Triumph Financial, Inc.

The Intelligence segment was launched at the beginning of the fourth quarter of 2024 to turn the over-the-road trucking data collected through our services into actionable insights for our customers. This launch coincided with our acquisition of the assets of Isometric Technologies Inc. that provides service and performance scoring and benchmarking capabilities to the over-the-road trucking industry. The operations of this segment were further supplemented with our acquisition of Greenscreens AI. Inc., a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights, during the quarter ended June 30, 2025. Intelligence offerings enable better decision making, market intelligence and automation. The revenue for Intelligence offerings is derived through access and subscription fees, as well as seat licenses where applicable.

2.Automation and AI Update

Automation and AI Update. Our instant decision platform continues to drive meaningful operating leverage. During the quarter, we purchased approximately 1.7 million invoices supported by 235 average full‑time employees, representing a 12.4% increase in the number of invoices purchased by 11.4% fewer FTEs compared to the same quarter last year. This improvement reflects the scalability of our operating model and our ability to support higher transaction volumes without proportional increases in headcount. These gains are being reinforced by broader efficiency gains across Factoring. We are simplifying workflows, reducing handoffs, and eliminating non‑value‑added activities which results in measurable improvements in both capacity and operating margin.

In parallel, we are advancing a focused set of AI‑ and LLM‑enabled automation initiatives alongside pragmatic process simplification across decisioning, cash application, and collections. While several initiatives remain in progress, early results are translating into improved throughput and workflow efficiency, supporting continued increases in invoices processed per employee. These efforts are strengthening the scalability of our operating model, enabling us to support growth without proportional cost expansion, with the benefits reflected in our expense outlook. As these initiatives mature, we expect further efficiency gains and sustained margin support as volumes scale.

Last quarter, we introduced a chart comparing client-facing headcount with the number of invoices purchased, highlighting meaningful efficiency gains from our investments. Going forward, as AI and automation extend beyond the back office to encompass the entire factoring operation, we will continue to demonstrate these efficiencies across the business.

How Triumph Intelligence Is Different. Triumph Intelligence is the only scaled market intelligence provider that delivers a holistic market price, considering performance and capacity, on both the buy and sell side of the broker transaction. We believe our offering is superior to incumbents not only because we have “more features,” but because we have better data, better analytics, and a forward-looking posture. Fully settled real-time transactions are THE best source of truth. This data set allows machine learning to build firm-specific views which create realizable value for customers. As we continue to integrate Intelligence into the Triumph Network, we expect a reinforcing feedback loop: adoption in other segments enriches the data asset, thereby increasing product value for those same customers.

AI, Disruption, and Our Approach

We generally think about AI in three categories: functionality, efficiency, and data. We participate in all three. Investors may see the data and product opportunity today; the efficiency benefits should be increasingly visible in 2026.

Our working view is that AI will compress “feature moats” across software, but it does not eliminate moats built on trusted networks, embedded workflows, and proprietary real‑time data. The most useful way to frame this — internally and with investors — is to distinguish between building a tool and operating a two‑sided network that sits inside the core transaction. In that construct, AI may replicate discrete product functionality faster, but it cannot reproduce the combination of network participation on both sides of the freight payment, the trust layer that enables audit and accurate settlement, the fraud‑mitigation guardrails, and the downstream intelligence embedded within those workflows. The defensibility is therefore less about “software we wrote” and more about the ecosystem we operate and the data asset we continuously create: settled, clean, real‑time transactions that can train models and drive actionable intelligence in a high‑confidence environment.

At the operating level, we view AI as both (i) a lever to lower cost‑to‑serve and (ii) an opportunity to redesign workflows from first principles rather than simply “bolting AI on” to existing processes. The strategic winners will be the companies that rebuild processes around AI and use it to do things better and differently — not just faster. We are applying that mindset alongside a lean operating posture that remains customer‑centric and innovation‑forward, while maintaining humility about the range of outcomes and resisting hype-driven narratives. We expect to continue innovating as we execute, with outcomes investors can see in our results, not press releases.

Our reportable segments are Banking, Factoring, Payments, and Intelligence, which have been determined based upon their business processes and economic characteristics. This determination also gave consideration to the structure and management of various product lines. The Banking segment includes the operations of TBK Bank. Our Banking segment derives its revenue principally from investments in interest earning assets as well as noninterest income typical for the banking industry. The Factoring segment derives its revenue from factoring services. The Payments segment includes the operations of TBK Bank's presentment, audit, and payment solution to Shipper, Broker, and Factor clients in the trucking industry. The Payments segment derives its revenue from transaction fees and interest income on factored receivables related to invoice payments. These factored receivables consist of both invoices where we offer a Carrier a quickpay opportunity to receive payment at a discount in advance of the standard payment term for such invoice in exchange for the assignment of such invoice to us and from offering Brokers the ability to settle their invoices with us on an extended term following our payment to their Carriers as an additional liquidity option for such Brokers. Our data intelligence segment was launched at the beginning of the fourth quarter of 2024 to turn the over-the-road trucking data collected through our services into actionable insights for our customers. This launch coincided with our acquisition of the assets Isometric Technologies Inc. that provides service and performance scoring and benchmarking capabilities to the over-the-road trucking industry. The operations of this segment were further supplemented with our acquisition of Greenscreens AI, Inc., a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights, during the quarter ended June 30, 2025. The revenue for Intelligence offerings is derived through access and subscription fees, as well as seat licenses where applicable.

Our business is conducted through four reportable segments (Banking, Factoring, Payments, and Intelligence). For the year ended December 31, 2025, our Banking segment generated 57% of our total segment revenue (comprised of interest and noninterest income), our Factoring segment generated 31% of our total segment revenue, our Payments segment generated 11% of our total segment revenue, and our Intelligence segment generated 1% of our total segment revenue. As part of our business operations, we use automated technologies, including those supported by artificial intelligence. This includes the instant purchase model we have deployed as part of our factoring operations.

Our data intelligence division, which we call Intelligence, was launched during the fourth quarter of 2024 to turn the over-the-road trucking data collected through our services into actionable insights for our customers. This launch coincided with our acquisition of Isometric Technologies Inc., a company that provides service and performance scoring and benchmarking capabilities to the over-the-road trucking industry. During the second quarter of 2025, we acquired Greenscreens AI. Inc., a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights, which further augmented the product and service offerings of our Intelligence division. Data has the ability to drive efficiency, enhance decision-making, and enable Shippers, Brokers, and Carriers to operate more profitably in a very competitive over-the-road trucking market. With our access to data from our Payments network and other sources, we believe we can develop products and services to offer to logistics service providers, allowing them to better plan for peak periods, competitively source freight capacity, and allocate resources efficiently, thus improving their profitability. Our Intelligence division operates in a highly specialized niche with unique processes and key performance indicators.

Our transportation payments products (i.e., factoring and payments) require specialized underwriting processes. For each factoring transaction, in addition to a credit evaluation of our client, we also evaluate the creditworthiness of underlying account debtors, because account debtors represent the substantive underlying credit risk. Transportation factoring also presents the additional challenge of underwriting high volumes of invoices of predominantly low value per invoice and managing credit requests for a large industry pool of account debtors. We facilitate this process through a proprietary web-based “Online Broker Credit” application, which processes invoice purchase approval requests for our clients through an online proprietary scoring model and delivers either preliminary responses for small dollar requests or immediate referral to our servicing personnel for larger dollar requests. In 2024, we launched our instant purchase decision model to augment and add efficiencies to invoice purchase underwriting in our factoring business. This model uses machine learning and artificial intelligence, based on a rule set established by our risk model that we have developed over our history, to screen invoices for compliance with appropriate criteria and risk scoring and approve them for purchase in seconds without human intervention. This model was launched at scale in our factoring division's small carrier group during 2024 and the Company anticipates continued deployment to other parts of its factoring business in the future. We also set and monitor concentration limits for individual account debtors that are tracked across all of our clients (as multiple clients may have outstanding invoices from a particular account debtor). For each Broker or Shipper client, for whom we will be originating quick pay or supply chain finance transactions, we conduct an in-depth credit evaluation and underwriting process. We facilitate this process by collecting detailed company and financial information, which we analyze to determine credit risk.

•the impact of recent and future legislative and regulatory changes, including changes in banking, securities and tax laws and regulations, such as the Dodd-Frank Act and their application by our regulators, as well as privacy, cybersecurity, and artificial intelligence regulation and oversight;

Our future Payments and Intelligence revenue will depend in large part on our success in attracting additional customers to our network platform. Our ability to attract additional customers will depend on a number of factors, including the effectiveness of our sales team, the success of our marketing efforts, our levels of investment in expanding our sales and marketing teams, success in developing and rolling out additional features and functionality on the platform (including integration of artificial intelligence tools and the development of products and services for our Intelligence business), and the availability of competitive transportation payments technology platforms. We may not experience the same levels of success with respect to our customer acquisition strategies as seen in prior periods, and if the costs associated with acquiring new customers materially rises in the future, our expenses may rise significantly.

Our current customer base consists primarily of third party logistics companies, or Brokers, making payments to their Carriers through our payments platform, as well as Brokers and Factors that process their invoices for payment or purchase, as applicable, through the audit functionality on the payments platform. We intend to continue to pursue growth within each of our target customer markets (Broker, Shipper and Factor) and to seek to convert customers using only a portion of the payments platform functionality (payments or audit), to use the other services on the payments platform to conduct end to end integrated payments transactions that create benefit for the other parties to the payment transaction on the platform, and to continue exploring the use and deployment of artificial intelligence as well as other new technology tools on the platform. This also includes the development of additional data driven products and services as part of our Intelligence business. Each of these efforts pose different sales and marketing challenges, and present different requirements. We cannot be sure that we will achieve the success in these expanded efforts and failure to achieve such success will hinder our growth prospects and strategic outlook.

Our Payments operations have grown significantly in recent periods both organically and through acquisition, and have a limited operating history, particularly at our current scale. Our Intelligence division was established in the fourth quarter of 2024, was supplemented by our acquisition of Greenscreens AI Inc. in the second quarter of 2025, an also has a limited operating history. These businesses operate in rapidly evolving industries. We have frequently expanded, and plan to continue to expand, our network platform features and services for these businesses and will likely continue to change and develop our pricing methodologies for these products and services. This limited operating history and our evolving business make it difficult to evaluate our future prospects and the risks and challenges we may encounter. These risks and challenges include, but are not limited to, our ability to:

•integrate new technologies on our platform, including artificial intelligence tools;

Our continued growth depends in part on the ability of our existing and potential customers to access our network and payments platform at any time and within an acceptable amount of time. Our platform is proprietary, and we rely on the expertise of members of our engineering, operations, and software development teams for our platform’s continued performance. We have experienced, and may in the future experience, disruptions, outages, and other performance problems related to our platform due to a variety of factors, including infrastructure changes, introductions of new functionality (including functionality that incorporates artificial intelligence tools), human or software errors, delays in scaling our technical infrastructure if we do not maintain enough excess capacity and accurately predict our infrastructure requirements, capacity constraints due to an overwhelming number of users accessing our platform simultaneously, denial-of-service attacks, human error, actions or inactions attributable to third parties, earthquakes, hurricanes, floods, fires, natural disasters, power losses, disruptions in telecommunications services, fraud, military or political conflicts, terrorist attacks and other geopolitical unrest, computer viruses, ransomware, malware, or other events. Our systems also may be subject to break-ins, sabotage, theft, and intentional acts of vandalism, including by our own employees. Some of our systems are not fully redundant and our disaster recovery planning may not be sufficient for all eventualities. Further, our business and/or network interruption insurance may not be sufficient to cover all of our losses that may result from interruptions in our service as a result of systems failures and similar events.

The markets in which we compete are characterized by constant change and innovation, and we expect them to continue to evolve rapidly. Our success has been based on our ability to identify and anticipate the needs of our customers and design and maintain a platform that provides them with the tools they need to operate their businesses successfully. Our ability to attract new customers, retain existing customers, and increase sales to both new and existing customers will depend in large part on our ability to continue to improve and enhance the performance, reliability, and features of our platform. To grow our business, we must develop products and services that reflect the changing nature of transportation logistics software and expand beyond our core functionalities to other areas of managing relationships with our customers, as well as their relationships with their customers. Competitors may introduce new offerings embodying new technologies, or new industry standards and practices could emerge, that render our existing technology, services, website, and mobile applications obsolete. Accordingly, our future success will depend in part on our ability to respond to new product offerings by competitors, technological advances, including automated technologies like artificial intelligence, and emerging industry standards and practices in a cost-effective and timely manner in order to retain existing customers and attract new customers. Furthermore, as the number of our customers with higher transportation volumes increases, so does the need for us to offer increased functionality, scalability, and support, which requires us to devote additional resources to such efforts.

Though the transportation factoring industry continues to fight headwinds due to higher cost of capital and lower average invoices, we have sufficient access to capital, manageable funding costs, and an ability to diversify transportation and factoring income. We continue to focus our efforts on technology initiatives to be more efficient, support the enterprise, and enhance our customer experience while delivering various products to strengthen our clients throughout their business lifecycle. Our plan is for managed growth in our factoring segment with a greater emphasis on enhancing efficiency and profitability. These plans may include use of new technology tools, including those that integrate artificial intelligence capabilities.

Our reportable segments are Banking, Factoring, Payments, and Intelligence, which have been determined based upon their business processes and economic characteristics. This determination also gave consideration to the structure and management of various product lines. The Banking segment includes the operations of TBK Bank. Our Banking segment derives its revenue principally from investments in interest earning assets as well as noninterest income typical for the banking industry. The Factoring segment derives its revenue from factoring services. The Payments segment includes the operations of TBK Bank's presentment, audit, and payment solution to Shipper, Broker, and Factor clients in the trucking industry. The Payments segment derives its revenue from transaction fees and interest income on factored receivables related to invoice payments. These factored receivables consist of both invoices where we offer a Carrier a quickpay opportunity to receive payment at a discount in advance of the standard payment term for such invoice in exchange for the assignment of such invoice to us and from offering Brokers the ability to settle their invoices with us on an extended term following our payment to their Carriers as an additional liquidity option for such Brokers. Our data intelligence segment was launched at the beginning of the fourth quarter of 2024 to turn the over-the-road trucking data collected through our services into actionable insights for our customers. This launch coincided with our acquisition of Isometric Technologies Inc. that provides service and performance scoring and benchmarking capabilities to the over-the-road trucking industry. The operations of this segment were further supplemented with our acquisition of Greenscreens AI. Inc., a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights, during the quarter ended June 30, 2025. The revenue for Intelligence offerings is derived through access and subscription fees, as well as seat licenses where applicable. Prior to the fourth quarter of 2024, there were no individuals allocated specifically to our data intelligence segment and an explicit data intelligence segment did not exist. Therefore, revision of prior period segment operating results is not applicable.

Expense Management, Goals for 2026 and The Role of AI in 2026

The Role of AI in 2026. The $30 million of expense savings we have already recognized, and the additional savings that are to come, are not just a result of across-the-board expense cuts. We have used and will continue to use artificial intelligence (AI) and machine learning (ML) to run our business more efficiently. We have already leveraged AI solutions to streamline document processing with our audit platforms, enable instant decisioning of invoice purchases, predict lane rates, and improve customer service response times even while generating cost savings. In the upcoming year, we will be deploying new models to further enhance our customer offerings and streamline internal operations. We will continue to be thoughtful about what we disclose publicly, as some capabilities are proprietary. In sum, we believe our data advantage and disciplined execution in this area are becoming a durable source of operating leverage.

2.Automation & AI Update

Automation & AI Update. I touched on this briefly above regarding enterprise initiatives, but I will include a few examples here that are germane to our Factoring segment. Our instant decision model continues to drive efficiency in our O/O segment. We are seeing 60%+ of these submitted invoices instantly verified and approved by our AI tools, and 20%+ of them proceed through purchase and funding without a human touching them. This is relatively flat with the prior quarter due in large part to a decision to tighten the automatic controls around this tool. The machines have adjusted, and we expect to continue to see progress and improvement in our touchless processing throughout 2026. Our average time to approve an invoice through these tools is about 12 seconds, with the purchase decision occurring within 23 seconds on average. The product team continues working on improving the model to better support our large fleet operations team, which is currently experiencing a 19% automated approval rate. This is a custom-built, AI driven process that I believe to be unrivaled in the industry (that is an educated guess and largely subjective). What I know for sure is that we will continue to update our AI models to improve performance, reduce friction, and increase margins. As an example of these efficiencies we have already realized, below is a chart that reflects the trend in invoice volume purchased vs operational headcount that supports invoice purchases since the instant decision model launched in June of 2024. Since that date, the number of invoices purchased (point-to-point) has increased 29%, while headcount has decreased 25%. That is largely how we have achieved margin expansion. There is much more to do, and we have several initiatives in flight.

As a final note on this, I would point out that we have primarily used AI to strengthen the pre-purchase side of our business for the past few years. As we close 2025 and move into 2026, our focus has shifted towards automation and AI-driven initiatives to enhance client service and streamline back-office functions in our Factoring segment. This transition includes a deliberate effort to standardize, optimize, and automate where it makes sense, enabling us to continually improve operating margin and reduce our cost to serve.

•Expense Reduction & Margin Improvement. We reduced total expenses by ~5% through our efficiency efforts in 3Q, and 90% of those savings are in our run rate as we begin 4Q. We expect the remaining savings to be in place by year-end. These restructuring efforts were not a one-and-done event; they are representative of the evolution of our business I've been alluding to for the last few quarters. We are now at the point of properly monetizing our platform and Network and improving efficiency. Our work to become leaner and more efficient is far from done, and we expect to progress on this front in 2026. The investments we have made in AI tools are bearing the expected fruit, and we continue to find new ways to incorporate these tools into our operations, as I discuss later. We project 4Q expenses to be $96.5 million, which is 4.5% below adjusted 2Q 2025 numbers. The following table shows the expense reductions segment and adjusts for the impact of non-recurring items.

2.Automation & AI Update

Automation & AI Update. We remain focused on initiatives that will drive revenue growth and improve operating margin. In 3Q, we launched automated cash posting, delivering a ~20% efficiency lift within our treasury department. We expect that additional enhancements that are underway in this area will further streamline back-office operations and reduce manual effort.

Our instant decision model continues to drive efficiency, with a pass and approval rate of 58% in our O/O segment, enabling these clients to get instant funding when they use LoadPay or our fuel card. Today, the product team is working on improving the model to better support our large fleet operations team, which is currently experiencing a ~15% approval rate. This is custom built AI that is, in my opinion, unrivaled in the industry. We will continue to update our AI models to improve performance, reduce friction, and increase margins.

In late September, we launched our integrated Pricing and Performance Intelligence solution. This product unifies pricing, performance, and capacity sourcing into a single data solution for freight brokers. With this tool, brokers can see rate, performance, and embedded capacity information with AI-driven predictive analytics and capacity sourcing suggestions. This solution builds on the AI models we acquired with Greenscreens. By combining real-world performance metrics with predictive analytics, brokers can sharpen their pencil on pricing, making smarter, more efficient decisions, balancing cost and reliability to improve load coverage and overall freight outcomes. Using AI and our proprietary data to do this makes our offering differentiated and highly scalable. I would also mention that we have ~400 Payments Network customers who do not currently use our Intelligence solution or its Greenscreens predecessor, and I would repeat here what I said earlier about treating customers the way we want to be treated to build loyalty and brand trust. If customers can trust you to do one thing well, they are likely to trust you with another. I would expect many of our existing Payments customers to become Intelligence customers in 2026.

Our data intelligence business, which we call Intelligence, was launched at the beginning of the fourth quarter of 2024 to turn the over-the-road trucking data collected through our services into actionable insights for our customers. This launch coincided with our acquisition of the assets of Isometric Technologies Inc., a company that provides service and performance scoring and benchmarking capabilities to the over-the-road trucking industry. During the second quarter of 2025, we also acquired Greenscreens AI, Inc. ("Greenscreens"), a company that provides a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights. Data has the ability to drive efficiency, enhance decision-making, and enable Shippers, Brokers, and Carriers to operate more profitably in a very competitive over-the-road trucking market. With our access to data from our payments network and other sources, we believe we can develop products and services to offer to logistics service providers, allowing them to better plan for peak periods, competitively source freight capacity, and allocate resources efficiently, thus improving their profitability. Intelligence operates in a highly specialized niche with unique processes and key performance indicators.

Though the transportation factoring industry continues to fight headwinds due to higher cost of capital and lower average invoices, we have sufficient access to capital, manageable funding costs, and an ability to diversify transportation and factoring income. We continue to focus our efforts on technology initiatives to be more efficient, support the enterprise, and enhance our customer experience while delivering various products to strengthen our clients throughout their business lifecycle. Our plan is for managed growth in our factoring segment with a greater emphasis on enhancing efficiency and profitability. These plans may include use of new technology tools, including those that integrate artificial intelligence capabilities.

Our reportable segments are Banking, Factoring, Payments, and Intelligence, which have been determined based upon their business processes and economic characteristics. This determination also gave consideration to the structure and management of various product lines. The Banking segment includes the community banking products and services offered through TBK Bank. Our Banking segment derives its revenue principally from investments in interest earning assets as well as noninterest income typical for the banking industry. The Factoring segment derives its revenue from factoring services. The Payments segment includes the operations of TBK Bank's presentment, audit, and payment solutions to Shipper, Broker, and Factor clients in the trucking industry. The Payments segment derives its revenue from transaction fees and interest income on factored receivables related to invoice payments. These factored receivables consist of both invoices where we offer a Carrier a quickpay opportunity to receive payment at a discount in advance of the standard payment term for such invoice in exchange for the assignment of such invoice to us and from offering Brokers the ability to settle their invoices with us on an extended term following our payment to their Carriers as an additional liquidity option for such Brokers. Our data intelligence segment was launched at the beginning of the fourth quarter of 2024 to turn the over-the-road trucking data collected through our services into actionable insights for our customers. This launch coincided with our acquisition of the assets Isometric Technologies Inc., a company that provides service and performance scoring and benchmarking capabilities to the over-the-road trucking industry. The operations of this segment were further supplemented with our acquisition of Greenscreens AI, Inc., a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights, during the quarter ended June 30, 2025. The revenue for Intelligence offerings is derived through access and subscription fees, as well as seat licenses where applicable. Prior to the fourth quarter of 2024, there were no individuals allocated specifically to our data intelligence segment and an explicit data intelligence segment did not exist. Therefore, revision of prior period segment operating results is not applicable.

25 Acquisitions: The Initial Core of Triumph's Intelligence Offering Greenscreens helps logistics companies make better, data- backed, AI-driven pricing decisions—boosting competitiveness and profitability while integrating seamlessly into existing systems. The platform embeds within existing TMS (Transportation Management Systems), offering embedded UI and automation streamlining quoting and pricing workflows. ISO is logistics performance intelligence platform designed to help shippers, brokers, and carriers better understand and optimize their transportation procurement. Real-Time and Predictive Pricing - Combines historical settled freight rates, real-time market intelligence, and machine learning to provide short-term predictive price guidance — for both buying and selling freight — predicting rates 2× more accurately than competitive products Market Data Aggregation - Aggregates settled transaction data from across the market feeding it back into the model as high-confidence pricing insights tailoring results to an individual company’s behavior, targets, and margin requirements Margin & Volume Optimization - AI powered predictive tools empower logistics service providers to quote with confidence, win more business, and protect or increase profit margins per transaction in keeping with their specific goals Total Cost of Procurement - ISO goes beyond paper freight rates by integrating service reliability data—such as on-time pickup, delivery and tender acceptance—to calculate the true landed cost of shipment, including penalties, delays, and inefficiencies Data Reconciliation & Accountability – Streamlines the reconciliation of performance data (e.g. missed pickups, chargebacks, OTIF penalties) across all parties—shippers, brokers, carriers— creating a unified "single source of truth." This helps identify which partner is responsible for service failures Performance Scorecarding & Benchmarking - Provides dynamic scorecards, internal/external benchmarks, and a cumulative "ISO Score" (akin to a FICO score for freight) including price and service, to compare carriers and lanes objectively Predictive Insights & Root-Cause Analysis - The platform uncovers hotspots in supply chains, predicts disruptions (e.g. penalties, delays), and allows drill-down into why issues occur— helping procurement teams address underlying problems © TBK Bank, SSB, DBA Triumph

1.Audit: We acquired the capability to audit broker invoices through Hubtran in 2021. We now provide audit services for over half of the top 100 brokers and we audit more brokered freight volume than any known competitor. Our NextGen Audit product will lead the industry for features and stability. It leverages AI-driven workflows and deep TMS integrations to automate tasks, transforming the back office from reactive to exception based. Audit is the tool that has already been and will most be affected by strides in machine learning and artificial intelligence. In the next few years, industry models will likely be good enough to handle almost all exceptions. At that time, it will probably not be enough to just offer audit as a standalone product. To be competitive, it will have to be part of a suite of services. This suite is one of Triumph's competitive advantages.

•Real-Time and Predictive Pricing: Combines historical settled freight rates, real-time market intelligence, and machine learning to provide near-term predictive price guidance for both buying and selling freight. Because we use more reliable ingredients in the input data I referenced above, our testing demonstrates that our prediction models are 2× more accurate than competitive products. Expect to hear more about this in coming quarters.

•Margin & Volume Optimization: AI-powered predictive tools empower logistics service providers to quote with confidence, win more business, and protect or increase profit margins per transaction in keeping with their specific goals.

Technology and FaaS updates. As of June 30, the instant decision model is active on the entire portfolio except Supply Chain Finance (SCF). Instant decision uses our data extraction technology, risk models and machine learning to make purchasing decisions on specific invoices without human intervention. We continue to believe that we have a 30% - 45% revenue growth opportunity before adding material expense as a result of the technological investments we have made over the last few years.

Our data intelligence business, which we call Intelligence, was launched at the beginning of the fourth quarter of 2024 to turn the over-the-road trucking data collected through our services into actionable insights for our customers. This launch coincided with our acquisition of the assets of Isometric Technologies Inc., a company that provides service and performance scoring and benchmarking capabilities to the over-the-road trucking industry. During the second quarter of 2025, we also acquired Greenscreens AI, Inc. ("Greenscreens"), a company that provides a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights. Data has the ability to drive efficiency, enhance decision-making, and enable Shippers, Brokers, and Carriers to operate more profitably in a very competitive over-the-road trucking market. With our access to data from our payments network and other sources, we believe we can develop products and services to offer to logistics service providers, allowing them to better plan for peak periods, competitively source freight capacity, and allocate resources efficiently, thus improving their profitability. Intelligence operates in a highly specialized niche with unique processes and key performance indicators.

Though the transportation factoring industry continues to fight headwinds due to higher cost of capital and lower average invoices, we have sufficient access to capital, manageable funding costs, and an ability to diversify transportation and factoring income. We continue to focus our efforts on technology initiatives to be more efficient, support the enterprise, and enhance our customer experience while delivering various products to strengthen our clients throughout their business lifecycle. Our plan is for managed growth in our factoring segment with a greater emphasis on enhancing efficiency and profitability. These plans may include use of new technology tools, including those that integrate artificial intelligence capabilities.

Intelligence. Finally, we will go to market with our new Intelligence offering and expect to take a ratings and performance product to market later this year following the anticipated closing of our Greenscreens acquisition. We expect this closing to occur in 2Q. Greenscreens is the fastest-growing rate intelligence provider in transportation. We are buying Greenscreens because we believe they have the best technology in the industry. Greenscreens’ core product is its freight-pricing software, which leverages AI/ML tools, and uses data from customer freight transactions to provide data analytics and insights regarding pricing to its customers. We can further empower that product suite with our own data and the data of our customers who wish to use our Intelligence services. When we layer that best-in-class technology onto those data assets, which we believe is the largest and cleanest repository of settled truckload payment data in the U.S., investors can get some idea why our payor clients are excited about the possibilities. Let me be clear (and this is my opinion, of course), Triumph will be producing the most trusted and comprehensive pricing tool in the market. We will be able to demonstrate that clearly and objectively. We have the strategy laid out, and we are laser focused on bringing it to reality post-closing on Greenscreens.

In previous letters, we introduced our Instant Decision model which is powered by AI/ML coupled with our Network data and our risk management program. The model continues to deliver measurable value, although material ROI will be recognized as client growth is achieved. This allows us to scale intelligently, particularly within our small carrier segment, consisting of carriers with 1 to 5 trucks, where we can grow by up to 30% without increasing headcount.

On February 26, 2025, the Company, through its wholly-owned subsidiary TBK Bank, SSB, entered into an Agreement and Plan of Merger providing for the acquisition of GreenScreens AI, Inc. (Greenscreens.ai), a pricing solution for the logistics industry that delivers short-term freight market pricing intelligence and business insights, by TBK Bank, SSB for $140,000,000 in cash and approximately $20,000,000 in the Company's common stock. The acquisition is subject to customary closing conditions, including receipt of regulatory approval, and is expected to close in the second quarter of 2025.

Though the transportation factoring industry continues to fight headwinds due to higher cost of capital and lower average invoices, we have sufficient access to capital, manageable funding costs, and an ability to diversify transportation and factoring income. We continue to focus our efforts on technology initiatives to be more efficient, support the enterprise, and enhance our customer experience while delivering various products to strengthen our clients throughout their business lifecycle. Our plan is for managed growth in our factoring segment with a greater emphasis on enhancing efficiency and profitability. These plans may include use of new technology tools, including those that integrate artificial intelligence capabilities.

Our business is conducted through four reportable segments (Banking, Factoring, Payments, and Intelligence). For the year ended December 31, 2024, our Banking segment generated 60% of our total revenue (comprised of interest and noninterest income), our Factoring segment generated 30% of our total revenue, our Payments segment generated 10% of our total revenue, and our Intelligence segment generated less than 1% of our total revenue. As part of our business operations, we use automated technologies, including those supported by artificial intelligence. This includes the instant purchase model we have deployed as part of our factoring operations.

Our transportation payments products (i.e., factoring and TriumphPay) require specialized underwriting processes. For each factoring transaction, in addition to a credit evaluation of our client, we also evaluate the creditworthiness of underlying account debtors, because account debtors represent the substantive underlying credit risk. Transportation factoring also presents the additional challenge of underwriting high volumes of invoices of predominantly low value per invoice and managing credit requests for a large industry pool of account debtors. We facilitate this process through a proprietary web-based “Online Broker Credit” application, which processes invoice purchase approval requests for our clients through an online proprietary scoring model and delivers either preliminary responses for small dollar requests or immediate referral to our servicing personnel for larger dollar requests. In 2024, we launched our instant purchase decision model to augment and add efficiencies to invoice purchase underwriting in our factoring business. This model uses machine learning and artificial intelligence, based on a rule set established by our risk model that we have developed over our history, to screen invoices for compliance with appropriate criteria and risk scoring and approve them for purchase in seconds without human intervention. This model was launched at scale in our factoring division's small carrier group during 2024 and the Company anticipates continued deployment to other parts of its factoring business in the future. We also set and monitor concentration limits for individual account debtors that are tracked across all of our clients (as multiple clients may have outstanding invoices from a particular account debtor). For each Broker or Shipper client, for whom we will be originating quick pay or supply chain finance transactions, we conduct an in-depth credit evaluation and underwriting process. We facilitate this process by collecting detailed company and financial information, which we analyze to determine credit risk.

•the impact of recent and future legislative and regulatory changes, including changes in banking, securities and tax laws and regulations, such as the Dodd-Frank Act and their application by our regulators, as well as privacy, cybersecurity, and artificial intelligence regulation and oversight;

Our future Payments and Intelligence revenue will depend in large part on our success in attracting additional customers to our network platform. Our ability to attract additional customers will depend on a number of factors, including the effectiveness of our sales team, the success of our marketing efforts, our levels of investment in expanding our sales and marketing teams, success in developing and rolling out additional features and functionality on the platform (including integration of artificial intelligence tools and the development of products and services for our Intelligence business), and the availability of competitive transportation payments technology platforms. We may not experience the same levels of success with respect to our customer acquisition strategies as seen in prior periods, and if the costs associated with acquiring new customers materially rises in the future, our expenses may rise significantly.

Our current customer base consists primarily of third party logistics companies, or Brokers, making payments to their Carriers through our TriumphPay platform, as well as Brokers and Factors that process their invoices for payment or purchase, as applicable, through the audit functionality on the TriumphPay platform. We intend to continue to pursue growth within each of our target customer markets (Broker, Shipper and Factor) and to seek to convert customers using only a portion of the TriumphPay functionality (payments or audit), to use the other services on the TriumphPay platform to conduct end to end integrated payments transactions that create benefit for the other parties to the payment transaction on the platform, and to continue exploring the use and deployment of artificial intelligence as well as other new technology tools on the platform. This also includes the development of additional data driven products and services as part of our Intelligence business. Each of these efforts pose different sales and marketing challenges, and present different requirements. We cannot be sure that we will achieve the success in these expanded efforts and failure to achieve such success will hinder our growth prospects and strategic outlook.

•integrate new technologies on our platform, including artificial intelligence tools;

Our continued growth depends in part on the ability of our existing and potential customers to access our network and payments platform at any time and within an acceptable amount of time. Our platform is proprietary, and we rely on the expertise of members of our engineering, operations, and software development teams for our platform’s continued performance. We have experienced, and may in the future experience, disruptions, outages, and other performance problems related to our platform due to a variety of factors, including infrastructure changes, introductions of new functionality (including functionality that incorporates artificial intelligence tools), human or software errors, delays in scaling our technical infrastructure if we do not maintain enough excess capacity and accurately predict our infrastructure requirements, capacity constraints due to an overwhelming number of users accessing our platform simultaneously, denial-of-service attacks, human error, actions or inactions attributable to third parties, earthquakes, hurricanes, floods, fires, natural disasters, power losses, disruptions in telecommunications services, fraud, military or political conflicts, terrorist attacks and other geopolitical unrest, computer viruses, ransomware, malware, or other events. Our systems also may be subject to break-ins, sabotage, theft, and intentional acts of vandalism, including by our own employees. Some of our systems are not fully redundant and our disaster recovery planning may not be sufficient for all eventualities. Further, our business and/or network interruption insurance may not be sufficient to cover all of our losses that may result from interruptions in our service as a result of systems failures and similar events.

The markets in which we compete are characterized by constant change and innovation, and we expect them to continue to evolve rapidly. Our success has been based on our ability to identify and anticipate the needs of our customers and design and maintain a platform that provides them with the tools they need to operate their businesses successfully. Our ability to attract new customers, retain existing customers, and increase sales to both new and existing customers will depend in large part on our ability to continue to improve and enhance the performance, reliability, and features of our platform. To grow our business, we must develop products and services that reflect the changing nature of transportation logistics software and expand beyond our core functionalities to other areas of managing relationships with our customers, as well as their relationships with their customers. Competitors may introduce new offerings embodying new technologies, or new industry standards and practices could emerge, that render our existing technology, services, website, and mobile applications obsolete. Accordingly, our future success will depend in part on our ability to respond to new product offerings by competitors, technological advances, including automated technologies like artificial intelligence, and emerging industry standards and practices in a cost-effective and timely manner in order to retain existing customers and attract new customers. Furthermore, as the number of our customers with higher transportation volumes increases, so does the need for us to offer increased functionality, scalability, and support, which requires us to devote additional resources to such efforts.

Though the transportation factoring industry continues to fight headwinds due to higher cost of capital and lower average invoices, we have sufficient access to capital, manageable funding costs, and an ability to diversify factoring income. We continue to focus our efforts on technology initiatives to be more efficient, support the enterprise, and enhance our customer experience while delivering various products to strengthen our clients throughout their business lifecycle. Our plan is for managed growth in our factoring segment with a greater emphasis on enhancing efficiency and profitability. These plans may include use of new technology tools, including those that integrate artificial intelligence capabilities.

Technology advances in extracting data from invoices and using AI/ML models to make instant purchase decisions have replaced some of the value of structured data for factors – or at least for our own Factoring segment. Other factors may attempt to replicate this with overseas manual labor if they do not have a form of the technology built for themselves. With the level of sophistication our Factoring segment has developed, combined with the historical data our models leverage, we can make very educated decisions almost instantly even when it is not a network transaction (and thus lacks the structured data that comes with it). I have seen first-hand that AI/ML is better than humans at recognizing patterns. The strides we have made with Instant Decision in the last 12 months are remarkable. Granted, not every factor has the same level of technology we do, but other large factors probably have a basic form of it or are working on one.

Even if a factor has an AI/ML driven decisioning model, it is still prudent to leverage a network transaction when available. It allows the factor to bounce the purchasing decision off the broker’s TMS for near absolute certainty pre-purchase. This assumes that the Factor’s tech stack has been modified to ingest this data, which is not universally true. Triumph's model (and thus, our FaaS offering) is built to prioritize structured network data first and then move on to data extraction and algorithmic techniques to complete the purchase decision process as necessary. This all happens in seconds.

The early results are impressive: Of those small carriers, 74.8% pass through the model and pay in full without dispute or human interaction. The average time to approve an invoice is nine seconds, a testament to the model’s ability to streamline operations while reducing risk during the purchase decision process. Investors should be wary of claims about “touch-free processing” that so many companies claim. Some AI tools in the market tout high percentages on easier tasks and then transfer the process back to a human for more challenging elements. When we say touchless, we mean it – from verification to validation – almost 75% of the invoices passing through our Instant Decision tool are never touched by a human in the purchase process. We continue training and improving the tools and have high expectations for the future.

© Triumph Financial, Inc TFIN launches Factoring as a Service (FaaS) 7Payments DataFactoring Legacy Factoring TFIN FaaS Closed inefficient market requiring extensive institutional knowledge A platform empowering new entrants, leveraging our industry expertise, leading technology and extensive network Off the shelf core systems lacking investment and features Purpose built, scalable core systems allowing instant purchase decisions of invoices and instant funding for carriers 24x7x365 People intensive to scale, little motivation to innovate Significant investments in technology to automate processes, reduce risk, improve the customer experience and efficiently scale new client business Significant risk of fraud AI/ML tools reduce legacy processes from hours to seconds improving accuracy and credit metrics while reducing fraud Capital intensive with significant funding requirements Leverages TFIN's balance sheet and cost of funds

we made enhancements to our original Audit product that can enable 85%+ touchless invoice processing. Additionally, our AI/ML technology is 94%+ accurate when extracting data from invoices. I will not go into specifics here about features we are releasing for competitive reasons. What I will say is that over the next few quarters, we will continue to invest, focusing on proactive features which prevent exceptions before they are created and further increase the percentage of invoices that are approved and paid without being touched.

Last, but certainly not least, is the rollout of what we believe to be the best-in-class decisioning model to make autonomous invoice purchase decisions. We call this “instant purchase” because the machine can make purchasing decisions instantly. At the core of this technology is an AI-based machine learning model. This has been developed internally and thoroughly backtested using our historical performance data. When augmented with TriumphPay integrated data, the machines are able to make purchase decisions in seconds without human interaction. The model observes various behaviors and trends which are markers for fraud. We couple that with data from our proprietary risk management program developed over the last 15 years. The model is now running in a pilot phase with roughly 600 clients today. Once fully released, we will see improved efficiency in our daily operations, and in conjunction with LoadPay, the ability to fund our clients instantly around the clock, regardless of bank hours and ACH windows. We expect that this technology will also be available for FaaS clients in 2025. The implications for efficiency and scale are very exciting.

•Aside from payment services, TriumphPay Audit provides automation and efficiency for the invoice auditing portion of a broker’s truckload a/p function. Through more traditional technologies such as OCR, and more recently enhanced with advanced AI/ML techniques, invoices received from carriers and factors are processed for a per-invoice fee.

Our teams continue to improve processes and enhance our technology using artificial intelligence and machine learning tools during this cycle. This positions Triumph for the freight rebound whenever that occurs — for our own account and as a future service provider to the factoring industry. We use artificial intelligence and machine learning in part as the engine for instant purchasing decisions, which gives us the ability to fund without time constraints if the carrier has a funding source integrated with us. As you might expect, we are looking forward to the additional value LoadPay will bring to this business. We also leverage these tools to create and improve operational efficiency for our existing business by reducing the amount of human intervention required in processes and procedures.

On the subject of operational efficiency, it is worth highlighting the results of some of our efforts this year in TriumphPay. In January, it cost our contact center and operations team roughly $0.64 to process an invoice.[1] As of September, utilizing artificial intelligence (AI), machine learning (ML), and some offshore resources, we have reduced that cost by 67% to $0.21 per invoice. In a prior letter, I referenced that we processed 26,000 Notice of Assignments (NOA) and Letters of Release (LORs) each day. Today, that figure is closer to 38,000, but with the AI/ML gains this year, over 25% of these submissions are now auto-processed with no human interaction on the first pass of a double blind process. Scale is a powerful force.