and controls in most cases. And so I think on average, the take rate there would probably be higher right now. But over time, that might kind of shift. And just a more general question around traffic, I think merchants are in a stage where they're trying to prepare and make sure that they're ready for the changes and put their best foot forward. But the traffic is probably not there yet.

Very helpful. And I guess just a follow-up, maybe for Agi. It's helpful when you go through the different segments that you're serving in the growth rates. Money transfer and payments, it was another exceptional year of growth as you're onboarding strategic accounts and they're growing. I'm curious, though, do you see that outsized type growth continuing in your guide for '26 on money transfer and payments versus the other end markets?

Terry, so money transfers and payments was an amazing category for us this year as the growth was really, really strong. Kind of looking into 2026, we have a number of opportunities in the pipeline. And I expect the category to continue to grow, but probably just to normalize in terms of like the total amount of growth.

Our next question comes from Ryan Tomasello with KBW.

Just following up on the agentic commerce topic. Can you talk about how you think about the potential for rising adoption there to either structurally reduce or increase the level of fraud in the system over the long run, notwithstanding kind of early days here? And then just your thoughts on the potential second order impacts. There's a lot of talk on agentic agents -- agentic AI agents utilizing alternative payments rails like stablecoins, just how you view that also impacting the structure of the system here?

Ryan, sure. Look, I think what we're seeing is that in order to be good at online commerce and payments specifically, you need to be doing a lot of things well. And right now, something that's added is agentic kind of commerce. And you can add crypto and stablecoin. So if historically, you would need to be able to manage credit cards and credit card acceptance, you now need to be able to support ACH and digital wallets and crypto and stable coins and agentic checkout. And with agentic, we're talking about a few different flows. And you don't just need to think about just the checkout, you also need to think about account creation and account login and you probably have some stored value in the account that people can transfer in and out. You obviously have the checkout experience, but you also have all these various post-checkout flows, returns, refunds, leveraging different discount codes and using that. You have the entire chargeback process, which is different between credit cards and ACH. It's not called the chargeback, there's insufficient funds and you have issues around scams that are popping up. So I think we're seeing an increase in complexity. And I would just tie in the agentic checkout into that overall increase in complexity, and we're seeing an overall increase in losses within the merchant ecosystem. And I think that as merchants are trying to solve these different use cases in these various fraud patterns, it just becomes more complicated, more quickly. So I think that's kind of a net benefit to Riskified as we see this more complex environment increasing in the years ahead. On a bit more targeted and specifically on agentic, like we just mentioned, we do see an increase in fraud right now in agentic channels, specifically when you have general-purpose LLMs. It could be a combination because it's newer and fraud tends to shift to that area and there's less control and saving there. So hard to say how that would kind of behave in kind of the quarters and years ahead as it gains more traction. But as of now, it's probably kind of net incremental to general take rates.

Great. Appreciate that. And then just an update, if you can provide on the mid-market expansion strategy, how that plays into your 2026 growth and just broader investment plans in that category?

Yes. I think one of the unique things about Riskified targeting the enterprises is that we're able to really customize to a high level, the modeling and the performance for each individual merchant. As we've been getting much better at completely automating the entire life cycle of doing that, I think that's going to present opportunities to kind of continue and refine this model in more of a down market and referral strategy. That's not something that's expected within our guide for the year. So the more we can accelerate that, that would be upside to current guidance.

Our next question comes from Will Nance with Goldman Sachs.

First of all, I hope all the teammates in Israel are at home and safe. I wanted to ask also on the agentic. It's kind of topic of the day. I was wondering if you could just speak to status on integrating into some of the latest agentic protocols, whether it's kind of ICP, Tribe, UCP, Google and any of the other relevant ones. I know a big part of the model is kind of taking all the different signals from the user behavior in those channels. So just maybe wondering if you could speak to that and maybe shed a little bit more light on kind of like the value of the data that might come through those protocols in detecting fraud vectors.

Yes. I think -- by the way, thank you for mentioning kind of the team in Israel. We appreciate that. I think the issue right now that the market is seeing is, to your point, there are a wide number of multiple protocols. And some of them, I think it's clear that they're already outdated in the months, maybe quarters ahead. There will be new protocols that are probably even kind of more updated that. So obviously, internally, we're doing everything we can to support everyone in that ecosystem, whether it's kind of AI Agent Approve, AWS Marketplace, Google's A2A protocol, just general RESTful APIs. We do see ourselves requiring to have that full spectrum to make sure we cover everything. Unfortunately, we do anticipate a somewhat continued fragmented approach here. So I think it's still early to say if there's anyone who's going to be a clear winner in that area. So there will probably need to be some kind of optimization between the various protocols.

Got it. That makes sense. And maybe just one for Agi. The FX headwind on the margins is helpful quantifying that. Could you just remind us of the -- it sounds like it's the FX exposure in the cost base that we should be thinking about there, shekel and otherwise. I was wondering if you could just kind of update us on major currency weightings as we try to fine-tune the model.

Yes, I will. So I mean, first of all, I'm so excited about the quarter, the guide, kind of the returning back to double-digit growth. And if I think about the FX headwind, we kind of spelled it out. That's approximately 400 basis points or $40 million to adjusted EBITDA. And it's frustrating. I mean, over the years, we kind of focused and we kind of ran on a flat expense base for a period of time. And this FX headwind is really obscuring some of the progress. But the truth is that the underlying business momentum is strong, and we'll continue to focus on optimizing. We'll continue to focus on growth, and I'm just excited about 2026.

Our next question comes from Cris Kennedy with William Blair.

I'll just echo Will's comment regarding Israel. The revenues from newer products, Policy Protect, Account Secure doubled in 2025. Can you talk about kind of the opportunity for that set of products in 2026?

Sure. So maybe just to refer back to kind of Ryan's question where we said, hey, we're seeing an increase in complexity of forms of fraud, we're seeing it across different channels like ACH, digital wallets, crypto stable coins, agentic checkout. And we're seeing it happen in different parts of the kind of shopping experience, not just checkout, but also account creation and use of policy rules and things around dispute management. All this to say, I think it's leading to an environment where there's kind of more demand and more value and just basically more necessity for merchants to leverage the wider product platform. So if I think about the revenue that we anticipate from kind of Policy, Account Secure, Dispute Resolve, some of the non-guaranteed payment flows that we now work with merchants on, anywhere from $15 million to $20 million in '26, I think, is a good range at this point.

Great. And then just one for Agi. If you think about the 2024 cohort, the CTB ratio really improved. Can you give us a little bit more color on what drove that improvement there?

Yes, of course. I'm very excited about some of the improvements of this cohort, and we kind of saw already the result of that in Q4. So there's kind of some merchants there that are specifically about the money transfer and payments category. And we were able to kind of do significantly better there as kind of evidenced in the cohort. And I think it's like a great base for some of the merchants that are in the pipeline there and just continuing to kind of optimize incoming merchants as well.

Our next question comes from Timothy Chiodo with UBS.

This one, we brought this up in the past, but I thought it would be a good one just to check in on to see if anything is different in the agentic channel. My guess is it's the same. But the question is really the services that you're providing to merchants, do you consider them and/or see them operating in addition to value-added services coming from the card networks or instead of value-added services coming from the card networks?

Tim, so sorry, could you rephrase our services in addition to the services from the card networks?

Sure. So if a merchant is working with Riskified, are they using Riskified in addition to some of the fraud-related value-added services coming from the card networks? Or are they using Riskified instead of some of the fraud tools that are coming from the card networks?

Okay. Thank you for clarifying. So look, I think there's no direct comparable in the stack of the card service providers right now to the spectrum of Riskified. One of them has more data-related features. So I think the Mastercard acquired Ekata. Visa probably has VisaVerify. So those are kind of what we consider data features. One of them has a more older generation scoring tool that we don't really view as competitive. No one has a policy product. Definitely no one has kind of what we would consider a modern machine learning type solution for fraud prevention. I think the dispute product, also there's nothing comparable. On the account side, there's nothing comparable. If you think about support for ACH, crypto stablecoin, kind of the fiat conversion and account storage, there's nothing comparable. On agentic checkout, there's definitely nothing that we've seen comparable to some of the releases we've recently made. So I think overall, it's -- on the Venn diagram, it's pretty distinct and different. There could be different services that they provide maybe more towards financial institutions, anything around tokenization and rails for really secure that plus in our wheelhouse. But hopefully, that gives kind of a good mapping of what we do that they don't do.

Excellent. That's a great answer. My follow-up is around -- you were talking around some of the other forms of payment, whether it be account-to-account, stablecoin and basically alternative payment methods in general. I know that it's early, but in your experience and with your position in the industry, do you have any reason to believe that the card mix within the agentic channel would be any different than the card mix is in traditional e-commerce? So whatever you believe that the mix is to be in traditional e-commerce, do you think through the agentic channel that it would be roughly the same, maybe the card mix is a little lower or maybe the card mix is a little higher? And what would be the reason that would lead you to believe the answer to the question?

Yes. Thank you. I think that's a great question and obviously, a lot of debate on that. I think there are specific industries and probably payments, remittance, kind of brokerages, which would probably see an increase over time on whether it's kind of stablecoins, cryptos or also direct ACH connections just because exchange fees, FX rates, everything that we know. So I think there, there probably is the potential for more to migrate maybe with long-term subscriptions. As things like ACH and others become easier, maybe merchants would have an easier time transferring some customers for kind of various discounts to that area. But overall, by and large, in most categories, I would not anticipate a shift. I think that overall consumer preference for cards, for rewards continues to be incredibly high and I think merchants adapt to that. I don't see that changing based on kind of the LLM channel or merchant native AI agents. I think merchants already have the ability to capture with extremely low interchange fees, debit cards. I think when you think about things like reward cards, the customer gets so much value from that. They have a clear preference. I think large merchants also have the ability to issue their own reward cards and take a meaningful portion of that interchange fee and usually through agreements with kind of networks or the issuers also take some kind of the potential float or value of kind of installments or late payments there. So from an ecosystem perspective, I think cards are still around to stay in most categories, but there's probably a few specific areas where we'll see an increased adoption in alternative payments. And I don't see a clear difference between kind of general purpose LLMs or merchant native AI that would make them specifically work on kind of stablecoins or anything else relative to the existing rails.

Our next question comes from Reggie Smith with JPMorgan.

Congrats on the quarter and achieving GAAP profitability. I guess I got another question about agentic as well. So I get and I appreciate that there's not a lot of transaction flow coming from, I guess, third-party LLMs today. And so it's early days, definitely get that. But I'm thinking about -- someone asked earlier about like how pricing may work there. I'm curious just like how that would roll out in general. And specifically, like would merchants need separate contracts for agentic? Or would it just be rolled into their standard e-commerce that occurs in our website? And then kind of beyond that, as you think about this new surface and these new potential risks, like how are you -- does that give you any pause at all or concern around like what early losses could be like and what kind of differentiates you there given that you won't have like a 13-year head start or back-testing history that you have on the traditional commerce side. So just curious like how you're thinking about that and like practically how this could actually roll out to your customers and yes.

Sure. Thanks, Reggie. That's a great question. So I think there are 2 ways this can go. One is with the client that's on various submission plans and not giving everything right now to Riskified. And usually, they would proactively come and say, "Hey, we're opening up this agentic channel or we're seeing some initial kind of traffic or maybe we're even reaching out to them." And then they say, "We would want you to manage this definitely because we're not prepared to do that." And we've seen some of the larger clients that we work with approach us proactively with that. We're also in contact directly with some of our other merchants. And the pricing there, it's just slightly more flexible pricing to start. I think merchants are very open to having a higher price initially, both because they understand there's an increased fraud in this day 1 and also because the absolute dollar amounts are still so small. It's less of an issue. And obviously, we would kind of better negotiate mutually the fees once we understand the actual risk profile and the volume there. So that's one instance. The other one is merchants that already exist and are providing all their volume to Riskified. Yes, it continues to be the case that we would just see this traffic. And based on the risk profile there, if there's a significant increase, we might need to have a discussion with the merchant what that means from a commercial perspective. As I think about how do we anticipate some of this fraud, on the one hand, you're right to say that it's still early stage and a single merchant might only see a single transaction. But by that same token, we're seeing it across the network of some of the largest merchants, and we're seeing some of the newer fraud MOs happen. And if you think about our system overall, what's unique and great about our system is that we're able to see fraud MOs in real time in one place and then adapt features or create new segments and deploy that relatively quickly to other parts in the model. So even though this is something that's kind of newer, our system really is adept at learning new fraud rings, new fraud MOs and pushing updates to the rest of the system based on that. It's what we've done is we've expanded into Lat Am into kind of other APAC regions. And you can continue to see that. I think Agi mentioned on some of the CTB cohorts, some of that continued quick improvements there. And agentic behaves the same, right? There's like new fraud trends, you need to stop bleeding there and then solve it for the rest of the portfolio.

Got it. Okay. And if I could ask one quick one on kind of FX. I appreciate that you guys are paid in dollars, but I was curious, is there any FX benefit to GMV growth next year? Or is that in U.S. dollars as well? FX is not much strong. So anything you could share there would be helpful.

Reggie, I'll take this one. So specifically, the way I kind of view the FX, the fluctuations over the years have been able that we were able to absorb specifically this year, where I see the FX impact and kind of isolated it in this 400 basis points, effect on adjusted EBITDA is around the strengthening of the Israeli currency, the shekel versus the dollar. And since half of our expenses approximately are in Israel, it's impacting it more materially. So that's the main kind of FX impact that I talked about and that it's worth mentioning. Without this, as I mentioned, on a constant currency basis, our expenses would have been flat year-over-year.

Got it. And so I guess just to put a finer point on it. Will there be a FX tailwind to revenue from the dollar just being weaker in general? Clearly, you've isolated the expense side, but I'm just curious, like is there anything we should think about at the revenue line?

On the revenue line, it's probably much minor. I'd imagine it's, if anything, probably from the euro, but that will be probably less than 0.5%, and it's something that's we've already incorporated in projections as kind of like we're basing our projections on what we see today.

Our next question comes from Clark Wright with D.A. Davidson.

Agi, on the beginning -- or I believe this actually might have been Eido, you spoke about the fact that your strategy is more oriented going forward on gross profit growth versus revenue growth. What does that mean from a go-to-market perspective and your risk tolerance for specific product categories?

Yes. Thanks for that question. Look, we've seen internally, I mean, we've always focused as a management team on gross profit, gross profit dollars, gross profit dollar growth, but it's probably been more of a focus recently over the past few quarters and will be over the next few quarters just because we're seeing more demand and more bundling strategies for the kind of wider product portfolio. And overall, there's a different margin profile within those products. So for us, it's clear we really want to focus on the gross profit dollars and that growth. From a sales perspective, anything from how they target accounts to how they think about -- how we think about commission structures is more oriented in this direction now.

Awesome. Appreciate that. And then just on another topic that was already discussed partially earlier, but I just wanted to understand the penetration rate on the nonchargeback guarantee products and the assumptions that you have for the 2026 guide. You referenced the $15 million to $20 million. But what does that mean in terms of the overall customer base and their willingness to accept or to adopt these offerings?

Yes. I think we shared on the script that we were seeing good progress of over...

Around 50%.

Around 50% kind of increase in adoption. We haven't really spelled it out by the specific product or what's dual product or what's triple product. We'll think about the best way to represent that to make it easier for investors to follow. But right now, we think that kind of revenue is probably the best proxy for that. And like we mentioned, went from -- I think it was really low single-digit millions to $10 million, and we think we can continue to grow that to $15 million to $20 million this year.

And I'm not showing any further questions at this time. I'd like to turn the call back over to Eido for any further remarks.

Thank you. Just before I conclude, I want to send my support to our team members in Israel and their families. Thank everyone for their hard work. And with that, just thank you, everyone, for joining us on today's call. I look forward to continuing to update you on our progress throughout the year.

Thank you. Ladies and gentlemen, this does conclude today's presentation. You may now disconnect, and have a wonderful day.