Cuộc họp công bố kết quả kinh doanh quý 3 năm tài chính 2026 của CS DISCO (LAW): EBITDA điều chỉnh dương và tăng trưởng AI

Các điểm chính

• Doanh thu quý 3 năm tài chính 2026 đạt 45,3 triệu USD, tăng 15% so với cùng kỳ năm ngoái nếu không tính vụ việc ngoài dự kiến được ghi nhận vào quý 3 năm tài chính 2025. Doanh thu phần mềm tăng 13% lên 38,6 triệu USD, trong khi doanh thu dịch vụ tăng 21% lên 6,7 triệu USD.
• EBITDA điều chỉnh đạt khoảng 0,1 triệu USD, chuyển sang mức dương sớm hơn một quý so với mục tiêu quý 4 trước đó của công ty. DISCO dự kiến EBITDA điều chỉnh trong quý 4 sẽ ở mức hòa vốn đến 1,5 triệu USD.
• Việc mở rộng tệp khách hàng lớn tiếp tục là động lực tăng trưởng chính. Số lượng khách hàng tạo ra doanh thu trên 100.000 USD trong 12 tháng gần nhất tăng lên 374 từ mức 326 khách hàng của một năm trước đó, đóng góp 133 triệu USD, tương đương 78% tổng doanh thu trong 12 tháng gần nhất.
• Doanh thu từ các tính năng AI tạo sinh và AI đại lý tăng hơn 200% so với cùng kỳ năm ngoái. Doanh thu từ AutoReview tăng hơn gấp 4 lần, trong khi số lượt câu lệnh (prompt) Cecilia AI tăng khoảng 40% trong quý 3.
• Doanh thu từ nền tảng DISCO tăng hơn 10 lần so với quý 1 năm tài chính 2026. Số lượng vụ việc lớn trên nền tảng đã tăng gấp ba trong quý 3, và mô hình này đã trở thành lựa chọn mặc định cho các hợp đồng mới và hợp đồng gia hạn kể từ ngày 1 tháng 10.
• Ban lãnh đạo đã nâng dự báo tổng doanh thu cho cả năm tài chính 2026 lên khoảng 176,1 triệu – 178,1 triệu USD. Công ty cho biết mảng kinh doanh cốt lõi có tiềm năng đạt mức tăng trưởng trên 20%, nhưng không đưa ra lộ trình cụ thể.

Dữ liệu tài chính cốt lõi

Kết quả kinh doanh và hoạt động

Chiến lược tăng trưởng của DISCO tiếp tục tập trung vào các vụ việc lớn hơn, mở rộng thị phần chi tiêu ở các khách hàng lớn và đẩy mạnh ứng dụng AI. Ban lãnh đạo cho biết các vụ việc lớn thường mang lại doanh thu cao hơn, duy trì trên nền tảng lâu hơn và quy mô tăng lên khi vụ án phát triển.

Số lượng khách hàng tạo ra doanh thu trên 100.000 USD trong 12 tháng gần nhất đã tăng lên 374 từ mức 354 khách hàng ở quý trước. Ban lãnh đạo cho biết phần lớn sự gia tăng này đến từ các khách hàng hiện hữu. Theo công ty, nhiều khách hàng trong nhóm này hiện chỉ dành 15%–20% ngân sách chi tiêu của họ cho DISCO.

Nền tảng DISCO là động lực lớn nhất thúc đẩy tăng trưởng doanh thu phần mềm. Mô hình định giá đơn giản hóa của nền tảng tích hợp Cecilia AI vào mọi vụ việc, và Cecilia đã được kích hoạt trên 96% số vụ việc trên nền tảng vào cuối quý. Ban lãnh đạo cho biết mô hình này đang giúp thu hút các vụ việc lớn hơn, kéo dài hơn và phức tạp hơn.

Mức độ ứng dụng AI cũng gia tăng nhanh chóng. Tính năng Nghiên cứu Nâng cao (Advanced Research) chiếm 91% lượng sử dụng hỏi đáp của Cecilia trong quý 3 và 97% trong tháng 10, trong đó mức độ sử dụng cao nhất ghi nhận ở các vụ việc lớn. Số lượng khách hàng sử dụng Cecilia trên nhiều vụ việc đã tăng hơn gấp đôi. DISCO cũng mở rộng thử nghiệm giải pháp tranh tụng hợp nhất và dự kiến sẽ tiếp nhận nhóm khách hàng thứ ba trước khi kết thúc năm.

Tăng trưởng mảng dịch vụ chủ yếu nhờ vào các dịch vụ chuyên nghiệp gắn liền với các vụ việc lớn. Công ty xem dịch vụ là cách thức để giành chiến thắng và tiếp nhận các vụ việc phức tạp, đồng thời hỗ trợ tạo ra doanh thu phần mềm tiếp theo.

Dự báo của ban lãnh đạo

Ban lãnh đạo cho biết điểm trung bình của dự báo doanh thu cả năm đã được điều chỉnh hiện cao hơn mức trần của khoảng dự báo đưa ra vào đầu năm tài chính 2026. Công ty tiếp tục kỳ vọng đạt EBITDA điều chỉnh dương trong quý 4.

Rủi ro và các yếu tố cần theo dõi

• Việc chuyển đổi sang nền tảng DISCO có thể làm giảm phí nạp dữ liệu một lần trong ngắn hạn. Ban lãnh đạo cho biết tác động này vẫn không đáng kể và kỳ vọng các khoản phí định kỳ cùng các vụ việc lớn hơn sẽ bù đắp theo thời gian.
• Hoạt động rà soát ban đầu dựa trên AI có thể đối mặt với chi phí thấp hơn và xu hướng bình dân hóa gia tăng. DISCO cho biết cơ hội chính của họ là chuyển đổi công việc hiện đang do con người rà soát thủ công sang giải pháp AutoReview dựa trên phần mềm.
• Ban lãnh đạo bày tỏ sự tự tin vào khả năng đạt mức tăng trưởng trên 20% ở mảng kinh doanh cốt lõi, nhưng không xác định cụ thể quý nào sẽ đạt được tốc độ đó.
• Việc đầu tư vào giải pháp tranh tụng hợp nhất đã làm tăng chi phí nghiên cứu và phát triển được vốn hóa trong quý 3. Giám đốc Tài chính (CFO) cho biết điều này phản ánh hoạt động phát triển sản phẩm mới chứ không phải do thay đổi chính sách kế toán.

Điểm nhấn phiên Hỏi & Đáp với chuyên gia phân tích

Khi được hỏi về lộ trình đạt mức tăng trưởng trên 20%, ban lãnh đạo đã chỉ ra các yếu tố như các vụ việc lớn hơn, sự mở rộng ở các khách hàng lớn hiện hữu, mức độ ứng dụng AI sâu hơn và tỷ lệ tiếp nhận nền tảng DISCO mạnh mẽ hơn dự kiến. Doanh thu phần mềm đã tăng 1,8 triệu USD so với quý trước, tương đương gần 5%, từ quý 2 sang quý 3.

Về khả năng sinh lời, công ty cho biết kết quả EBITDA điều chỉnh đạt mức dương sớm hơn kế hoạch chủ yếu nhờ tăng trưởng doanh thu, biên lợi nhuận gộp sản phẩm ổn định và sự kỷ luật trong chi phí hoạt động. Việc tăng vốn hóa chi phí R&D gắn liền với quá trình phát triển giải pháp tranh tụng hợp nhất.

Đề cập đến nguy cơ bình dân hóa việc rà soát bằng AI, ban lãnh đạo nhấn mạnh rằng doanh thu từ AutoReview đã tăng hơn gấp 4 lần so với cùng kỳ năm ngoái và cơ hội tiếp cận thị trường chủ yếu nằm ở mảng rà soát do con người thực hiện. Công ty cũng đang sử dụng các dịch vụ và chương trình ứng dụng AI tại chỗ nhằm giảm bớt rào cản triển khai.

Ban lãnh đạo cho biết sự tăng trưởng từ nhóm khách hàng lớn chủ yếu đến từ việc mở rộng quy mô hợp tác với các tài khoản hiện có thay vì thêm mới khách hàng. DISCO cũng đang hỗ trợ các công ty luật là khách hàng của mình quảng bá các dịch vụ tích hợp AI trong nội bộ và tới các khách hàng doanh nghiệp của họ.

Toàn văn cuộc họp báo cáo kết quả kinh doanh

Toàn văn cuộc gọi công bố kết quả kinh doanh

Phần trình bày của ban lãnh đạo

Operator

Ladies and gentlemen, thank you for standing by, and welcome to CS DISCO's Third Quarter of Fiscal Year 2026 Conference Call. [Operator Instructions]

I would now like to hand the conference over to your first speaker today, Bill Roberts with DISCO Investor Relations. Please go ahead.

Unknown Executive

Good morning, and thank you for joining us on today's conference call to discuss the financial results for DISCO's Third Quarter of Fiscal year 2026. With me on today's call are Eric Friedrichsen, DISCO's Chief Executive Officer; Aaron Barfoot, DISCO's Chief Financial Officer; and Richard Crum, DISCO's Chief Product, Technology and Strategy Officer.

Today's call will include forward-looking statements within the meaning of the safe harbor provisions of the Private Securities Litigation Reform Act of 1995, including, but not limited to, statements regarding our financial outlook and future performance, our future capital expenditures, market opportunity, market position, product and go-to-market strategies and growth opportunities and the benefits of our product offerings and developments in the legal technology industry. In addition to our prepared remarks, our earnings press release, SEC filings and a replay of today's call can be found on our Investor Relations website at ir.csdisco.com.

Forward-looking statements involve known and unknown risks and uncertainties that may cause our actual results, performance or achievements to be materially different from those expressed or implied by the forward-looking statements. Forward-looking statements represent our management's beliefs and assumptions only as of the date made. Information on factors that could affect the company's financial results is included in its filings with the SEC from time to time, including the section entitled Risk Factors in the company's annual report on Form 10-K for the year ended December 31, 2025, and the company's upcoming quarterly report on Form 10-Q for the quarter ended September 30, 2026.

In addition, during today's call, we will discuss non-GAAP financial measures. These non-GAAP financial measures are in addition to and not a substitute for or superior to measures of financial performance prepared in accordance with GAAP. Reconciliations between GAAP and non-GAAP financial measures and a discussion of the limitations of using non-GAAP measures versus their closest GAAP equivalent is available in our earnings release.

And with that, I'd like to turn the call over to Eric.

Eric Friedrichsen

Thanks, Bill. Good morning, everyone, and thank you for joining us. I just passed my 2.5-year anniversary at DISCO, and it's been a period of historic change for the legal industry. Just over the past few quarters, we've seen the legal industry go from a technology laggard to 1 that's rapidly being transformed by AI. DISCO has been an innovator in this space, and we're primed to capitalize on the large and growing appetite among law burns and corporations for the best legal AI capabilities. And we're doing so by forging our own path. While many legal tech companies are trying to be everything for everyone, DISCO is hyper focused on being the AI for litigators. We believe that we are uniquely positioned to win this category, thanks to our long-held advantages in data and evidence through our core eDISCOvery product and our latest innovations in building a unified AI solution tailored for the demands of modern litigation.

In short, the legal world is rapidly evolving and DISCO is helping drive that change. Before I dive into the quarter, I am incredibly excited to welcome Mitch Razon as DISCO's new Chief Technology Officer; and Johan Reed as DISCO's new Chief Marketing Officer. As you saw from our press release this morning, the impressive resumes speak for themselves. Both Mitch and Johann are established leaders who are deeply experienced in growing and running large businesses in highly competitive industries. Given the high aspirations we have for DISCO, that perspective and experience will be enormously helpful. As we're building AI for litigators, I am thrilled that mission resonated so strongly with both Mitch and Johann.

Now let's dive into some more detail. DISCO delivered a strong third quarter. We won more of the large and complex matters that sit at the center of our strategy. We grew with our largest customers and we saw AI become a bigger part of how litigators run their matters on DISCO every day. In Q3, total revenue was $45.3 million, up 15% year-over-year. Software revenue was $38.6 million, up 13% year-over-year. And services revenue was $6.7 million, up 21% year-over-year. All of these are adjusted for the contingent matter that we previously reported in Q3 of 2025.

Adjusted EBITDA was $79,000, we reached positive adjusted EBITDA 1 quarter ahead of the Q4 target, and we expect to be adjusted EBITDA positive in Q4. In Q3, we saw a strong performance in the 3 areas that matter most to our near- and medium-term growth. First, continued growth in large matters. Second, expanding wallet share with our largest customers; and third, accelerating usage of AI across our platform. Let's start with large matters because they remained a thread that connects our strategy. We saw another quarter of strong double-digit growth in a number of large matters, which is an important focus because they generate more revenue, they expand as the case develops, and they stay on our platform longer.

In short, large matters are worth significantly more to DISCO over time. They are also where DISCO delivers the most value to our customers, where our AI does is most valuable work and where our services teams continue to make the biggest difference. The rapid uptake of DISCO platform, our new commercial model has become a powerful engine to help drive large matters to DISCO. The number of large matters on DISCO platform tripled over the course of Q3 and Customers tell us that the inclusion of AI on every manner and the straightforward pricing make it easier to bring their biggest cases to DISCO.

More importantly, DISCO platform is changing the way these firms approach their matters. Much Shells is a well-regarded full-service business law firm that is a long-time DISCO customer. They recently moved to the DISCO platform on an enterprise contract and put it to work on their largest construction matter where Cecilia AI has proven immediately useful compared to the general AI legal tools that they tried to use. The manager of eDiscovery noted, I anticipate Cecilia AI will really transform how we work, the more we use it, especially on our larger matters.

Larger matters are also where services and software work hand-in-hand. The biggest, most complex cases demand nuanced, forensic collection, dedicated project management and review expertise and our global team of experts deliver all of it alongside the platform. Our services are important to our customers' success particularly in large matters where the additional resources and expertise make a difference on complex cases and are an accelerant to our software growth.

Services helps us win larger matters and onboard them faster, which is reflected by our service performance this quarter. Services interwoven with our software is the core of our with you and every case value proposition, software and services that evolve alongside the customer and the case and continually add value across the life of the matter.

Alongside large matters, we also made significant progress with large customers. The largest, most sophisticated firms and corporations not only have more large matters, they have consistent litigation needs and frequent spending, so we believe growing this segment gives us a more efficient path to a predictable and durable revenue base. The number of customers generating more than $100,000 over the last 12 months grew to 374 up from 326 a year ago and 354 just last quarter. These customers represent $133 million or 78% of our last 12 months of total revenue. We were also very pleased with the continued growth in AI usage across our platform.

Revenue attributable to generate and Agentic AI capabilities grew over 200% year-over-year and engagement is deepening. Cecilia AI prompts increased approximately 40% during Q3, and the number of customers using Cecilia on multiple matters more than doubled. Advanced Research is another success story. As a reminder, Advanced Research is our next-generation genic tool that automates deep, complex interrogations of the facts to produce litigation quality work product at scale. This new feature is delivering measurable returns for our customers.

A partner at Kutak Rock, an AmLaw 200 litigation leader called Advanced Research, his best associate and noted it stands out among legal AI solutions for the value it adds. He said that, -- now that AI is everywhere, ROI is what matters. We need solutions that are scalable and drive meaningful improvements in how we prepare and take a case to trial. Advanced Research has been a force multiplier on complex matters by handling difficult tasks across huge volumes of data to deliver real insights to our case teams.

In Q3, we made advanced research generally available across our customer base. In that time, it has grown to account for 91% of all Cecilia Q&A usage and our largest matters are seeing the most use. Additionally, auto review has also continued to gain momentum, and we're seeing customers move from trying out a review to building it into how they work. One of our corporate customers recently directed all of its outside counsel teams to use auto review on every manner going forward, absent a compelling reason for an exception. That is a powerful signal of where eDiscovery review is headed.

We are also investing directly in adoption. This quarter, we launched our AI adoption program which puts DISCO litigation engineers on site with customers to work hands on with case teams on advanced research and auto review in their live matters. Alongside it, our Pathways training program continues to expand. At 1 leading international law firm, we trained more than 200 people, most of them attorneys in a single quarter. The more confident litigators are with these tools, the more matters they bring to DISCO. Repeat usage is central to our strategy. And as I mentioned previously, our customers are really feeling the value these tools bring to the litigation teams.

In the near to midterm, we see AI-driven first pass review becoming the default approach as the capabilities of this technology and overall readiness of the market continue to evolve. The quality of AI review has risen so dramatically that there is now more risk and human review and the massive speed advantages mean that our customers see significantly more value for their dollar. Being able to review in 48 hours, what it would take a team of 20 to do in 2 months and with higher quality completely reset the economics of eDiscovery review. This area of e-discovery is forever changing, and we are building auto review to lead that shift.

Speaking of where the world of litigation is going, I also want to speak a bit about our unified litigation solution, which we announced going into pilot with key customers last quarter. The unified litigation solution is an incredible opportunity for DISCO to become the AI solution for litigators. While Richard will cover our progress in a bit, I want to share what we're learning from the litigators using it. The thing we hear most from our pilot customers, and these are some of the largest and most sophisticated firms in the world, is the need for purpose-built technology that can handle the speed and scale of modern litigation, manage their data natively and securely and provide deep contextual insight into both the facts of matter and the controlling law.

One of our pilot customers and Amla top 50 firm with multiple matters in the pilot program, has used our solution in a large medical matter after recognizing that general legal AI tools had too many limitations. Two things brought them to our unified solution. First, our ability to easily manage the size and the complexity of the matter, and second, the ability to apply our advanced AI capabilities to both the facts and the law with any single secure environment. without relying on external data connectors or separate tools.

All of this adds up to a very exciting story for DISCO. I believe now more than ever that our core business alone can drive DISCO to be a 20%-plus grower. Our core business is accelerating with larger matters larger customers and deeper AI usage. Our unified litigation solution offers enormous upside on top of that momentum as we broaden our capabilities to become AI for litigators and to serve even more needs for litigators across the full litigation life cycle.

With that, I'll turn it over to Richard to dive into some impressive product updates.

Richard Crum

Thank you, Eric. This quarter, we made great strides in our product road map for our core capabilities, our generative and agentic AI capabilities and our unified litigation solution. Let's start with DISCO platform because that is where customers are seeing the impact of our work today. From a product perspective, DISCO platform does something simple but important. It removes friction. With a straightforward commercial model and Cecilia AI included on every matter, case teams get the very best DISCO technology for a single price. Customers are responding so positively. They are buying DISCO platform at rates above our expectations, and they use it to gain a strategic advantage. .

As of the end of the quarter, Cecilia was enabled on 96% of DISCO platform matters, which means AI is part of the entire workflow to give litigation teams the power to deeply understand their evidence and leverage that insight to best serve their clients. DISCO platform is also built for the large matters Eric described. Large matters bring more data and more complex data from the ever-growing list of collaboration and messaging tools and beyond. DISCO is built to handle it all and perform at scale. Our ability to ingest and structure massive data volumes and to do it with the speed and reliability our customers count on when their largest matters are on the line is a massive differentiator.

The next area where DISCO is leading the market is advanced research. In Q3, we made advanced research generally available to all customers. Advanced Research is an agentic AI capability built deeply into our eDiscovery product. It goes well beyond simple question and answer by performing multistep reasoning across the matter's full evidentiary record. -- litigators need to understand what the evidence means for their case, and that is what Advanced Research is built to deliver. The adoption data tells the story.

In October, advanced research accounted for 97% of Cecilia Q&A usage and its use is highest on our largest matters. That's exactly what we expected. The bigger and more complex the record, the more valuable it is to have a system that can reason with you across all of it.

Turning to auto review. As Eric mentioned, we believe AI review will soon be the industry standard, and we are building for that moment. In Q2, we began rolling out an enhancement that simplifies and accelerates the process of creating the inputs needed for effective AI review. That capability is now in the hands of customers, and the results have been very positive. This enhancement reduces the learning curve for AI review, making AutoReview even more appealing.

Now to our unified litigation solution. As a reminder, our unified litigation solution unites matters evidence and the law in a single pane of glass. with native access to the full eDiscovery record, metadata and work product included paired with DISCO's license to the full corpus of U.S. case law court rules, statutes and regulations. It is a powerful combination that puts the litigator in full command of the things that are most critical to winning their matter.

We built our unified litigation solution because more than a decade of innovating in e-discovery has repeatedly shown us that litigation is simply different from other areas of law. It has a specific language, unique workflows at incredibly high stakes. As we've said before, we're not building software for lawyers. We're building solutions for litigators. Litigators need purpose-built technology to handle the scale and complexity of modern litigation. Simply put, litigation isn't a feature you add to your legal AI stack. It's an architecture you build from the ground up and expertise you gained from years of working directly with top litigators on their most critical matters.

Last quarter, we announced our initial pilot for our unified litigation solution to learn quickly on real cases with leading firms, including 4 [ AML ] 50 firms. Fast forward 3 months, and those pilot customers are now actively working on the product on live matters and providing exactly the kind of detailed feedback we'd hope for.

One pilot firm has already successfully used the solution on a matter requiring the scale of the evidence collected, something they told us many of the general purpose legal AI simply cannot do today. In less than a week, they were able to perform a thorough risk assessment and advise their clients on the best legal path forward. Work that would have taken an entire team of associates weeks to accomplish. The firm now has multiple live matters in our pilot demonstrating the value the solution adds. That is exactly the sort of insights the pilot is designed to provide and we're using it to further refine our road map.

We've expanded the pilot with a second cohort of customers and expect to onboard a third cohort before the end of the year. We will share more on broader availability as our plans evolve. One theme that stands out in what we're hearing from our pilot customers is the value of deep context and a true native mastery of the facts because fax is where most cases are on or lost. Much of the industry is reaching a portion of the evidence and the law through connectors such as an API call to a research service and MCP integration into someone else's platform or more often and potentially more risky, a copy of the documents stripped out of the context that lives in the source system.

At their core, many of these connections are a workaround for capabilities a platform can access or build itself. And for customers, it means having to license and maintain multiple solutions to do their work. DISCO doesn't have that problem. We hold both the evidence and the law natively and our unified litigation solution reasons across them together. That is what gives case teams confidence in the output, and it's what our customers tell us they really want a unified litigation solution that was built specifically for their unique and complex needs rather than a solution cobbled together across MCP connections.

To be clear, we will always partner where it makes sense for our customers. But when it comes to the evidence and the law native integration is what substantive case work requires. We don't believe that shallow connections, the void of context and that expose sensitive data to multiple systems are the right things for our customers.

More importantly, as we've learned from our pilot, it is not what our customers want. Taken together, -- we believe DISCO platform, advanced research, auto review and our unified litigation solution shows clear evidence that our AI native stack is compounding. The platform our customers rely on the AI they increasingly built into their work and the data that flows through DISCO every day are what gives us confidence today and in the future.

With that, I'll hand it over to Aaron.

Aaron Barfoot

Thank you, Richard. Q3 results reflect continued progress in the areas Eric and Richard just walked through. When excluding the contingent matter from Q3 2025, total revenue was $45.3 million, up 15% year-over-year at the high end of our guidance range. Software revenue was $38.6 million, up 13% year-over-year and services revenue was $6.7 million, up 21% year-over-year. When including the contingent matter, those year-over-year growth rates were 13%, 11% and 18%, respectively. This contingent matter was initially reported in our Q3 2025 earnings. It was revenue associated with the matter that has been on our platform for more than 5 years and was contingent on the successful outcome of the case.

At the time, we recognized $1.3 million of total revenue from the case $1.2 million of which was related to software. I want to touch on the quarter in more detail. First, DISCO platform was again our biggest driver of software revenue growth. DISCO platform revenue has grown more than 10x compared to Q1 of 2026. We've already exceeded our full year internal target for DISCO platform and as of October 1, DISCO platform is the default commercial model for all new and renewal contracts.

Second, AutoReview continued to scale with total AutoReview revenue growing more than 4x year-over-year. Third, services. Services growth was driven primarily by professional services tied to our largest matters. We view services as an accelerant to software. And in many instances, our services offerings are an important reason customers brought those cases to DISCO. I also want to revisit a dynamic I've mentioned the past 2 quarters. As DISCO platform adoption grows and now that it is our default commercial model, we may see some short-term revenue impact from lower onetime ingest fees.

To date, it continues to have an immaterial impact on the business. We expect that to be more than offset over time by DISCO platform's ongoing fees and by the larger, longer and more complex matters that come with it. In addition to larger matters, DISCO platform is also driving improved consideration and better unit pricing. We believe this is a positive trade-off for the long-term revenue profile of our business and far outweighs onetime ingest fees.

Turning to profitability metrics. As a reminder, unless otherwise specified, references to gross margin operating expenses and net loss are on a non-GAAP basis, and adjusted EBITDA is also a non-GAAP financial measure. Gross margin in Q3 was 77%, consistent with the prior year. Sales and marketing expense was $15.2 million or 34% of revenue compared to 33% in the prior year. Research and development expense was $12.5 million or 27% of revenue compared to 28% in the prior year. The dollar increase reflects continued investment in our unified litigation solution advanced research and auto Review to capitalize on the legal industries AI transformation. General and administrative expense was $7.9 million or 17% of revenue compared to 19% in the prior year.

We are also pleased to announce that we were adjusted EBITDA positive a quarter earlier than expected. Adjusted EBITDA was $0.1 million in Q3, representing an adjusted EBITDA margin just above breakeven compared to negative 1% in Q3 of the prior year and negative 8% last quarter. Net loss in Q3 was $0.3 million or negative 1% of revenue compared to a net loss of $0.6 million or negative 1% of revenue in Q3 of the prior year. Net loss per share was breakeven compared to negative $0.01 in Q3 of the prior year.

Turning to the balance sheet and cash flow statement. We ended Q3 with $96.8 million in cash and short-term investments and no debt, maintaining our strong financial position. Operating cash flow year-to-date through Q3 was negative $15.8 million compared to negative $15.7 million in the same period last year.

Turning to guidance. For the fourth quarter of fiscal year 2026, we are providing total revenue guidance in the range of $45.8 million to $47.8 million and software revenue guidance in the range of $39.4 million to $40.4 million. We expect adjusted EBITDA to be in the range of breakeven to $1.5 million. For fiscal year 2026, we are raising our total revenue guidance to a range of $176.1 million to $178.1 million, reflecting the Q3 beat with software revenue guidance in the range of $149.5 million to $150.5 million.

I want to point out that the midpoint of this revenue range is above the high end of the range we provided at the start of the year. We are updating full year adjusted EBITDA guidance to the range of negative $6.8 million to negative $5.3 million. We continue to expect to be adjusted EBITDA positive in Q4 of this year.

Now I'd like to open the call up for Q&A. Operator?

Operator

[Operator Instructions] Your first question comes from the line of Scott Berg from Needham.

Phần hỏi đáp

Scott Berg

Nice quarter. Eric, I want to start on the demand growth side. Your -- the company's revenue growth has been really stable in the 12% to 14% range the last couple of quarters now. A lot of new products. But how do you think about the pushes and pulls to get to that 20-plus percent level that you all are targeting now that you've seen this kind of higher, more stable level of growth? .

Eric Friedrichsen

Scott. Look, I am more confident than ever that the core business at DISCO can be a 20%-plus grower, if you look at the total revenue growth this quarter, we accelerated from 13% in Q2 to 15% in Q3. It was the best growth quarter that we've had in 4 years. And that's double-digit growth over double-digit growth from a year ago. So obviously, those are top line numbers, which I'm really, really pleased with, but it's not actually the top line numbers that give me the confidence, Scott. It's how we got there. It's the fact that we set our strategy, our core strategy really around wallet share expansion in our largest customers, winning larger matters and driving AI adoption.

And then you layer on top of that, the progress that we made with DISCO platform that is far outperformed our expectations for this year. And when you think about the fact that there's some short-term trade-off with DISCO platform, in terms of short-term revenue. But as that data expands, that means that our revenue grows, which gives us even more downstream opportunities.

So the combination of those pieces alone give me a lot of confidence in predictable and durable accelerated revenue growth. Just software alone grew $1.8 million from Q2 to Q3, which is almost 5% sequential growth quarter-over-quarter. So look, I think all the leading indicators give us a lot of confidence that we will get to be a 20%-plus grower in our core business. I'm not ready to call any kind of a specific quarter at this point where we'll get there. But my confidence that the core business will be the 20%-plus grower is higher than it ever has been. And then the early returns on the unified litigation solution give me confidence that we've got a path to even faster growth beyond that.

Scott Berg

Excellent. Helpful. And then, Aaron, you achieved adjusted EBITDA positive here for the first quarter, as you all mentioned the quarter ahead. I have to notice that your PA expense capitalized take down pretty meaningfully in the quarter relative to kind of every historical period. Was you were reporting of adjusted EBITDA Direct result of change in any sort of capitalization versus expensing R&D. Is there kind of -- or is that just related to something else? .

Aaron Barfoot

Yes. I mean let me kind of unpack the adjusted EBITDA performance for the quarter, right? And I think what you just said, we were very pleased to have come in positive -- and I think, as Eric actually alluded to a second ago, really, it's how we did it in some ways, right? We kind of grew our way into -- we talked about from the beginning, we want to orient towards growth. And so the 15% growth adjusted for the contingent matter is the main driver, right? And then you see the kind of gross margins with our product profitability have kind of stayed in line with historical norms, and we've been disciplined in our operating expenses. And those are really the elements that have driven the profitability you are correct that our R&D has picked up, the capitalization, if you will. It's not from an accounting policy change or from anything else that really is us investing in the unified litigation solution.

So we've been investing, if you think about it all year, we've been investing in advanced research in auto review and really what you saw is in Q3, the investment in the unified litigation pickup. And that's a new product. And so that's something that you would capitalize.

Operator

Your next question comes from the line of B.J. Heim from Canaccord Genuity.

Ryan Shanahan

This is Ryan on for Didi. So I guess pretty recently, we've seen some advances in would be decision models and, I guess, other first-pass review products from other AI native competitors. In the event work costs come down there and first pass if you gets a little bit more commoditized, how do you protect revenue per matter for DISCO.

Eric Friedrichsen

Yes, I'll start off the Ryan and then Richard can feel free to jump in here. But look, we've got multiple revenue streams. We've got our core business. We've got our auto review business, which is for first pass review. And then we've certainly got medium- to longer-term opportunity with the Unified litigation solution. As far as first pass review, I mean, keep in mind, this is a several billion dollar market and DISCO has never produced more than about $10 million in managed review or human-based review, and our auto review is growing significantly.

So We, first of all, have a very differentiated product capability that our customers absolutely love and they're starting to use it time and time again. But we're also building on frankly, not that big of a base considering the enormous size of that market. So I've got a ton of confidence that auto review for first pass review can be an important driver for us in the future. .

Richard Crum

Yes, I'll build on that just to express why we're so confident. And you saw some of it in the data that shared in our prepared remarks. Aaron noted that total auto review revenue grew 4x year-over-year. Eric pointed out a customer who's now mandated the use of auto review across all of its outside counsel. And so to Eric's plan, we know we have a strong fit with our customers with the product we have today, and we keep making it better. .

We rolled out an enhancement over this last quarter that really helps teams speed up the way in which they create the inputs for an AI-based review. -- and that both lowers the learning curve so more customers can adopt it but also broadens the fit of where auto review first pass makes sense for both law firms and corporate customers and then you layer the services that we put on top, and Eric talked about the AI adoption program, which puts real AI experts, right, in partnership with our customers to give them the confidence that they can continue to do it. And so -- as we pointed out, we fully expect that software auto review will take over a lot of what humans are doing today, and we're really set up well for as that unfolds.

Eric Friedrichsen

I think that's the key, really, Ryan, is it's not it's not the competition from a technology perspective that is -- we're thinking a lot about. It's the human review. That's where the real opportunity is to -- for services to become software. That's where the tremendous upside is.

Ryan Shanahan

Yes, that's helpful. And then a quick follow-up. A really impressive quarter with the large customer additions and the cat platform growth. Is there any way you could maybe just attribute how much of that growth is coming from net new customers versus maybe more incremental matters at existing customers?

Eric Friedrichsen

Sure. I mean I could touch on the large customer momentum. First of all, it is a mix of new customers and existing customers, but the majority are really existing customers that have grown -- as you know, we've got -- we've grown now to 374 customers that have spent more than $100,000 with us over the last 12 months. That's up from 326 a year ago and up from 354 just last quarter. So there's significant improvement in terms of that group of customers. And a lot of it is because we're focused on those accounts where we already have a relationship, where there's a tremendous amount of upside -- many of these customers that spend $100,000 or more with us per year, we've got 15% to 20% of their wallet share. So we are very much focused on becoming more strategic, winning their larger matters and gaining more and more of that upside. Now of course, we're acquiring customers along the way. But much of our focus is on expansion of the relationships within our existing customers.

Richard Crum

Think of it -- I guess 1 more way to think about this, Ryan, is that our law firm customers are really a channel for us. And so they've asked us to help them to sell DISCO within their own law firms. And they've asked us to help them sell their services empowered by AI capabilities to their corporate clients to help them expand their own legal services business which is a much more strategic position than we've ever been in the past. And so we've put programs recently in place that really are focused around enabling our champions within these law firm customers to help expand their business and ours. .

Operator

At this time, there are no further questions. I would now like to turn the call back to Eric Fredrichsen, CEO for closing remarks.

Eric Friedrichsen

Thanks very much, everyone. I am extremely proud of our team members here at DISCO. It was just excellent execution in a really pleased and happy with our customers for trusting us with their most important legal matters. Look, it was a great quarter getting to 15% revenue growth, the best that we've had in over 4 years getting to adjusted EBITDA positive a quarter earlier than we talked about several quarters ago are things that I think our team should be really proud of.

I think more importantly, my confidence that DISCO and our core business alone can be a 20%-plus grower is more than it ever has been. And if you think about the progress that we're making with the Unified litigation solution, that opens up a path to even larger and faster growth. Q3 just was confirmation that we're making real progress on the drivers that are going to get us there. We're winning more large matters, growing with our largest customers, we're seeing AI becoming more and more central to the way litigators are working, which is really giving us we're earning that trust from our customers to think about DISCO to do much, much more. So that momentum puts us in a position of strength. I believe that, that strength, combined with the tremendous assets that we have when you think about our core platform, our AI capabilities, our comprehensive access to the U.S. case law.

And then more than a decade of experience on building innovative solutions on top of those assets puts us in a position to win in this new category that we're defining called AI for litigators. So I'm more excited than I ever have been. We're having a lot of fun, and I really appreciate your time today. We'll see you next quarter.

Operator

This concludes today's call. Thank you all for attending. You may now disconnect.