Workday (WDAY.US) Q2 Conference Call: The current focus is to accelerate AI adoption and unlock future growth space with intelligent agents

The Zhitong Finance App learned that the performance guidelines announced by the office management software manufacturer Workday (WDAY.US) fell short of expectations, curbing the upward momentum accumulated recently due to the market's renewed optimism about the software industry. According to financial reports, Workday's second-quarter revenue increased by about 13% year-on-year to US$2.65 billion, slightly better than analysts' average expectations of US$2.63 billion; adjusted earnings per share were US$2.75, which fell short of analysts' average expectations of US$2.61. Subscription revenue for the second quarter was $2.47 billion, up 14% year over year. The subscription revenue backlog for the past 12 months was $9.034 billion, up 14.2% from the same period last year. The total subscription revenue backlog was $27.403 billion, up 8.0% year over year.

However, Workday expects the total subscription revenue backlog to increase 8% year over year to $27.4 billion at the end of the third quarter. It falls short of analysts' average expectations of $28.6 billion; third-quarter subscription revenue is expected to be $2.52 billion, in line with analysts' average expectations.

Workday is part of a group of application software companies that continue to be plagued by investors' concerns about artificial intelligence (AI) disruption. However, as sales remained stable, investors' concerns that emerging AI companies and tools might steal business from traditional software vendors began to ease in the past month. The stock is up 31% over the past month, but is still down about 10% since this year.

The company's management stated during the earnings call that AI is shifting from product innovation to a core growth driver. AI products contributed more than 100 million US dollars to the new ACV, accounting for more than 25% of the new ACV; more than 5,500 customers used self-developed intelligent agents, an increase of more than 35% over the previous month; more than half of the new winning customers purchased AI solutions. Relying on a “deterministic track” and security, authority and compliance systems, the company promotes intelligent agents from pilot to large-scale production, and builds an open enterprise AI platform through products such as Sana Enterprise, Flex Credits, Data Cloud, and Developer Agent. The company's outlook for future performance also shows that AI is expected to be an important driver for future growth and profit expansion.

Here are the details of Workday's Q2 earnings call.

I. Management statement

Aneel Bhusri — Co-Founder, CEO and Executive Chairman of the Board of Workday

The second quarter was another strong quarter, and AI is playing an increasing role. AI products alone brought in over $100 million in new annual contract value (ACV), accounting for more than 25% of all new ACVs added this quarter, and it is also beginning to boost our core business, including our win rate.

We achieved commercial success in AI this quarter, including the strong early performance of Flex Credits (a pay-as-you-go mechanism for AI functionality launched by Workday in 2026), and we expect this trend to accelerate further in the future. What I'm watching more closely is adoption rates. Currently, more than 5,500 customers are using one or more of our self-developed intelligent agents. This figure is up more than 35% from the previous quarter.

While we continue to build our own smart agents based on Workday, we are also deeply integrating acquired smart agents — such as those from Paradox, HiredScore, and eVisort — into our platform. We've also added some of the world's biggest brands to our Lighthouse program, which offers strategic customers a year of free access to Sana Enterprise.

We promoted it to internal Workday employees last month, and they are totally in love with it. I expect these customers to do the same. And we know that as long as we can provide value, commercialization will follow.

After two strong quarters, I'm very optimistic about our momentum as we enter the second half of this year. Gerrit and our product and technology teams have once again set the innovation engine up and running at high speed, and Robert Enslin (President and Chief Commercial Officer) and his team have also done a great job in establishing sales pipelines, which will drive more new business.

Zane Rowe (Chief Financial Officer) will later present specific data, including the impact of AI adoption on our outlook for the second half of the year and the 2028 fiscal year.

I've spent a lot of time talking to customers over the past 6 months. I told you last quarter that I haven't met a single customer who wants to replace Workday with products they've developed in-house or purchased from a startup. It's been a quarter and that hasn't changed.

The reason is the “deterministic trajectory” I've talked about before. Our smart agents are “law-abiding.” They operate within the permissions, policies, and business processes that the enterprise already uses. That's why our customers can confidently hand over important work to them, and why our self-developed smart agents are really rapidly gaining popularity this year.

One of the customers is BMO, one of the top ten banks in North America. They tested the Self-Service Agent for 500 employees in May, and successfully rolled it out to all 55,000 employees in June. Self-Service Agents provide BMO employees and managers with a personal, intuitive way to get answers to HR questions and get work done faster and easier. And because it's native to Workday, this smart agent can understand employee rights and what HR policies apply to them. This built-in contextual information is the key to enabling BMO to deploy AI in a responsible and scalable manner.

BMO is just one of many customers. More and more customers are moving our smart agents from pilots to formal production environments. But we know that no one company will build all the intelligent agents. The future will be open, and the Chief Technology Officer (CIO) needs a trusted platform to connect it all. We're building this future with Data Cloud and Developer Agents. Data Cloud allows customers to use Workday data in conjunction with other systems they are running without copying or moving the data. Customers have seen the value, and demand is growing.

We released the Developer Agent at the DevCon conference in June, which allows developers to use natural language to develop on Workday, and the speed and convenience have been greatly improved. It's one of the innovations I'm most looking forward to.

I'll add one more point because I don't think this is fully understood. Workday is more than just an enterprise application software company. We're an enterprise context platform, and we've built it in this direction from the beginning. Intelligent agents need context to do anything useful — who reports to whom, what policies, how money flows. So whether customers run our smart agents or build their own smart agents on top of Workday, we can benefit from them. Gabriel Monroy (Chief Technology Officer) will go into more depth on this later.

Finally, Workday Rising will be held in Las Vegas from October 12 to 15. I have to tell you all that Gerrit and his team told me about what will be announced at the Rising conference, and I was blown away by the scale of innovation we have — not slides, but real presentations, and customers personally telling us about the work our intelligent agents are already doing within their company. This is the proof I care most about. We'll also be hosting a Financial Analyst Day during Rising, and we hope to see you all there.

Finally, as many of you know, I'm an undisguised optimist, but this time I'm not relying on optimism. We are launching AI products. Our customers are rapidly adopting these products, and this trend covers the entire product system, from our self-developed smart agents to the smart agents we have acquired. This is a critical time for Workday, and I've never been more confident about our future.

Gabriel Monroy — Chief Technology Officer

As Aneel said, Workday is just like it was built 20 years ago for the AI era. Having a data model, a security model, and a single version for every customer right from the start was an incredible advantage and one of the main reasons I joined Workday.

Over the past few years, we've modernized this infrastructure to handle the complexity of running AI at scale. We built relevant tools to query Workday using industry-standard analysis solutions, and also built an MCP layer to enable AI models to interact with our API using natural language.

But the visit was only half of it. In a field involving people and money, the risks and requirements are completely different. That's why we've updated our security model and built the Agent System of Record to enable our intelligent agents to use the same permissions, rules, and controls that our customers already trust.

An intelligent agent is more than just a model. It is a synergy of identities, permissions, protective measures, logs, and governance mechanisms. This is the difference between an “unbound smart agent” and a “law-abiding smart agent.” An unfettered smart agent will try to act on its own. A law-abiding smart agent acts because the system allows it to do so. Within Workday's deterministic trajectory, every step of the operation must be checked against the security model, business process framework, and compliance logic before execution.

In the second quarter, we advanced our platform strategy in three areas: making the platform more open, scalable, and more “law-abiding.” The first is openness. CIOs won't be satisfied with a single, rigid AI technology stack. Therefore, we have built three access paths for Workday. Developers using their AI technology stack to build custom intelligent agents need secure and governed access to Workday, whether it's querying data or taking action.

Data Cloud provides external intelligent agents and zero-copy access to HR and financial data for partners such as Snowflake, AWS, and Google. In the second quarter, we signed 6 deals, and all of our customers chose our Premium Data Cloud Pro Edition. Data Cloud is still officially launched in the third quarter as planned.

At the DevCon conference, we launched an MCP-based tool for intelligent agents. These Agent APIs allow third-party intelligent agents on any technology stack to securely complete work within WorkDay, such as updating employee records or approving invoices, without violating corporate rules. Currently, the tool for intelligent agents is open to early adopters through Workday Extend Pro, and is scheduled to be officially launched in the third quarter. Over 1,600 customers used Workday through third-party services such as Teams and Slack, as well as Copilot and Gemini, the latter two of which we joined in the first quarter. Over 100 of these customers already call Self-Service Agents directly into the tools they use every day.

Whether managers approve expenses directly in chat or record expenses, the underlying policy checks and compliance mechanisms are securely anchored within Workday. And some jobs are beyond the scope of the chat bubble — such as travel and IT service management mentioned earlier by Gerrit, which require a canvas. These jobs will go to Sana, where we have both the inference engine and the user experience. Whether customers use their own smart agents, use our smart agents through their own portal, or use Sana, we can meet them wherever they work. And since all of these operations run through Flex Credits, we're able to commercialize every action that takes place in our tracks, no matter which path the customer chooses.

Custom apps built on Workday Extend grew more than 90% year over year in the second quarter, even though developing these apps in the past required professional engineers. Our new Developer Agent removes this bottleneck. Developers can now generate policy-compliant workflows using natural language directly in tools such as Cursor, Codex, and Claude Code. This reduced development time from weeks to minutes and made our platform accessible to all developers.

Customer adoption was immediate. Since DevCon, developers have built over 3,000 custom apps and intelligent agents. As we roll out these natural language tools to our entire user base, we're fundamentally changing our business model. You don't need to be an engineer, and you can build on Workday.

In the end, if the deal itself isn't compliant, then none of this makes sense. In Q2, we launched Agent Passport. It provides security teams with a verified, auditable record that an intelligent agent has been tested for critical risks before deployment and is continuously monitored after deployment. Cisco joined as an initial partner and introduced Cisco AI Defense to test intelligent agents according to industry-leading standards and protect them in production environments. Over the next few months, we'll be adding more security partners and certification marks to provide CISOs with the confidence they need to run higher value work through Workday. Simply put, our customers can work where they want, build with natural language, and trust every execution.

Robert Enslin — President and Chief Commercial Officer

Our customers entrust their most important business to Workday, and we see this on the front lines of the market every day. Businesses that are still using traditional HCM and ERP systems are realising they can't get value from AI without modernizing core systems. They want a platform they can trust while ensuring security and reliability, which is an important driver of Sana Enterprise's strong demand. AI is amplifying the value of software technology stacks that customers already trust.

You can see these tailwind factors in our second quarter results. Over 65% of Fortune 500 companies are running Workday, and we continue to attract the world's leading brands. In the second quarter, we established new partnerships with companies such as KPMG America, Danske Bank, BWX Technologies, and Guidehouse. In the midsize business sector, as Gerrit just mentioned, Workday Go is growing rapidly. The number of customers increased more than fivefold from the first quarter. In large and medium enterprises, AI is a key reason for companies to modernize their core systems on Workday. In fact, over half of our new winners in the second quarter signed up for one or more AI solutions.

As customers use our unrivaled HR and finance context to truly unlock the value of AI in the enterprise, we're also seeing a further acceleration in AI adoption across our entire customer base. Genesis is using our Financial Audit Agent. Its chief accountant described it as the first step towards the dream of “achieving auditing without human intervention.” Seminole Hard Rock Services is using Payroll Agent to automate complex tax and compliance calculations for 28,000 employees. Crestline Investors was one of our first 20 customers in 2007, and now they have added Sana Enterprise to provide a layer of AI capabilities for the employee experience. This builds on other AI intelligence agents they've recently added, including Agencies, Contract Intelligence, and Planning.

Sana Enterprise had an excellent launch performance in the second quarter. New customers include AstraZeneca, Novartis, Caterpillar, and Delivery Hero Group. Since we integrated Sana Learn with our core learning management system, our overall learning business has grown significantly, more than tripling month-on-month.

Adaptive Decision Intelligence has helped the entire planning business achieve strong performance, and our ecosystem is rapidly growing around it. Partners have helped build more than 100 industry-specific application scenarios over a period of more than a month.

At the end of May, we opened Sana for Workday and Sana Self-Service to all customers in accordance with the AI Terms of Service. This has driven a significant increase in the number of customers who have signed the Universal Main Service Agreement (UMSA), which enables customers to use our intelligent agents and AI capabilities.

This strategy has clearly worked. More than half of our customers have already migrated to UMSA, and the trend is accelerating. Our focus now turns to driving adoption through Flex Credits. We have signed 200 customers this quarter, and we expect this number to grow significantly in the second half of the year as R2 adds more officially launched smart agents and expands our platform and Data Cloud capabilities.

We continue to maintain strong execution around the world. North America is our biggest market. The quarter's performance was excellent. Major US companies performed well, and Canada once again achieved strong quarterly performance. EMEA experienced steady growth, with strong performance in France, Germany, and the Nordic region. Currently, AI accounts for nearly one-third of the new ACV in EMEA. Japan also performed well, which is further proof that our continued investment in Japan is paying off. I'm proud of the results our team achieved throughout our business in the second quarter.

Finally, I want to go back to the “trust” I mentioned at the beginning. Organizations have been handing over the most critical work to Workday for over 20 years, and now they are also handing over AI to Workday. This is a lasting advantage, and it also prepares us for an even stronger second half of the year.

Zane Rowe — Chief Financial Officer

As Robert mentioned earlier, our second-quarter results reflect the continued strength of the platform as organizations rely on Workday to support their most critical HR and financial operations. Subscription revenue for the second quarter was $2,471 million, up 14% year over year. Professional services revenue was $178 million, so total revenue was $2,649 million, up 13% year over year.

Here's a look at the performance by region. US revenue totaled $1.97 billion, up 12% year over year; international revenue was $682 million, up 17% year over year, thanks to further improvements in performance over the past few quarters.

Let's take a look at the contract backlog. The 12-month subscription revenue backlog, or CrPO, reached $9.03 billion at the end of the quarter, up 14.2% year over year. Growth was once again driven by the expansion of existing customers, with AI increasingly becoming one of the driving forces, while new customer contributions remained stable.

At the end of the second quarter, the total subscription revenue backlog reached US$27.4 billion, up 8% year over year. The year-on-year growth rate was affected by continuing structural changes between basic customer orders and net new orders, as well as changes in the industry structure driving net new orders.

The total revenue retention rate (GRR) remained strong this quarter, reaching 97%; the net expansion of existing customers once again contributed about 60% of subscription revenue growth.

Non-GAAP operating profit for the quarter was US$824 million, and the non-GAAP operating margin was 31.1%, mainly driven by revenue exceeding expectations and continued cost discipline. GAAP results for the quarter include one-time tax revenue of $374 million related to internal intellectual property transfers.

Operating cash flow for the quarter was $520 million and free cash flow was $460 million. The year-on-year decline was affected by the timing of the payroll payment calendar, as wages were paid once more this quarter.

We repurchased $1.3 billion of shares this quarter and completed the $5 billion share repurchase plan discussed on Financial Analyst Day in September last year, six months ahead of our target. Stock buybacks will continue to be an important part of our capital allocation philosophy, and our board of directors has recently approved a $4 billion indefinite share repurchase plan.

At the end of the quarter, we had $3.4 billion in cash and marketable securities.

By the end of the quarter, we had 20,896 Workmates worldwide.

Let's take a look at the performance outlook. We are satisfied with the results for the first half of the year. Currently, we expect subscription revenue for the 2027 fiscal year to be between $9.94 billion and $9.95 billion, an increase of 13% over the previous year. For the third quarter, we expect subscription revenue of approximately $2,515 million, up 12% year over year. We expect CrPO to grow 11% to 12% in the third quarter. We'll compare the third quarter to the first comparable quarter after the completion of the Paradox acquisition, which contributed more than 1 percentage point to CrPO growth in the third quarter of last year. We expect professional services revenue of $175 million for the third quarter; $710 million for the full year.

We continue to prioritize investment in AI while making strategic investments in our core business and driving efficiency improvements across the business. Based on this, we raised our FY2027 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 to expand our profit margins while positioning ourselves for future growth.

We expect the GAAP operating margin for the third quarter to be about 18 percentage points lower than the non-GAAP operating margin; the GAAP operating margin for the full year of fiscal year 2027 is expected to be about 18 to 19 percentage points lower than the non-GAAP operating margin. Our estimate for the 2027 non-GAAP tax rate is still 19%.

We maintain our forecast of operating cash flow of US$3.45 billion for fiscal year 2027 and continue to expect capital expenditure of approximately US$270 million for fiscal year 2027, which will generate US$3.18 billion in free cash flow, an increase of 15% year over year.

As evidenced by developments in the second quarter, embedding AI into every aspect of the Workday platform provides significant opportunities to increase customer value. While we're still in the early stages, the market demand for our smart agent product portfolio continues to grow.

We focus on driving customer discovery and adoption through initiatives such as the Lighthouse Program, which we believe will drive subscription revenue growth over time. At the same time, we continue to execute in accordance with the established framework to expand profit margins while driving revenue growth.

Based on this, and ahead of the upcoming Financial Analyst Day, I'd like to share some of our thoughts on next year in advance. Our current target for FY2028 subscription revenue growth is roughly in line with the expected growth rate for the second half of FY2027, or about 11%. We're also seeing potential upside for new products, including Sana Enterprise, Workday Extend with Data Cloud, and our AI intelligent agent, all of which are currently showing very strong early demand.

Additionally, we expect our non-GAAP operating margin to increase by at least 2 percentage points next year. We are excited to continue to achieve long-term profitability and free cash flow growth in the future. We look forward to further explaining our platform innovations at Financial Analyst Day in Las Vegas on October 13th, and hope to see many people there.

II. Q & A session

1. Gabriela Borges, Goldman Sachs Group Research Department:

Zane, I would like to thank you very much for letting us know in advance about the growth rate of the exit phase in the second half of FY2027, as well as the long-term upward potential. There are two aspects to the question I want to ask. First, how do you consider commercializing “headless” (headless) transactions? Second, as a management team, how do you view the risk that some upward opportunities may eventually coexist with some core businesses?

I mean will customers end up bargaining more strongly on the core product or putting pricing pressure on the core product so that even with the addition of new features, the overall results are still at a similar level? So I would be very grateful if you could combine these aspects to share your views.

Aneel Bhusri:

OK, I'll answer the first part first. When it comes to headless trading, we have basically no difference in preference for this from a profitability perspective. From a revenue perspective, whether the customer buys our smart agent — we can commercialize from it; using our API — we can also commercialize; or through the Data Cloud — we can do the same. Another option is for customers to use Workday Extend AI to build their own intelligent agents. As a result, we feel that no matter how AI is used, we can cover it without losing any opportunity. If anything changes, then our market opportunities are expanding rather than shrinking.

For the second part, I might ask Robert to talk about it.

Gabriela, what I've actually seen is that customers have been making decisions about new platforms for 5 or 7 years, and AI is a very important decision factor. If anything, AI is actually increasing our win rate at the platform level because compared to traditional competitors we are all very familiar with, we are considered significantly stronger in terms of AI vision and AI intelligent agents.

Robert Enslin:

Gabriela, from my side, we don't see our core business being squeezed. What I've seen is that customers are starting to switch from AI topics and want us to become their intelligent agent AI platform for HCM and finance. I think this has spurred a broader discussion, and a broader discussion around the value we provide. You can actually see this in the adoption of the smart agents we started announcing and bringing to market. What customers really want to see is where HR will go in the future and where the future of finance will go.

Aneel Bhusri:

Let me add to what Robert just said. Our competitors don't use a uniform data model across multiple app versions. These traditional software companies may have 6 or 7 different versions. As a result, they are unable to aggregate data models to drive AI models. So from an AI perspective, we're way ahead of them too much. Not only are they ahead of where they are now, but from the perspective of using AI to drive business results, it will be difficult for them to reach our level in the future.

Justin Furby — Vice President of Investor Relations

Gabriela, I'll add one more point. We are disclosing the 2028 fiscal year information in advance, just to give you a goal and let everyone understand how we think. As you can see, we've implemented a number of initiatives, and we're very excited about the early metrics AI products are showing right now. Not all of these factors are factored into our 2028 fiscal year outlook. So as you can see, we're very excited about the upside in it. But I just want to give you at least a benchmark before Financial Analyst Day to let everyone understand how we think.

2. Michael Turrin, Wells Fargo Securities Research Department:

The software industry has changed a lot over the past few months. We are seeing a shift in market discussions to an open source model and the rise of open weighting models, and there are also some signs of cooperation between cutting-edge laboratories and existing software vendors. Aneel I'm curious where Workday is in these discussions? Will these changes affect how you think Workday should focus on? Also, I'd like to give you an opportunity to respond to some recent news about private equity interests, because recently we've also received a lot of related questions.

Aneel Bhusri:

I obviously can't comment on the latter question. As for the previous question, we will do what is in the best interests of our customers. So I'll ask Gerrit to talk about how we view the different models. We work with all of these models, but what we strive to do is provide our customers with the best options in terms of both performance and cost.

Gerrit Kazmaier — President of Product and Technology:

As Aneel said, our real concern is to bring the right return on investment and economic benefits to our customers. So as of today, we've deployed a large number of models from multiple vendors. Our model portfolio includes small models, open weight models, and large cutting-edge models. We use these models to build our AI systems and intelligent agents.

Frankly, we're very excited about the open weighting model. They have given us new opportunities, such as doing our own reinforcement learning and building our own adapters on top of these models. As we consider the requirements of internationalization and sovereignty, they also provide us with a stronger margin of choice.

So, as Aneel said, model agnosticism is a core principle for us. Since the different models have different characteristics, we have benefited a lot from it. Now, as the results of various benchmarks are getting closer, we are joining the Open Weights Initiative along with other companies. We've really seen huge upside, allowing us and our customers to get a better return on investment from their AI investments.

Finally, I'll add one more point. We've also set up our own research foundation within Workday. Therefore, we have our own research laboratory dedicated to building highly accurate HR and financial AI systems. One thing this team is currently doing is actively exploring not only using an open weighting model, but actually customizing it professionally for our use, and seeing how much improvement we can get as a result.

3. Kirk Materne, Evercore ISI institutional stock research department:

I think this question might be for Aneel or Rob. Obviously, we can all hear your enthusiasm for the early advancements in AI. I just wanted to know: is AI currently crowding out space for some of your other products?

I mean, when you talk to customers, they obviously want to talk about AI, and you obviously want customers to use AI. Does this mean that sometimes your sales people have to temporarily put aside some projects they might have considered moving forward a year ago?

I'm asking this question because your tone and enthusiasm is very clear. But when I look at cRPO as a proxy metric, it's basically at the expected level. Furthermore, Zane gave initial guidance for next year. Although it is only an early guide, it is also basically in line with the market's previous predictions. So it looks more like an alternative than an incremental one.

So I want to know, from your market entry perspective, is there any kind of strategy to promote AI: in other words, we're not going to try to pack everything into every customer, but let the customer successfully use AI first, even if that means not seeing that many incremental benefits in the short term?

Aneel Bhusri:

Thanks for the question, Kirk. First, you're right, it's still in its early stages. The difference between these smart agents and traditional applications is that they iterate and improve much faster. So I'm very optimistic that we'll see the usage of these smart agents grow faster than any of our previous apps.

At the same time, we are also just starting to adopt the Flex Credits spending model. Frankly speaking, this is a “delayed satisfaction” model, which is also relatively new to Workday. Again, I'm very optimistic. We're seeing very good early signs. But I think this is one of the reasons we are cautiously optimistic as we enter next year, as we are actively shifting to a hybrid model that combines a subscription model with a consumption model.

I don't think this is crowding out the core business. I actually think this is the new reason customers chose Workday because of our smart agent strategy. For existing customers, they have AI budgets. Now we have products that actually fit into this AI budget. This is a major benefit for us. But once again, many of our new AI products are based on consumption models. As a result, we may not see the impact on revenue until months or even a year.

Robert Enslin:

I'd like to add one more point. The market is very excited about AI. But our focus has always been on driving adoption. When you launch an intelligent agent, you have to polish it to maturity, so you have to let customers adopt it. The more customers participate in early access and the more customers adopt these products, the better these products will get better very fast. That's why we've always organized our work around “how to achieve rapid adoption.”

The more customers we have access to smart agents, the more they use them. That's how we measure these products and drive the company in this direction.

I think you can actually see that from the number of UMSA we've signed. Our UMSA has indeed begun to grow rapidly. Customers really need it. They must have UMSA to really enter the AI world. Then connect that to the data business, Extend Pro, and what we're doing with the Developer Agent.

As Aneel said, there's a lag period here. But our customer communication is excellent. No matter what level of customer I talk to, almost every conversation involves that. Whether it's a C-level executive, an HCM team, or a finance team, people are beginning to really understand: Workday has entered the era of intelligent agents in a very important way.

Justin Furby:

Kirk, let me add one more thing. As we mentioned, 5,500 customers are currently using our self-developed smart agents, and I think only slightly more than 200 customers have actually signed up for Flex Credits. This is what Aneel refers to as the lag effect. This has been factored into our CRPO guidance and revenue guidance for the rest of the year, and is also the reason we remain cautiously optimistic as we enter the 2028 fiscal year and beyond.

4. John DiFucci, Guggenheim Securities Research Division:

My question should be for Aneel, Gerrit, and possibly Gabe.

It sounds, and to me, Workday's approach to AI is sincere and, quite frankly, more thoughtful than some of your peers. But at the same time, it's also more pragmatic. In other words, it makes sense. My simple understanding is that AI will be part of everything you do. It sounds simple, but it's actually a huge task, and frankly, I think it's the right thing to do. Judging from the speeches you have prepared, it is clear that you are already on this path.

But how much effort do you think it would take to achieve this goal? Of course, this is an ongoing job. But when do you think you can reach such a stage where you can say “we have become an AI platform company”? Will this take years? I'm just curious what you guys think about this.

Aneel Bhusri:

I'll talk first, then I think both Gerrit and Gabe should get involved.

We tried to build our smart proxy solution in a very thoughtful way. They aren't meant to simply solve problems. Anyone can solve simple problems. When I saw the Self-Service Agent we were working on, it was actually very difficult. The Financial Audit Agent we're working on is also very difficult. But by building these intelligent agents and going through the process, they can add tremendous value to customers.

Frankly speaking, they're also very, very difficult to replicate from a competitive perspective because they're deeply embedded within the core of Workday. I don't know when we actually reached the finish line. I think we're in the process of arriving now. And I think it's only going to get better in the next few years.

But the key is adoption rate and customer success. We must have intelligent agents that generate a real return on investment. It's also the standard I've been using to measure them since I came back. When I came back, the company had more smart agents. We cut down a lot of smart agents and merged some smart agents into larger smart agents. And with the smart agents we currently have, I am very optimistic that they are all of great significance to our customers.

But let Gerrit and Gabe add more.

Gerrit Kazmaier:

If you look at the upper level of the entire technology stack, I'd like to add a few points and then let Gabe talk about it from a platform level. But first, let me give you some specific information about the key milestones we're focusing on. These milestones are not only iconic moments for Workday, but for the entire industry to enter the AI era of enterprise SaaS. One of the biggest milestones is how Sana will change the way we work in the work experience.

As you've all heard in our prepared speeches, Sana will be Workday's default home screen at this year's Rising conference. Rob talked about Sana Enterprise's Lighthouse plan, and I also shared how this plan has changed the way we work within Workday. We've created 24,000 smart agents in just 3 weeks.

We really think this year's Rising Conference will be a critical moment as the SaaS interface and the way work happens will fundamentally change. Frankly speaking, we think this work experience is clearly different from the work experience provided by a generic chat-based Copilot because Sana is deeply connected to the mobile system and the work system.

Second, you've heard that our smart agents are making great progress. When we say “smart agents,” we really mean AI systems that can automate much of the HR and financial value chain. We've just officially released Decision Intelligence. This is actually a reimagining of how enterprise data and AI collaborate. We're about to bring new features to Adopt Agent. Everyone has already heard about the tremendous momentum it has achieved so far.

We've also made significant progress in one of Aneel's most anticipated areas — Financial Audit and Financial Compliance Agents. All of these smart agents are being launched — some are already live, and some will be launched at the Rising conference at the end of this year. I think when we talk again in the next conference call and everyone sees the momentum we've built up until then, the question about whether Workday is actually an AI company will no longer exist. Because the SaaS world has changed by then.

Gabriel Monroy:

If you look at the timeline from a platform perspective, this is pretty obvious: AI technology is evolving at a crazy pace, right? What we're seeing is a rapid and continuous evolution. And I don't think this evolution will come to an end. It will continue to evolve.

So our idea is: as these new integration models change, as protocols change, and as identity management methods change, we'll continue to acquire new capabilities, bring them into the platform, drive adoption — as Rob just said — and then deliver business results and return on investment.

The key is to continuously form a closed loop. This is the art of building AI systems. And it's going to be a long journey. It's not something that's going to end anytime soon.

5. Aleksandr Zukin of Wolfe Research:

This is probably another AI issue. I'm excited to hear you guys talk about smart agents and the Data Cloud opportunities you're promoting to customers. But can you help us be more clear, how are you commercializing it? For example, how many Flex Credits will an Onboarding Agent or Procurement Agent actually consume? How much of a net expansion, or increase in customer spending, might that mean for customers deploying these smart agents? For example, Saifushi mentioned yesterday that in order for customers to get AI features, they must upgrade to a premium version, which in some cases means a 60% to 80% increase in costs.

Rob, what do you think about the MSA agreement you just talked about? When should we look at it as a driver for financial data? Is this a factor for fiscal year 2027 or next?

Gerrit Kazmaier:

You said it was an AI question, but you actually asked all the questions about sales, finance, and core technology. So let me answer that first, then I'll leave it to Rob, and maybe ask Zane to answer questions about the outlook.

What we're seeing is that these smart agents drive a huge amount of work. You've heard our workload data before. In the field of recruitment, for example, we have promoted a lot of work. And environmental smart agents — that is, those that run in the background — we think there's a huge opportunity. But we're not comparing it to software spending. In fact, when we modeled, we were referring to the company's labor expenses in relevant positions. We'll think: In these specific task groups, how much of the workload will shift from labor spending to AI intelligent agent spending?

But now, as Rob said, I want to repeat: our focus is really on adoption, adoption. This is how you actually build great AI systems. So what we're doing now, Rob has already talked about: we provide incentives through projects like the Sana Lighthouse Program. Basically, we allow our customers to use it for free for the first year. The Self-Service Agent is one of our most important smart agents, and we've launched a promotion where we won't be charging Flex Credits until the end of August or September.

So, you're asking the question — “How is this translating into sales and revenue now?” It's hard to answer because that's not our focus right now. But if you ask me how big our potential opportunities are, because these smart agents actually handle a huge amount of work, then the answer is: the scale is amazing. That's why Zane said we're very optimistic about this and see it as an upside opportunity. Because on the one hand, we're seeing increasing usage; on the other hand, we see from internal data that this represents a huge commercialization opportunity.

Robert Enslin:

I mean, you've actually stated the point. We used to focus on adoption. Now we're starting to really shift our focus to consumption. As we move into this phase, the sales model will change. When you see the widespread adoption of the Sana Self-Service Agent, it will also drive every other intelligent agent, because it gives customers a completely different way to see how Workday will be used in the future, and which users can use Workday — basically anyone can. I saw the CFO and COO experience Decision Intelligence Agent, and they were completely shocked. So I think our chances are excellent.

The second half of this year looked very positive for us. I can't see this changing. And we're working closely together to ensure that the quality of these smart agents continues to improve as customers adopt and consume them. I think this is what Gabe called a “cycle, a closed loop.” I think we've established a very good process to define this closed loop. So our sales team is really excited about it. Every conversation I have with my clients has been very, very interesting. Even those customers I've known for years who aren't really part of HCM or finance are eager to know what we're doing and want Workday to lead this direction in HCM and finance.

Finally, I'd like to say, don't forget that we also have talent acquisition agents and document-related intelligent agents, and we measure them based on different metrics. If I look at Talent Acquisition Agent, more than 30 million candidates participated in the interaction in the second quarter, and 8 million interviews were scheduled at the same time. These numbers are already starting to show and will continue to improve in the coming months and quarters.

Justin Furby:

Yes Alex, I'll add one more point. We talked about around $600 million in AI ARR this quarter, compared to just over $150 million a year ago. So we expect this trend to continue. As Aneel said earlier on the call, it all comes down to customer success and customer value, which will ultimately drive our revenue growth.

As you'd expect, we've included some of this in our results for the second half of this year. But the real growth will come from FY2028 and beyond, as the AI component will be an important part of adding ARR at that time. So you can tell we're very excited about the future. We just want to commercialize it in a prudent way.

6. UBS's Karl Keirstead:

Maybe I'll give this question to Zane. Zane, you gave an initial forecast for next year's profit margin expansion of 200 basis points ahead of time, which is slightly higher than what I expected in my previous model. In fact, this expansion is also faster than your guidance for this year. My assumption is that your investment priorities haven't changed significantly. So maybe you'll talk more about this on Investor Day. But several factors may be driving this change.

Zane Rowe:

Yes, Karl, thanks for the question. As Aneel mentioned earlier this year, this is a year where we intend to invest heavily in AI, we've brought in some really great talent, and we think we've done a lot of work in this field. I think we've done a pretty good job in setting priorities, leveraging the size and size of the company, and being very careful about how to drive costs in the future. So it's an ongoing job. I think we're also doing a pretty good job using AI internally, and I expect this trend to continue.

So it really comes down to the team focusing on what's important, rethinking what we're doing, and more importantly, rethinking what we're not doing, and where we won't spend money in the future. I think it was this focus that drove profit margins even higher. We think a 2 percentage point increase next year is a good starting point. As you know, we've raised our margin guidance to 31% this year. We believe we can continue to see this improvement while continuing to invest as heavily as we do now in key areas such as AI and platforms.

Aneel Bhusri:

I just wanted to add one more point. We're using AI internally, and our goal is to get more done while keeping the number of employees largely the same. I think this is a very important development direction for us.

7. J.P. Morgan's Samik Chatterjee:

Today you've talked a lot about self-developed smart agents and the adoption curves you've seen. If so, can you talk more deeply about the AI-related acquisitions you have completed, especially the intelligent agents you have acquired? How are you currently considering their future product roadmap? Specifically, how are they ultimately integrated into products, and how can they be commercialized based on this? In particular, how should we expect these factors to affect your financial outlook for FY2028? It would be really helpful if you could share some thoughts on this.

Aneel Bhusri:

All I can say is that they all performed very well. But more importantly, they're all now deeply integrated. We never buy a technology and then leave it outside. We'll immediately deeply integrate it to provide a unified experience for our customers. Gerrit, maybe you can talk about the future direction of these products?

Gerrit Kazmaier:

This is a great question because in fact we've already talked about this before. Two important products you can think of are HiredScore and Paradox in the recruitment field. They are now actually all part of our Talent Acquisition Agent. In other words, they have now become intelligent agent skills that we have in the process of continuously developing AI systems. Both are currently gaining very strong momentum.

I'll take the question to Rob right away and let him talk about that. But we don't really see them as something separate from our core business. As Aneel said, we have a very mature M&A system, and we look for assets with a strong technical fit. As a result, as we advance Workday Initiatives, HiredScore, and Paradox, they are all an integral part of it. They all eventually converge under role-based talent acquisition agents.

Again, the whole idea is to create a true recruiter personification and equip it with all of these skills. On Sana's side, the quarter was also excellent — growth was very strong. We have officially launched it as an integrated product. As a result, Workday Learning and Sana Learning are now actually one product. It is now officially launched and is driving customers to achieve very strong growth. But when we introduce it to customers, we don't describe it as “Workday and other products.” This is Workday's Learning Agent. This is Workday's Prescribed Agent. These acquired products only broadened the range of skills these smart agents possess.

Robert Enslin:

Yes, I just wanted to say: our AI products brought in over $100 million in new ACV, accounting for 25% of all new ACVs. If you just look at these numbers, you can see what they mean. Then our AI ARR is now close to $600 million. More than half of the new winners in the second quarter signed up for AI solutions. So our AI solutions are really paying off.

As Gerrit said, I think it's very important that these products are integrated into Workday's basic platform. It's a core, a unified view that customers see. This allows us to continue delivering intelligent agents on this foundation.