Teleconferência de Resultados do 2º Trimestre do AF2027 da Workday (WDAY): ARR de IA se Aproxima de US$ 600 Milhões

Principais Destaques

• A Workday relatou receita de US$ 2,649 bilhões no segundo trimestre fiscal de 2027, um avanço de 13%, incluindo receita de subscrição de US$ 2,471 bilhões, alta de 14%.
• Os produtos de IA geraram mais de US$ 100 milhões em novo ACV e representaram mais de 25% do ACV novo total. O ARR de SKUs de IA aproximou-se de US$ 600 milhões, saltando mais de 200% na comparação anual e mais de 20% em relação ao trimestre anterior.
• Mais de 5.500 clientes estavam utilizando pelo menos um agente orgânico da Workday, uma alta de mais de 35% em relação ao trimestre anterior. Mais da metade das novas conquistas líquidas de clientes incluiu uma ou mais soluções de IA.
• As obrigações de desempenho remanescentes atuais (cRPO) atingiram US$ 9,03 bilhões, alta de 14,2%. O backlog total de receita de subscrição aumentou 8%, alcançando US$ 27,4 bilhões.
• A margem operacional não-GAAP foi de 31,1%. A Workday elevou sua projeção de margem operacional não-GAAP para o ano fiscal de 2027 para 31%, mantendo sua estimativa de fluxo de caixa operacional em US$ 3,45 bilhões.
• A administração prevê receita de subscrição para o ano fiscal de 2027 entre US$ 9,94 bilhões e US$ 9,95 bilhões, representando um crescimento de 13%. Também espera que a margem operacional não-GAAP se expanda em pelo menos 2 pontos percentuais no ano fiscal de 2028.

Resultados Financeiros Principais

A Workday recomprou US$ 1,3 bilhão em ações durante o trimestre, concluindo seu plano de recompra de US$ 5 bilhões seis meses antes do previsto. Em seguida, o conselho aprovou um novo programa de recompra de ações sem prazo determinado de US$ 4 bilhões.

Os resultados GAAP incluíram um benefício fiscal não recorrente de US$ 374 milhões relacionado a uma transferência interna de propriedade intelectual.

Desempenho Operacional e dos Negócios

A IA foi o principal impulsionador de crescimento no trimestre. O ARR de SKUs de IA atingiu quase US$ 600 milhões, enquanto mais de 5.500 clientes utilizaram os agentes orgânicos da Workday. A administração afirmou que a IA também está sustentando as taxas de vitória da plataforma principal, em vez de substituir a demanda por produtos consolidados de HCM e finanças.

O uso de agentes continuou a se expandir. Mais de 30 milhões de candidatos interagiram com o agente de aquisição de talentos, que agendou automaticamente mais de 8 milhões de entrevistas. O agente de implantação da Workday teve mais de 4.600 clientes e quase 24.000 usuários, com o volume de consultas subindo quase 500% durante o 2º trimestre.

O Sana Enterprise também registrou adoção inicial. Após um lançamento interno em 4 de agosto, a Workday relatou 12.000 usuários ativos e 22.000 agentes personalizados criados em três semanas. Os usuários ativos mensais do Sana aumentaram quase 190% na comparação anual, e a IA gerou 78% dos cursos publicados.

O Adaptive Decision Intelligence atingiu a disponibilidade geral em 31 de julho, quatro meses após o início do desenvolvimento do conceito. A Workday informou que 170 clientes já haviam adquirido o produto. O Adoption Agent também alcançou a disponibilidade geral, atraindo mais de 200 adesões de clientes em seus primeiros três dias.

A Workday continuou abrindo sua plataforma para sistemas externos de IA. O Data Cloud, que fornece acesso sem cópia (zero-copy) a dados de RH e finanças, fechou seis contratos durante o 2º trimestre, todos para sua versão premium Pro Edition. A administração afirmou que o Data Cloud e as ferramentas Agent Ready permanecem previstos para disponibilidade geral no 3º trimestre fiscal de 2027.

Mais de 1.600 clientes usam a Workday por meio de serviços de terceiros, incluindo Teams, Slack, Copilot e Gemini. Aplicações personalizadas construídas na Workday aumentaram mais de 90% na comparação anual, enquanto desenvolvedores criaram mais de 3.000 aplicações e agentes personalizados após a DevCon.

Geograficamente, a administração destacou a forte execução na América do Norte, especialmente nas grandes empresas dos EUA e no Canadá. A região EMEA entregou um crescimento sólido, com a IA representando quase um terço do novo ACV na região. O Japão também registrou um trimestre forte.

Projeções da Administração

A administração prevê que o crescimento do cRPO no 3º trimestre fiscal enfrentará um efeito de comparação proveniente da aquisição da Paradox, que adicionou mais de um ponto percentual ao crescimento do período no ano anterior.

A Workday continua a priorizar o investimento em IA e em sua plataforma principal, enquanto utiliza a adoção interna de IA e a disciplina de custos para apoiar a expansão das margens. A administração descreveu sua perspectiva preliminar para o ano fiscal de 2028 como uma linha de base e afirmou que nem todos os benefícios potenciais das iniciativas atuais de IA foram incluídos.

Riscos e Áreas de Foco

A adoção de IA está atualmente avançando mais rápido do que a monetização. Mais de 5.500 clientes usam os agentes orgânicos da Workday, mas apenas pouco mais de 200 aderiram aos créditos Flex. A administração disse que o modelo baseado em consumo pode adiar o reconhecimento de receita por meses ou potencialmente por um ano.

A empresa também está usando avaliações gratuitas e períodos promocionais para acelerar a adoção, incluindo acesso de um ano no âmbito do programa Sana Lighthouse. Como resultado, a administração afirmou que continua difícil quantificar a conversão de receita no curto prazo a partir do uso atual dos agentes.

O backlog total de subscrição cresceu mais devagar que o cRPO. A Workday atribuiu a diferença a uma mudança contínua no mix de novos contratos (bookings) em direção a clientes existentes e ao mix de setores das novas contratações líquidas.

A administração também ressaltou que a IA corporativa exige investimento contínuo conforme modelos, protocolos de integração e estruturas de identidade evoluem. A Workday está usando modelos pequenos, modelos de pesos abertos (open-weight) e grandes modelos de fronteira em vez de depender de um único fornecedor.

Destaques das Perguntas e Respostas dos Analistas

Monetização de IA e preços principais: A administração disse que a Workday pode monetizar a atividade dos clientes por meio de seus próprios agentes, APIs, Data Cloud ou aplicações construídas com o Workday Extend AI. Os executivos disseram que não estavam observando compressão de preços no negócio principal e argumentaram que a IA estava ampliando as conversas com os clientes e melhorando as taxas de vitória da plataforma.

Demanda incremental versus substituição de produtos: A administração disse que a IA não está canibalizando os produtos principais. Novos clientes estão considerando a estratégia de agentes da Workday ao selecionar uma plataforma, enquanto clientes existentes podem usar orçamentos de IA separados. No entanto, a mudança em direção ao consumo de créditos Flex significa que o uso pode não se refletir imediatamente na receita ou no cRPO.

Modelos de IA de pesos abertos: A Workday afirmou que já utiliza modelos de múltiplos fornecedores, incluindo modelos pequenos, de pesos abertos e de fronteira. A administração enxerga benefícios potenciais dos modelos de pesos abertos no aprendizado por reforço, especialização, implantação internacional e soberania de dados.

Desenvolvimento da plataforma de IA: A administração descreveu a transição como contínua, e não como um projeto com uma data de conclusão fixa. Os marcos de curto prazo incluem tornar o Sana a tela inicial padrão da Workday, expandir os agentes de finanças e recrutamento, e disponibilizar o Data Cloud e as ferramentas Agent Ready para disponibilidade geral.

Produtos de IA adquiridos: A Workday informou que os recursos do HiredScore e da Paradox estão integrados ao seu agente de aquisição de talentos, enquanto o Sana Learning foi combinado com o Workday Learning. A administração caracterizou essas aquisições como extensões da plataforma unificada da Workday, em vez de produtos isolados.

Transcrição Completa da Teleconferência de Resultados

Transcrição completa da teleconferência de resultados

Comentários da administração

Operator

Ladies and gentlemen, welcome to Workday's Second Quarter Fiscal Year 2027 Earnings Call. [Operator Instructions]

I will now hand it over to Justin Furby, Vice President of Investor Relations. Please go ahead.

Justin Furby

Thank you, operator. Welcome to Workday's second quarter fiscal 2027 earnings conference call. On the call, we have Aneel Bhusri, our CEO; Garrit Kazmaier, our President, Product and Technology; Gabe Monroy, our Chief Technology Officer; Rob Enslin, our President and Chief Commercial Officer; and Zane Rowe, our CFO.

Following prepared remarks, we will take questions. Our press release was issued after close of market and is posted on our website where this call is being simultaneously webcast.

Before we get started, we want to emphasize that some of our statements on this call, particularly our guidance, are based on the information we have as of today and include forward-looking statements regarding our financial results, applications and solutions, customer demand, operations and other matters. These statements are subject to risks, uncertainties and assumptions that could cause actual results to differ materially. Please refer to the press release and the risk factors and documents we file with the Securities and Exchange Commission including our fiscal 2026 annual report on Form 10-K for additional information on risks, uncertainties and assumptions that may cause actual results to differ materially from those set forth in such statements.

In addition, during today's call, we will discuss non-GAAP financial measures, which we believe are useful as supplemental measures of Workday's performance. These non-GAAP measures should be considered in addition to and not as a substitute for or in isolation from GAAP results. You can find additional disclosures regarding these non-GAAP measures, including reconciliations with comparable GAAP results in our earnings press release, in our investor presentation and on the Investor Relations page of our website. The webcast replay of this call will be available for the next 90 days on our company website under the Investor Relations link.

Additionally, the prepared remarks of this call and our quarterly investor presentation will be posted on our Investor Relations website following this call. Our third quarter fiscal 2027 quiet period begins on October 15, 2026, and Unless otherwise stated, all financial comparisons in this call will be to our results for the comparable period of our fiscal 2026.

With that, I'll pass the call over to Aneel.

Aneel Bhusri

Thanks, Justin, and thanks, everyone, for joining us today. It's hard to believe I've been back in the CEO seat for 6 months. Time flies, we're having fun, and I'm having a lot of fun. This is a fun time to be back in the tech industry. Q2 was another strong quarter, and AI played an increasingly bigger role. AI products alone drove more than $100 million of new ACV, which accounted for more than 25% of all new ACV closed in the quarter and is starting to lift our core business, too, including our win rates. We had commercial success with AI in the quarter, including strong early returns with Flex credits, and we expect this to accelerate going forward. But I'm watching more closely is adoption. Today, more than 5,500 customers are using one or more of our organic agents. That's up more than 35% from last quarter.

As we continue to build organic agents on Workday, we're also deeply integrating acquired agents like those from Paradox, HiredScore, Evisort into our platform. And we've signed some of the world's largest brands to our Leidos program, which gives strategic customers Sana Enterprise free for a year. We've rolled it out to our own work made last month, and they're hooked. I expect these customers will be, too, and we know that if we deliver the value, monetization will follow. After two strong back-to-back quarters, I really like the momentum we have going into the second half of the year. Gerrit and our product and technology teams have our innovation engine humming again. While Rob and his team have done a phenomenal job building pipeline that will drive more new business. Dave will cover the numbers shortly, including the impact of AI adoption on our outlook for the second half and fiscal year '28.

I spent much of the past 6 months with customers. Last quarter, I told you I had met a single customers looking to replace Workday with something they're building internally or buying from a start-up. A quarter later, that hasn't changed. The reason is the deterministic rails I've talked about before. Our agents are lawful. They work inside the permissions, policies and business process a company already runs on. That's why our customers can trust them with the work that matters, and that's why we're seeing our organic agents really take off this year. One of those customers is BMO, a top 10 bank in North America. They piloted self-service agent with 500 employees in May and successfully rolled it out to all 55,000 employees in June. SSA gives BMO employees and managers a personal, intuitive way to get HR questions answered and work done faster and easier. And because it's native to Workday, the agent understands employee permissions and which HR policies apply to them.

That built-in context is exactly what makes this a responsible and scalable AI deployment for BMO. BMO is one of many. More customers are moving from pilots to production with our agents. But we know that no one company will build every agent. The future will be opened and CIOs need a trusted platform to connect at all. We're building for that world with Data Cloud and Developer Agent. Data Cloud lets customers use their Workday data alongside the other systems they run on, without copying or moving it. Customers see the value and demand is building. Developer Agent, which we announced at DevCon in June, makes it dramatically faster and easier for developers to build on Workday using natural language. It's one of the innovations I'm most excited about. I'll add just one thing, because I don't think it's fully understood. Workday is not just an enterprise apps company. We're an enterprise context platform. We were built that way from the start.

Agency contacts to do anything useful. Who reports to whom, what the policies are, how money moves. So whether a customer runs our agents or builds their own on top of Workday, we win either way. Gabe will go deeper on this shortly.

Finally, Rising is coming to Las Vegas, October 12 through the 15th. I have to tell you, Gerrit and his team have what me through our rising announcements, and it was loan away by how much innovation we have to show, not slides, live demos and customers talking about work our agents are already doing inside their business. That's the proof I care about most. We'll also be hosting our Financial Analyst Day at Rising, and I hope to see you all there.

To close, as many of you know, I'm an unabashed optimist, but I'm not leaning on optimism here. We're shipping air products. Our customers are adopting them rapidly. It's happening across the board from the agents we built ourselves to the ones we've acquired. This is Workday moment and never felt better about where we're headed.

With that, I'll hand it to my Vulcan Mine Mill friend, Gerrit.

Gerrit Kazmaier

Thanks, Aneel, and hello, everyone. In Q2, once again, we have accelerated our road map and added new AI capabilities across our platform. All of it builds on our unique growth model of work, and we keep moving more agents into production at scale. So let me put this into numbers for you. Nearly $600 million in ARR from our AI SKUs and that is up more than 200% year-over-year and up over 20% from last quarter.

Here is what this looks like in practice. In recruiting, more than 30 million candidates interacted with our talent acquisition agent and the automatically scheduled more than 8 million interviews last water alone. In functions like procurement and legal Workday Contract Intelligence Agent drove nearly 70% year-over-year growth in agreements processed. And NHR case resolution, more than 100 customers move Sana self-service agent into production last month. Since February, we have nearly doubled the tools and skills in the agent that run in production.

So now let's talk about three key innovation highlights from last quarter, Sana, Adaptive Decision Intelligence and our deployment and adoption agents. Let's talk about Sana Enterprise first. Sana Enterprises our AI workbench [indiscernible] HR, finance and IT to build, orchestrate and run AI across the entire end rise. In Q2, we have shipped 23 new capabilities in Sana that open up what customers can do with it. For example, our brand-new Sana Agent Builder lets any user create in action in natural language and then run it on a schedule or trigger it on a business event. Every one of those agent runs in a secure sandbox.

We now have one shared agent task inbox in Sana, where work is managed in one place, human or AI. And we have added custom MCP connectors. So Sana can connect to any third-party application out there. This innovation, it's already changing AI adoption inside and outside of Workday. Here is what we have done. We rolled out Sana Enterprise internally on August 4 and already have 12,000 active users on it. These workmates have built 22,000 custom agents in 3 weeks alone. Now this is what AI looks like when it's embedded into real work. We have also brought Workday Learning and Sana learn together into one product. It is in GA now with a learning admin agent, AI-based course generation and AI tutor for self-paced learning and MCP support as part of our open platform, and the results are clear.

In Sana, monthly active users grew almost 190% year-over-year and 78% of all published courses were created of AI. And we are taking the next major step to Sana Workday. As we have shared in the past, Sana is our vision for an eye driven work experience across Workday and beyond. With our unique meta-driven platform, we can now unlock all Workday processes for generative UI in Sana. So by work their [indiscernible] [ Gober, ] Sana's conversation AI experience will be the default home screen to Workday.

The second big AI innovation is Decision Intelligence. As you know, frontier models are good at analyzing data, but they sit on the sidelines of the operational system, they are detached from business [indiscernible]. They are working off downloaded spreadsheets, and they are unable to close the loop back to work actually happens. Adaptive decision intelligence is different. It is our AI agents for analytics and blending that runs analysis, generates plans and takes government action directly inside Workday, all based under our security model, audit trails and governance rules.

On July 31, Adaptive Decision Intelligence entered into GA. And this was one of the fastest products we have ever delivered from concept to general availability in four months, and it's built organically within Workday. Here is what we have shipped all AI agent-based, life analysis of blender scenario modeling, sensitivity analysis, recression models, version comparison and the recommendations that you can submit as government plan updates with approval, audit trails and staleness detection.

And as of today, 170 customers have already purchased it. And one AI customer told us, it gives finance teams answers they can actually defend. And you're not stopping here. We just kicked off the early adopter program, connecting adaptive decision intelligence directly into Workday Finance, to give it access to general lines, planned data and to our accounting center. Our customers can now tie AI to financial outcomes with transaction level evidence.

And next, we're going to move to the workforce analytics, connecting live people data with financial decisions in One AI agent experience. And our third AI highlight for this quarter is deployment and adoption agent. Deployment age momentum has been dynamite. As of today, we have more than 4,600 customers, almost 24,000 users and query volumes surging nearly 500% in Q2.

Mohegan Gaming is now live, and Mohegan says the downstream impact is priceless. Deployment age and deploys updates so rapidly that employees remain focused on their core business. And our long-term target is reduce deployment time with this agent by 80% to 90%. And once customers are live, adoption agent helps keeping them ahead of what's new. It surfaces the releases that matter and recommends the next step for their environment. Adoption agent entered GA just last week. We had more than 200 customers signed on in the first 3 days alone. With 20,000 release notes evaluated and achieved a 94% customer satisfaction rate. One customer solution Health used adoption agent on their latest release. They ran it across every functional area, and they have cut that release time down by 70%. And here is why this matters so much faster deployments, faster adoption and a lower cost to serve a feeling Workday goes momentum in the medium end price.

Workday go win rates and deal volume are both increasing in Q2. And here is one more note for you. Travel agent and Sana for ITSM have both moved into early adopter with the first customers already boarded. And the bottom line of this is this. The broad model of FERC is not a thesis. It is in production, it is compounding and you can see it in the numbers. A key part of this question is our open platform strategy, unlocking the power of Workday for our customers and for the entire ecosystem around us.

I'm going to pass it to our very own Gabe Monroy, Workday's CTO and the leader of our technology platform to share our progress here. Gabe, take it away.

Unknown Executive

Thanks, Gerrit. It's good to be here. Building on what Aneel said, it's almost as the Workday was built for AI 20 years ago. Having one day to model, one security model, one version for every customer from the start is an incredible advantage, and it's a big reason why I came to Workday. Over the last several years, we've modernized that foundation to handle the complexity of running AI at scale. We built tools that can query Workday using industry-standard analytics solutions and an MCP layer, so AI models can talk to our APIs in natural language. But access is only half of it. The stakes are different in the world of people and money. That's why we've updated our security model and built the agent system of record. So our agents act with the same permissions, rules and controls our customers already trust.

And Asian isn't just the model. It's identity, permissions, guardrails, logs and governance, all working together. That's what separates a lawless agent from a lawful one. A lawless agent will attempt to take action on its own. A lawful agent takes action because the system allows it. Inside Workday's deterministic rails, every step is checked against the security model, the business process framework and the compliance logic before it runs.

Now in Q2, we advanced our platform strategy in three ways, making it more open, extensible and lawful. Starting with being open. CIOs aren't going to settle for a single rigid AI stack. So we built Workday with three paths in. Developers building custom agents on their own AI stacks need safe governed access to Workday, both for querying data and for taking action. Data Cloud gives external agents and partners like Snowflake, AWS and Google, zero-copy access to HR and finance data. In Q2, we closed 6 deals with all customers choosing our premium Data Cloud Pro Edition. Data Cloud remains on track for GA in Q3. And at DevCon launched agent ready tools over MCP. These agent APIs allow third-party agents on any stack to safely get work done inside Workday, like updating an employee record or approving an invoice without breaking corporate rules.

Agent Ready tools are now available to early adopters through Workday Extend Pro with GA in Q3. Today, more than 1,600 customers use Workday across third-party services like Teams and Slack as well as Copilot and Gemini, which we added in Q1. Over 100 of those customers are already leveraging self-service agent directly in the tools they use every day. Whether a manager approved spend or logs and expense right in chat the underlying policy checks in compliance stay securely anchored inside Workday. And some work simply outgrows the chat bubble, travel and IT service management, which Gerrit mentioned earlier, they'll need a canvas. That work was in Sana, where we own both the reasoning engine and the experience. Whether a customer uses their own agent uses our agents with their own front door or Sana, we meet them wherever they work. And because all of it runs through Flex credits, we monetize every action happening on our rails, regardless of the path.

In Q2, custom apps built on Workday spend grew over 90% year-over-year. Even when building required specialized engineers. Our new developer agent removes that bottleneck. Now builders can generate policy-compliant workflows using plain language directly inside tools like cursor, codecs and quad code. This takes development time from weeks to minutes and opens up the platform to all builders. Customer uptake was intent. Developers have built more than 3,000 custom apps and agents since DevCon.

As we expand these plain language tools across our user base, we fundamentally shift our business model. You don't have to be an engineer to build on Workday. Finally, none of this matters if the transactions are not lawful. In Q2, we launched Asian Passport. It gives security teams a verified auditable record that an agent was tested against critical risks before it deployed and is monitored after. Cisco joined as a launch partner, bringing Cisco AI Defense to test agents against leading standards and protect them in production. We'll add more security partners and out of station stamps in coming months, giving CISOs the confidence required to write higher value work through Workday.

In short, our customers can work where they want build in plain language and trust every execution.

With that, I'll hand it to Rob.

Robert Enslin

Thanks, Gabe, and hello, everyone. Our customers trust Workday with the most important parts of their business, and we see it in the field every day. Companies on legacy HCM and ERP systems are realizing they can't get value from AI with modernizing their core. They want one platform they can trust with security and reliability, which is a big driver behind the demand for Sana Enterprise. AI is amplifying the value of the software stack customers already trust. You can see those tailwinds in our Q2 results. With more than 65% of the Fortune 500 running on Workday, we continue to bring on some of the world's leading brands.

In Q2, we formed new relationships with companies such as KPMG U.S., Danske Bank, BWX Technologies and Guidehouse. In medium enterprise, as you heard from Gerrit, Workday is taking off. Customer volume increased more than 5x over Q1. Across large and medium enterprises, AI is a key reason companies are modernizing their Core in Workday. In fact, over half of our net new wins in Q2 signed up for 1 or more AI solutions, and we're seeing even faster AI adoption across our base. As customers leverage our unmatched HR and finance context to truly unlock the value of AI in the enterprise.

Genesis is using our new financial audit agent, a Chief Accounting Officer described it as a first step towards the dream of getting a touchless audit. Seminal hard rock services is using payroll agent, automate complex back and compliance calculations for 28,000 employees. In [indiscernible] investment, one of our first 20 customers back in 2007 added Sana Enterprise to power an AI layer over the employee experience. This bolds on the other AI agents they recently added, including recruiting, contract intelligence and planning. Sana Enterprise had an exceptional launch in Q2. New customers included AstraZeneca, Novartis, Caterpillar and Delivery Hero Group. Since we combine Sana Learn with our core learning management system, we've seen a sizable jump in our overall learning business, which more than tripled quarter-over-quarter. Adaptive Decision Intelligence helped drive strong performance across the entire planning business and our ecosystem is moving fast with it. Partners have helped build more than 100 industry-specific use cases in a little over a month.

In late May, we made Sana Workday and Sana Health Service available to all our customers on our AI terms of service. That drove a surge in the number of customers that have signed universal main service agreement, which gives them access to our agents and our AI capabilities. That strategy clearly worked. More than half of our customer base has already migrated to the UMSA, and that momentum is accelerating. Our focus now shifts to driving adoption through Flex credits. We've already signed 200 customers this quarter, and we expect to significantly grow that number in the second half as R2 adds more GA agents and [indiscernible] platform and data cloud capabilities. We continue to see strong execution across the globe. North America, our largest market, had an exceptional quarter, anchored by U.S. large enterprise and another strong quarter in Canada. EMEA drove solid growth, the strong performance in France, Germany and the Nordics. AI now accounts for nearly 1/3 of new ACV in EMEA.

Japan also had a standout quarter, further proof that our continued investment there is paying off. I'm proud of the results our teams delivered across the business in Q2. Let me close where I started with trust. For more than 20 years, organizations have trusted Workday with their most critical work and now with the AI. That's a durable advantage, and it sets us up for an even stronger second half.

Now over to Zane.

Zane Rowe

Thanks, Rob. Good afternoon, everyone, and thank you for joining us. Building on Rob's remarks, our second quarter results reflect the continued strength of our platform as organizations rely on Workday to power their most critical HR and finance operations.

Subscription revenue in Q2 was $2.471 billion, up 14%. Professional services revenue was $178 million, resulting in total revenue of $2.649 billion, growth of 13%. Looking at our results by geography. U.S. revenue totaled $1.97 billion, an increase of 12%, while international revenue was $682 million, up 17%, benefiting from stronger performance over the last few quarters.

Turning to backlog. 12-month subscription revenue backlog or CRPO, ended the quarter at $9.03 billion, an increase of 14.2%. Growth was again fueled by expansion within our existing customer base with AI increasingly a driver alongside a steady contribution from new logos.

Total subscription revenue backlog ended Q2 at $27.4 billion, up 8% from a year ago. The year-over-year growth rate was impacted by a continued mix shift towards customer-based bookings versus net new and the mix of industries that drove our net new bookings. Gross revenue retention remained strong at 97% for the quarter and net expansion from existing customers once again led to about 60% of our subscription revenue growth. Non-GAAP operating income was $824 million for the quarter, representing a non-GAAP operating margin of 31.1%, driven by a combination of revenue outperformance and ongoing cost discipline. Our GAAP results in the quarter included a $374 million nonrecurring tax benefit related to an internal IP transfer. Operating cash flow totaled $520 million in the quarter, and free cash flow was $460 million. The year-over-year decline was impacted by the timing of the payroll calendar, which had an additional payroll run this Q2.

We repurchased $1.3 billion of shares during the quarter, completing the $5 billion repurchase plan that we discussed at our Financial Analyst Day last September, 6 months ahead of our target. Buybacks will continue to be an important part of our capital allocation philosophy, and our Board has recently approved a $4 billion open-ended share repurchase program. We ended the quarter with cash and marketable securities of $3.4 billion. Our headcount as of quarter end stood at 20,896 workmates around the globe.

Turning to our outlook. We are pleased with our first half results, and we now expect FY '27 subscription revenue of $9.94 billion to $9.95 billion, growth of 13%. For the third quarter, we expect subscription revenue of approximately $2.515 billion, growth of 12%. We expect Q3 CRPO growth of 11% to 12%. We lapped the Paradox acquisition in the third quarter, which added over 1 point to last year's Q3 CRPR growth. For Q3, we expect professional services revenue of $175 million. And for the full year, we expect $710 million. We continue to prioritize investment in AI alongside strategic investments in the core while driving efficiencies across the business. With that, we're increasing our FY '27 non-GAAP operating margin guidance to 31%.

For the third quarter, we expect a non-GAAP operating margin of approximately 30%. We expect to continue expanding margins while positioning ourselves for future growth. We expect Q3 GAAP operating margin to be approximately 18 percentage points lower than our non-GAAP operating margin. And the full year FY '27 GAAP operating margin to be approximately 18 to 19 points lower. Our FY '27 non-GAAP tax rate estimate remains 19%.

We are maintaining our FY '27 operating cash flow outlook of $3.45 billion, and we continue to expect FY '27 capital expenditures of approximately $270 million, resulting in free cash flow of $3.18 billion [indiscernible] 15%. As our Q2 progress demonstrates embedding AI across Workday's platform provides a significant opportunity to drive customer value.

[Audio gap]

extend with Data Cloud and our AI agents all of which are seeing great early demand. In addition, we expect our non-GAAP operating margin to expand by at least 2 percentage points next year. We're encouraged by the significant opportunity ahead to continue to deliver long-term earnings and free cash flow growth. We look forward to diving deeper into our platform innovation at our Financial Analyst Day on October 13 in Las Vegas, and we hope to see many of you there.

With that, I'll turn it back over to the operator to begin Q&A.

Operator

[Operator Instructions] And our first question comes from the line of Gabriela Borges with Goldman Sachs.

Perguntas e respostas

Gabriela Borges

Zane, I really appreciate this early look into growth rate exiting '27 and then the longer-term upside potential. There are -- the question I want to ask about business twofold. One is tell us a little bit about how you're thinking about monetization for headless. And the second is, how do you as an executive team, think about the risks that some of the upside opportunities end up coming as a trade-off between some of the core. What I mean by that is do customers end up essentially negotiating hard on the core products or exerting pricing pressure on the core products such that you end up at a similar place even when adding new functionality. So maybe just those pieces together, I would appreciate your thoughts.

Zane Rowe

Yes. So I'll take the first part. On the head list transactions, we are pretty much indifferent whether somebody uses from a profitability perspective, I think it will show on a revenue perspective, whether a company buys our agent uses our APIs, which we get to monetize or goes through data cloud, which we get to monetize. So -- or the other option is to build their own agents using Workday Extend AI. So we feel like we're pretty covered in all ways that AI gets used, and it doesn't take away any opportunities. If anything, it grows our market opportunity than where it has been before. Second part, I'll probably ask Rob to weigh in. What I actually see, Gabriel, is that customers are making 5-year, 7-year decisions on new platforms and AI is a huge decision point for them. And if anything, it's impacting our win rates on the platform, because we're viewed as the much stronger player in terms of an AI vision and AI agents than our legacy competitors are that we all know so well.

Robert Enslin

Yes, Gabriel, from my side. We don't see compression on the call. What I see is customers starting with the AI discussion and wanting us to be the AI platform for agentics on HCM and on finance. And I think that's driving a broader conversation. And a much broader conversation to the value we offer. You can see it actually in the amount of uptake on the agents that we started to announce and bring to the far where customers -- this is where they actually want to see where HR is going and where finance is going in the future.

Aneel Bhusri

Yes. And just to add to what Rob said, our competitors don't have one data model across there are multiple versions of their applications. These legacy companies might have 6 or 7 different versions. And as a result, they can't aggregate the data model to drive the the AI models. And so we are just way ahead of where they can be, not just where they are now, but where they can ever be from that AI perspective in terms of driving outcomes using AI.

Unknown Executive

Hi, Gabriel, I'll just add. The early look into FY '28 is just to give you a target on how we're thinking about it. As you can tell, we have a number of initiatives in place, and we're very enthused by what we're seeing as early indicators in our AI products. Not all of that has been factored into our FY '28 outlook. So were as you can tell, very enthused on the upside there, but I just wanted to ahead of Financial Analyst Day, give you at least a baseline to see to let you see how we're thinking about it.

Operator

And our next question comes from the line of Michael Turrin with Wells Fargo Securities.

Michael Turrin

Appreciate you taking the question. A lot's changed over software over the past couple of months. We've seen the topics shift to the rise of open source, open weight models, some scientific cooperation between the Frontier labs and existing vendors. And Aneel, I'm curious where Workday fits within those discussions, if any of that, if any of those shift your view on where to focus, and I want to also give you a chance to respond to just some of the recent headlines around private equity interest, given we've all been feeling a number of questions there as well.

Aneel Bhusri

I obviously can't comment on the latter. On the former, we're going to do what's in the best interest of our customers. And so I'm going to ask Gerrit to talk about how we think about the different models, and we work with all of them, but we're trying to do what's best for our customers, both from our performance but also from a cost perspective.

Gerrit Kazmaier

As Aneel has said, right, we are really focused on driving the right ROI and economics for our customers. So as of today already, we deploy a large set of models for multiple vendors. We have small models, open rate models and large frontier models, all in the set of models that we use to build our AI systems and agents with. And quite frankly speaking, we are really excited about open weight models. So if they open up new opportunities for us for having own reinforcement learning, building on adapts over them they give us a much stronger optionality when you think about international and sovereignty. So it's a core tenant for us, as Aneel has said to be more agnostic and because of the different characteristic already, we get great benefit from them. And now as the benchmarks that are getting closer, we have signed the open [indiscernible] initiatives as well as auto companies. we truly see a real big upside for us and our customers to drive better ROI out of our investment. And last [indiscernible], we also started our own research foundation inside of Workday. So we have an own research lab, which specifically actually focuses on getting high accuracy HR and finance, AI systems in place. One of the things that this team is doing right now is actively actually exploring opportunities for us to not only use open weight models, but really to specialize them to our purposes and see what [indiscernible] get out of that.

Operator

And our next question comes from the line of Kirk Materne with Evercore ISI.

S. Kirk Materne

I think this is probably for Aneel or Rob. Obviously, we can -- we hear the enthusiasm around the early progress on AI. I was just kind of curious if you can give us an idea, is AI crowding out some of your other products at this point in time, meaning when you go in and talk to a customer, you obviously want to talk about AI, you obviously want to get AI into their hands. Does that mean sometimes your salespeople have to put something they might have been thinking about a year ago on the shelf for now? And the reason I ask that is because your tone and your enthusiasm is obviously very apparent. But when I look at CRPO just for a proxy, that's more or less in line and Zane gave a guide for early guide in the last. For next year, that seems more or less in line with where people were kind of forecasting. So it seems more substitutive than incremental. So I was just wondering is there something from just a go-to-market perspective where there's a purposeful push with AI, whereas we're not going to try to load everything in every customer right now. Let's get them successful with AI. Even though that might mean we're not seeing as much incremental benefit. I hope that makes sense.

Aneel Bhusri

Yes. I'll start. Thank you for the question, Kirk. So first of all, you're right in that it's early days. The difference on these agents versus traditional applications they iterate and become better so much faster. So I'm very optimistic that we're going to see ramping up usage of these agents faster than we ever saw of our apps. And we're also new to the Flex credit consumption model. And that's a delayed gratification model that is, frankly, newer to Workday again, very optimistic we're seeing great early signs. But I think that's one of the reasons for cautious optimism going into next year because we are aggressively moving towards a hybrid model between subscription and and this consumption model. I don't actually see it as crowding out. I actually see it as customers -- new customers choosing Workday because of our Agentic strategy, existing customers, they have an AI budget. And now we have products that actually fit in that AI budget, which is a big win for us. But again, a lot of these AI products are consumption based. So we won't see the impact from revenue until months or a year down the road. So Rob, what do you want to add to that?

Robert Enslin

Yes, I would add there's a lot of excitement around AI. But our focus has really been about getting adoption. And when you look at bringing an agent, you got to harden them, so you've got to get them adopted. The more customers that go through early access to more customers that are adopting these products just get better really fast. And so that's how our merger has been around how do we get adoption for us. The more customers we have access to our agents, the more they're using those agents, and that's how we're measuring them and moving the company in that direction. And I think you can actually see it in the amount of new UMSAs we've signed. UMSAs have really started to take off. Customers really want that they need to have the UMSA to actually get into the AI world. And then you link that to the data play and the Extend Pro and what you're doing with the Developer Agent. As Aneel said, this is a [indiscernible], but our conversations are really good. And it's in almost every conversation, irrespective of the level of customer speaking to whether it's a C-level executive or the ACM team or the finance team really start to understand that Workday has really entered agentic world in a big way.

Unknown Executive

Kirk, and just to add to that. As we mentioned, we've got 5,500 customers using our organic agents and just over 200, I think, that have actually signed up for Flex credit. So that's the delay that Aneel is alluding to, and that's been factored into our CRPO guide as well as our revenue guide for the remainder of this year, but also our cautious optimism heading into FY '28 and beyond.

Operator

And our next question comes from the line of John DiFucci with Guggenheim Securities.

John DiFucci

My question, I think, for Aneel and Gerrit and maybe Gabe, listen, Workday's approach to AI seems, at least to me, seems sincere and frankly, more thoughtful than some of your peers. But it's also more pragmatic. In other words, it makes sense in my understanding is simply it, AI is going to be part of everything you do, which again, sounds simple, but also an immense task. And frankly, the right move. It sounds like you're certainly on the way from your prepared remarks. But how much effort do you think it takes to accomplish that? And of course, it's a continuous effort. But when do you think you'll get to the point where you can see we've arrived as an AI platform. Is that going to take years, or just curious how you think about that?

Aneel Bhusri

I'll start. And I think both Gerrit and Gabe should weigh in. We are trying to be very thoughtful in the way that we're building our Agentic solutions. They're there meant to solve a anybody can solve the simple problems when I look at what we're doing with the self-service agent is really hard. What we're doing with the financial audit agent is really hard. But by building these agents and coming out the other end, they add tremendous value to our customers and frankly, from a competitive perspective, they're really, really hard for anybody to compete with because they're so deeply embedded in the bowels of Workday. I don't know when we arrive. I think we're arriving now, and I think it just gets better over the next couple of years, but it's all about adoption and customer success. We have to have agents that have real ROI, and that's been the lens we've been looking at it since I came back. We had a lot more agents when I came back. We killed a lot, or we rolled them into [indiscernible] agents. And the ones that we have right now, I'm very optimistic they're all very meaningful to our customers. But let's have Gerrit and Gabe add to that.

Gerrit Kazmaier

Yes, I had a couple of points on the top of the stack, if you will, and then Gabe on the platform side. But to give you a few concrete points on what are the key milestones that we are looking at that we truly believe our landmark moments for Workday as well as for the industry of moving into the AI era and enterprise SaaS. And one of the biggest one for us is how work is going to change in the work experience with Sana. You heard in the prepared remarks that Sana is going to be the default home screen for Workday coming rising this year. Rob spoke about the lighthouse program at Sana Enterprise, and I have shared this changed the world of work inside were there already. We have 24,000 agents being built in just 3 weeks. We really see this coming this rising as a key moment because the phase of SaaS and how work happens is going to fundamentally range. And frankly, we believe it's a stark difference to what that work experience is than what you get from generic jet Copilots because they're deeply tattered into the system of action and a system of work. Secondly, you heard you know that our agents are making a lot of progress. And when we say agents, we really mean AI systems that automate large parts of the value chains in HR and Finance. We just put Decision Intelligence into GA, which is truly a reimagination of how you collaborate with AI on enterprise data. We are about to bring new functionality to recruiting age, and you have heard earlier how much momentum that has already. We are making great progress in one of the areas that Aneel is most excited on financial audit and financial compliance agents, all of these agents are coming out either day out already or coming out of rising at the end of the year. And I think when we come together in the next call back and you look at the momentum we have built until then. I think there will be no question mark left anymore about Workday being in an AI company or not because the world of SaaS will have changed.

Gabe, over to you for platform.

Unknown Executive

Yes. And thanks for the question. In terms of just the time line view on this on the platform side, it's pretty obvious that the AI technology evolution has been at a frantic pace, right? We're seeing rapid and continuous evolution. And I don't really see an end to that. It's going to be continuous, constant evolution. And so the way we're looking at this is as these new integration patterns change, as protocols change, as identity approaches change, we're on a process of taking the new capabilities, pulling them into the platform, driving adoption, as Rob was mentioning, and then delivering outcomes and ROIs. And the key is going to be doing that continuously in a loop. That is the art of building AI systems. And that's going to be a long journey. This is not something that's going to end anytime soon.

Operator

And our next question comes from the line of Alex Zukin with Wolfe Research.

Aleksandr Zukin

Probably another AI question here. But just it's great to hear about the agents and the data cloud opportunities that you're getting front of customers, but maybe just help crystallize us how you're monetizing? And maybe any specifics around like how many Flex credits does an onboarding or procurement agent actually go through, and what that could mean in terms of a net expansion in terms of spend at a customer that's deploying them or like a sales force yesterday talked about in order for customers to unlock AI functionality, they had to upgrade to a premium version that's 60% to 80% more in some cases. How do you see that with some of the MSA agreements we were talking about, Rob? And when should we think about that as a tailwind to numbers? Is that a 2020 -- fiscal '27 dynamics or the following year?

Aneel Bhusri

You said it's a question about AI, but you are hitting all of the cards here from sales of finance to car technology. So maybe let me start and then I hand it over to Rob and probably saying on the outlook question. What we are seeing is that the workload that these agents are driving quite substantial. When you heard about early the numbers and our volumes, we are driving recruiting already. And the key of the ambient agents, the agents that are running in the background, we see a substantial opportunity, but we are not comparing it to software spend. Actually, what we are modeling it against is the labor spend that companies have in those given roles already and think about what is the share of the agent that actually will get basically transferred from a labor spend into an AI agent spend for that specific task group. Right now though, and Rob has said it, so I want to repeat that. Our focus is really on adoption, adoption adoption, right? This is the way how you actually build great AI systems. And so what we are doing right now, and Rob spoke about it, right, we're incentivizing this with programs like the Sana Lighthouse program that we basically allow customers to use it for free in the first year self-service age and one of the most consequential agents we do we put out a promotion that we are not putting on a flexed meter until the end of August or September. So the question that you're asking, how does this translate now to sales and dollars right now, it's difficult to answer because this has not been our focus point. But then you ask me for the potential that we have because of how much work or it actually processes. It is incredibly substantial. And this is why San has said, right, we are so bullish as this being an upside because we can see on the 1 side, the usage increasing, and we sit on the inside that this represents significant monetization opportunity for us.

Robert Enslin

I mean you kind of said that, look, I'm super excited. We focus on the adoption side of things. Now we're starting to focus really on the consumption side of things. So as we move into this, the sales motion changes, and what you see with broad adoption of Sana self-service agents, it's going to rise all the other agents as well because it gives it a completely different look and feel on how you utilize Workday in the future? And what kind of users can work with Workday pretty much anyone. I watched the CFO and a COO look at the the intelligence agent. And it was just -- they were completely blown away. So I think the opportunity for us is really good. The back half of the year looks really positive for us. And I can't see that changing. And we are lockstep in making certain that these agents are to adopt them and consume them. The quality of the agent continues to improve all the time. And I think that's what he meant by it's a circle. It's a loop. And I think we've got a really good process on how to define that loop. So the field is excited by it. Every single one of my customer conversations is really, really interesting. And even customers that I've known for years that are not really in the HCM or finance space. We really want to understand what we're doing and one Workday to lead in the space of HCM and finance. And then lastly, I'd just say, don't say that we've also got agents like talent acquisition agent and a document where we actually measure by different statistics. And that -- if I look at the [ Telemacquisition ] agent, we had more than 30 million candidate interviews in Q2 with 8 million in view schedule. So the numbers are starting to show up, and it will continue to improve over the next months and quarters.

Unknown Executive

Yes, Alex, I would just add, this quarter, we talked about roughly $600 million in AI ARR, which is up from around just over 150 just a year ago. So we expect that trend to continue. As Aneel mentioned earlier on the call, this is all about customer success and customer value, and that's going to be the ultimate driver of our revenue. So we've got some of that built into the back half of this year, as you would expect, but really growing from FY '28 and beyond where the AI component becomes a significant part of our incremental ARR. So we're excited about the future, as you can tell, and we just want to be thoughtful as we monetize it.

Operator

And we will now take two more questions. Our next question comes from the line of Karl Keirstead with UBS.

Karl Keirstead

Maybe I'll direct this one to Zane. Zane, you gave us a preliminary look at 200 bps of margin expansion next year, a little bit more than I was modeling. And actually, a greater pace of expansion than you're guiding to this year. I'm assuming there's no big change in the prioritization on investing. So perhaps you'll probably address this at the Investor Day, but a couple of things that might be driving that.

Zane Rowe

Yes, Karl, thanks for the question. As Anil mentioned earlier this year, this was the year where we intentionally invested heavily in AI and brought in some great talent and feel like we're doing a lot in that area. And I think we've done a credible job prioritizing and then leveraging size and scale and really being thoughtful on how we drive just thoughtful expenses in the future. So it's a continuation of that doing, I think, a pretty good job utilizing AI ourselves internally and expect that trend to continue. So it's really been about the team focusing on what matters and rethinking what we are doing and as importantly, what we're not doing where we're not going to be spending money in the future. And it's that kind of focus that I think is also driving the increase. And we believe that the 2 points for next year is a good starting point. As you know, we've increased to 31% this year, and we believe we can continue to see that improvement as well as leaning in as much as we have in investments in critical areas around AI and our platform.

Unknown Executive

Yes. I would just add, we're leveraging AI internally, and we're doing more with flat headcount. That's the goal. And I think that's a really important direction for us to head.

Karl Keirstead

And our final question comes from the line of [ Samik Chaterjee ] with JPMorgan.

Unknown Analyst

You talked a lot today about the organic agents and the adoption curve you see there. Maybe if you can sort of dive into the acquired acquisitions that you've done and particularly the acquired agents and how you're thinking about the road map there in terms of integration finest of making to a product and then the monetization on that front? And particularly, how should we expect that to impact your fiscal '28 sort of financial outlook as well? Any thoughts on that front would be helpful.

Aneel Bhusri

Well, I'll just say that they're all doing well, but most importantly, they're all deeply integrated already. We don't ever buy technology and just leave it on the outside. We deeply integrate it right away, so we can deliver that unified experience. Maybe, Gerrit, you want to talk about where the products are headed?

Gerrit Kazmaier

Yes, awesome. So it's a great question because actually, we kind of spoke about it earlier, the big products that you can think of is a higher score on the recruiting side in paradox. They actually both are part of our talent acquisition agent now. So they are basically agentic skills as we evolve our AI systems. And both of them have tremendous momentum. I'm going to hand it over to Rob in a second to talk about that. But we are not really thinking about this as, hey, this is separate from the core. As Aneel said, we have a really good M&A regime that we look for strong technical fit. And so as we move forward with Workday recruiting, HiredScore and Paradox, they are an integral part of that, and they're all coming together under the row-based agent for talent acquisition. Again, right this whole idea. This is the embodiment of an actual recruiter with all of these skills. And on the Sana side, which had an absolutely blowout quarter as well gangbuster growth. We put it into GA in a combined product. So Workday Learning and Sana Learning is actually one product now. It's in GA already, and it's driving substantial growth for our customers. But as we take it to them, we don't put rates as Workday and others. This is the learning agent from Workday. This is the recruiting agent from Workday and those acquisitions, they are just widening the breadth of skills these agents have. Rob?

Robert Enslin

Yes. I'd simply say, I mean, we drove more than $100 million of new ACV with our AI products, just 25% of all new ACV. So if you just look at those numbers. And then we -- our ARR numbers are up at 600 -- close to $600 million now. More than half of the new into signed up for AI solutions. So our AI solutions are really paying off. And as Gerrit said, I think it's really important to point out, they are integrated into the Workday Foundation. It's one core, it's one view for our customers, and that allows us to deliver the agents even on top of that.

Operator

And ladies and gentlemen, thank you for your participation on today's conference. You may now disconnect.