UiPath (PATH.US) Q2 conference call: Increased AI penetration rate leverages larger order models irrespective of process orchestration to provide competitive advantage

Zhitong Finance App learned that after the US stock market on Thursday, enterprise automation software company UiPath (PATH.US) announced financial results for the second quarter of the 2027 fiscal year. Financial reports show that in the second fiscal quarter ending July 31, 2026, UiPath achieved revenue of US$410.2.6 million, an increase of 13% over the same period last year of approximately US$363 million, and higher than the general market expectation of US$397.7 million. The company said that revenue growth was mainly boosted by stable demand and weakening foreign exchange resistance. The company achieved GAAP operating profit of $32 million for the quarter and GAAP profit for the fourth consecutive quarter; adjusted earnings per share were $0.15, in line with analysts' expectations of $0.15.

In terms of annual recurring (ARR) revenue, UiPath reported ARR of $1.938 billion, up 12% year over year. The net increase in ARR increased from $31 million in the previous quarter to $37 million, showing an improvement over the previous quarter. Although the absolute size of the net increase in ARR is still not high, management emphasized that the trend of this indicator is “moving upward in a steady manner.”

At the customer structure level, UiPath said that its customer structure is being concentrated towards large accounts. The number of ARR customers reaching $100,000 and above increased 10% year over year to 2,666; the number of large ARR customers reaching $1 million and above increased 21% year over year to 387. The company added that customer attrition was mainly concentrated in smaller customer groups.

Looking ahead, UiPath raised its annual guidance. The company's estimated revenue is between 1,789 billion and 1,794 billion US dollars, higher than the previous guidance range of about 1,776 billion US dollars to 1,781 billion US dollars, and higher than the market forecast of about 1.78 billion US dollars. Non-GAAP operating profit is expected to be approximately $445 million; adjusted free cash flow is approximately $425 million; and gross margin is approximately 84%. Furthermore, the company expects ARR to reach US$2,065 billion to US$2.07 billion for the full fiscal year, which is higher than the previous forecast of approximately US$2.06 billion. Management said the company is expected to achieve the ARR milestone of surpassing $2 billion in fiscal year 2027.

For the third quarter of fiscal year 2027, the company expects revenue to be between $440 million and $445 million, higher than analysts' expectations of approximately $441.8 million; non-GAAP operating profit of approximately $100 million; and ARR is expected to be between $1,992 million and $1,997 million.

In its earnings report, UiPath highlighted the extent to which AI has penetrated its business. CEO Daniel Dines said that 18 of the top 20 deals of the quarter covered AI, highlighting the central position of AI in major customer partnerships. “Use AI where intelligence creates value, and use deterministic automation where accuracy is critical,” he said. This strategy is designed to help customers achieve a higher return on investment at scale while keeping processes under control.

Although the results of the second fiscal quarter exceeded expectations and the full-year guidance was raised, UiPath's stock price fluctuated sharply in after-hours trading. This trend shows that investors' focus has clearly surpassed single-quarter revenue exceeding expectations, and is shifting to the company's ARR growth momentum, AI monetization path, and overall growth prospects.

A series of executive changes were also announced at the same time as the financial report. Hitesh Ramani was officially promoted to Chief Financial Officer, having previously held important roles in the finance team. Ashim Gupta continues as COO. Brad Brubaker has been appointed as Chief Legal and Administrative Officer, and his responsibilities have further expanded upon his original role. Additionally, Kaiser Permanente executive Yazdi Bagli joined UiPath's board of directors.

Financial data shows that UiPath's performance growth is once again accelerating, while operating efficiency and profitability have improved markedly. The company's management pointed out that UiPath's biggest competitive advantage is model irrelated+process orchestration+corporate governance. In other words, customers can freely choose different AI models and AI frameworks without being locked into a specific AI ecosystem. UiPath believes that its core value is not to produce one of the strongest AI models, but to become an “infrastructure/track” AI model for enterprise AI operation, which is responsible for intelligence, while UiPath is responsible for processes, orchestration, automation, and governance. As AI capabilities become stronger, the importance of orchestration ability, governance ability, and model choice will increase. Furthermore, AI has begun to actually enter major customer transactions, indicating that AI has gradually entered the actual procurement of large enterprises from the conceptual and experimental stages. This means that UiPath has the opportunity to increase the contract amount for a single customer and further expand the platform penetration rate. The company's management is also particularly optimistic about coding agents, saying they may become new engines of growth and efficiency.

The following are the details of UiPath's second fiscal quarter earnings conference call.

I. Management statement

Daniel Dines — Co-founder, CEO and Executive Chairman of the Board

We had another strong quarter of execution. ARR grew 12%, the non-GAAP operating margin expanded to 22%, and achieved the fourth consecutive quarter of GAAP profit.

Over the past two years, we've been transitioning for the next phase of growth. We have continuously developed our platform, business process orchestration, agentic (agentic) capabilities, and software testing; significantly improved our go-to-market execution and operational discipline; and re-accelerated the company's pace of innovation.

Today, we're a stronger company, and more and more customers want UiPath to not just automate individual tasks, but to orchestrate complex, long-running, and anomalous business processes and be a key partner in their AI transformation.

We've talked a lot about how AI is changing software in the past. The bigger question now is how can companies turn AI into real commercial value. Customers aren't choosing between AI and deterministic automation. They are choosing the best way to achieve a certain outcome.

AI is very good at reasoning, but it is probabilistic and can be expensive when applied on a large scale. Many business processes don't require reasoning at every step. What they need is accuracy — getting the exact same results every time, while being safe, reliable, and done at the lowest possible cost.

That's why we let our customers choose deterministic automation (deterministic automation) or tokenless automation (tokenless automation) in addition to AI. Our approach is simple: use AI where intelligence creates value, and deterministic automation where accuracy matters. In this way, customers can reap the benefits of AI without having to pay for every step of AI inference, and ultimately achieve better economic efficiency and higher return on investment (ROI) in large-scale applications.

This is what makes UiPath unique. We deliver business results by orchestrating end-to-end processes between agents, robots, API systems, and people, and use the most appropriate technology for each step to achieve the best combination of intelligence, reliability, and cost.

We are also model agnostic (model agnostic), which gives customers the freedom to choose the AI model and technology that best suits their work without being locked into a single ecosystem. As AI expands the scope of automation that enterprises can achieve, we believe this combination of choice, orchestration, and governance capabilities will become more valuable.

Therefore, the opportunity now is to expand the business we have established, expand adoption among our customer base, go deeper within the company, and continue to transform our innovation into lasting growth. And as we scale up, strong execution and synergy between departments within the company become more important.

That's why Ashim Gupta will next focus on his role as Chief Operating Officer (COO). Ashim has been one of my closest partners and one of the leaders most responsible for establishing our financial and operational discipline over the past few years. As COO, he will focus on the company's day-to-day operations, drive greater discipline and consistency across the market entry organization, strengthen cross-functional execution, and lead key strategic priorities within the company's business. As Ashim fully focuses on the company's operations, we will have a planned leadership transition in the finance department, with Hitesh Ramani taking over as chief financial officer.

This is a logical next step, and it also reflects the strength and depth of the leadership team we have built. Hitesh joined us in 2021 as Chief Accounting Officer and has been Deputy Chief Financial Officer for the past two years, working closely with Ashim throughout the finance organization. He was a key partner in every major milestone for the company, including our IPO, and helped establish the financial rigour and discipline we have today. Given Hitesh's current responsibilities and in-depth knowledge of the company's business, we expect the finance organization to have a very smooth transition and maintain a high level of continuity. Meanwhile, Ashim continues to be COO, so he and Hitesh will continue to work closely in their respective positions.

These changes will allow us to focus more on operations and finance, and place two experienced leaders in key positions as the company grows. I'm excited to continue working closely with Ashim and Hitesh, and I'm confident in our current leadership team and our ability to execute future opportunities.

Now back to our quarterly results. We delivered strong second-quarter results, once again exceeding guidelines in terms of both revenue and profit. ARR reached $1,938 million, up 12% year over year, including net additional ARR of $37 million; revenue was $410 million, up 13% year over year. Non-GAAP operating profit increased to $89 million in the second quarter, and the operating margin was 22%, up more than 400 basis points year over year, thanks to improved operational efficiency and strengthened overall business execution discipline.

Behind these results, the strategy we just described is gradually being reflected in customers. Of our 20 biggest deals this quarter, 18 included AI, which shows that AI is increasingly a core component of our biggest customer projects. Customers are expanding from a single automated use case to a wider range of end-to-end processes, adopting more UiPath platform features, and in some cases integrating automation and AI workloads onto the UiPath platform.

We have also seen that when AI is added to transactions, the scale of customer expansion will be even greater. A global insurance company is a great example. The company has carried out a seven-digit expansion project, is modernizing the beneficiary claims process and expanding the use of IXP, Maestro agents, and robots. With the support of UiPath pre-deployment engineers, Maestro integrates file reception, beneficiary analysis, orchestration, exception handling, and human-in-the-loop (human-in-the-loop) work into a unified, end-to-end process. As UiPath is already embedded in its ecosystem, the customer was able to rapidly advance this use case and continue to transform other processes within the organization on the same basis.

In the public sector, the US Department of War has expanded its partnership with UiPath to support military wide clean audit (clean audit) programs. Based on determinism, the department is adding Autopilot, our IDP solution, and test automation to automate critical audit and reconciliation tasks.

We are also seeing that governance capacity and reliability are becoming real competitive differentiators. A leading financial institution chose UiPath over other orchestration vendors as a single platform for end-to-end processes. Maestro is the only solution that can orchestrate between its homegrown applications while meeting its large-scale governance and compliance requirements. Currently, the solution has been deployed in a key revenue channel process, combining deterministic automation with human environmental safety mechanisms.

These are not isolated examples. Whether it's a new customer or an existing customer expansion, we've seen customers standardize on UiPath and integrate various dot solutions into our platform. A leading US regional bank is integrating its entire automation project on UiPath, using Test Cloud for conversion testing, and using intelligent processes in the area of fraud and compliance to help manage risk through an important module. One of Canada's largest financial services companies is partnering with Ashling Partners to migrate its entire automated business to UiPath and plans to use coding agents to drive the migration, with the goal of reducing maintenance costs and time to value. In terms of expanding its business, a Fortune 500 financial services company is moving all of its automation requirements to UiPath through a CIO-led multi-million dollar program while expanding the use of Test Cloud to test investment management software deployed to customers.

What these success stories have in common is integration. As customers begin to consider a combination of automation and AI, we increasingly see that they want a unified platform to build, orchestrate, test, and govern the entire process. I'm very excited about the Coding Agent, the pilot projects, and the results of the implementation so far. Our early results showed that front-end deployment engineers and coding agents were able to reduce the workload by nearly 60%. As we continue to build on this, this will have a transformative impact on our customers' time to value and overall total cost of ownership (TCO).

We see similar potential in a leading American energy company. They are using Cursor and UiPath throughout the automation lifecycle, from architecture and development to testing and code review to production deployment. Coding agents can directly create UiPath workflows, and our platform ensures that the development process is governed and standardized. So it's not just about making AI write code faster. It means making the entire automation lifecycle faster.

This is one of the important reasons why we think AI will expand the automation market. AI will not only create new use cases, but also reduce the cost and effort required to build these use cases. Additionally, to further speed up implementation, we are announcing the launch of a new developer-friendly workflow automation tool, which is currently in public preview. The tool allows developers to use coding agents they are already familiar with, such as Claude Code, Codex, Cursor, and GitHub Copilot, to orchestrate business processes and automate manual tasks through APIs and agents. It combines the speed of native AI development with the governance capabilities required by enterprises.

Our horizontal platform remains a core strength, enabling customers to automate and orchestrate processes across functions, systems, and technologies. More and more, we are combining this horizontal advantage with vertical, outcome-oriented solutions to directly target specific business outcomes to reach buyers across business lines while creating natural entrances for wider platform adoption. We've seen strong growth in our customers this quarter, including a Fortune Global 500 manufacturer. We're modernizing their accounts payable business with the CFO's Office Invoicing Solution to automatically process approximately 700,000 invoices per year.

What really impressed the customer was the results our approach could deliver: 96% document processing accuracy during the proof of concept phase, automated supplier communication, rich operational dashboards, and an estimated 50% reduction in invoice processing time and support effort. In healthcare, a leading US healthcare system chose our chargeback solution to automate medical claims chargebacks in its revenue cycle management process. The solution will help it automate the creation and submission of complaints in its inpatient and outpatient businesses, enable it to recover millions of dollars of claims that were previously unprocessed due to falling below manual review thresholds, and potentially recover significant additional revenue.

WorkFusion further extends this approach to the financial services sector. The integration is progressing according to plan, and we are encouraged by customer feedback and the sales pipeline that is being developed. Its intelligence, built specifically for financial crime and compliance, enables customers to obtain more complete, outcome-oriented product solutions right out of the box. Testing is another area where we continue to expand our coverage, particularly through our partner ecosystem.

We recently expanded our partnership with Cognizant. Cognizant will embed UiPath Test Cloud into its Testing as a Service and many service products to help customers move from human script-based testing to intelligent testing. Cognizant will also help expand customer rollout and adoption of Test Cloud through its global delivery model.

Before we finish, I'm also excited to welcome Yazdi Bagli to our board of directors. Yazdi has deep technical, operational, and corporate transformation experience at Kaiser Permanente, Walmart, and Procter & Gamble, and I'm looking forward to his perspective on UiPath.

Finally, we're looking forward to meeting many friends in Las Vegas next month. We'll kick off with Investor Day on September 22nd, when we'll share more information about our long-term strategy and product roadmap. Subsequently, our annual user conference FUSION will be held from September 23 to 25.

Ashim Gupta — Chief Financial Officer and Chief Operating Officer

I'm incredibly proud of what our finance team has achieved, and I'd like to congratulate Hitesh, who has always been an outstanding partner and leader in our organization. Hitesh and I have been working side by side for many years, and no one is better suited to lead our finance organization than him.

As I fully focus on my role as Chief Operating Officer, I'm looking forward to working closely at all levels of the company to drive consistent execution and help the business scale. A large part of this work is to continue strengthening our market entry execution.

We spent a lot of time segmenting with sales leaders and customer accounts to ensure we deployed the right resources and strategies for the right opportunities, while collaborating with the leadership team to enhance the overall synergy of bringing the breadth of our platform to market. The same focus extends to how we drive adoption and use among our customer base and how we work with partners. These have always been our top priorities, and we're continuing to strengthen the connections between the sales line, partners, and customers to drive business expansion and make it easier for customers to adopt more platform features.

We have a strong leadership team, huge innovation capabilities on the platform, and huge market opportunities ahead. I'm very excited about what we can achieve together.

In a few minutes, Hitesh will give you guidance for the third quarter and the rest of the year, but first, I'll cover the results for the second quarter. Back to this quarter. Unless otherwise specified, I will be discussing non-GAAP performance, and all growth rates are year-over-year. I would also like to point out that since we price and sell in local currencies, exchange rate fluctuations can affect performance. In the future, we will also disclose the incremental impact of exchange rate changes compared to previous guidance, as well as the year-over-year impact.

Second-quarter revenue grew to US$410 million, up 13% year over year. Excluding the year-on-year exchange headwinds of approximately $8 million, revenue increased 16%. This includes an additional $1 million in exchange rate headwinds since we released our last guidance and the first quarter results call. The year-on-year exchange rate headwinds mainly came from the yen, the Romanian leu, and the Indian rupee.

Total ARR was $1.93 billion, up 12% year over year. This includes a favorable year-on-year exchange rate of $1 million, with no additional impact since the guidelines were issued during the first quarter results call. The net increase in ARR was $37 million, up from $31 million in the same period last year. The year-on-year exchange rate trend was mainly due to the euro.

At the end of the quarter, our cloud ARR was approximately $1.3 billion, including hybrid cloud and SaaS, an increase of over 19%. At the end of the quarter, we had approximately 10,350 customers, and customer attrition was still mainly concentrated in the smallest customer segment, while the number of ARR customers over $30,000 increased 6% year over year. This quarter, we signed one of the biggest new customers in the company's history — a top Canadian bank. The bank is looking for a platform that can support its evolution towards an intelligent workflow.

We demonstrated this capability through an intelligent proof of concept for its third-party requirements process, combining agents, robots, people, and systems, all orchestrated by Maestro, and providing the governance and compliance capabilities required for large-scale applications. This win reflects our customer strategy of adding new enterprise customers with huge potential for expansion.

This quarter, we also added additional customers including Flexsteel, [Azul], and Purdue Federal Credit Union. Our strategy is increasingly focused on winning over the world's largest enterprise customers and expanding our business within those customers, and we're seeing this strategy work. The number of ARR customers reaching $100,000 or more increased 10% to 2,666; the number of ARR customers reaching $1 million or more increased 21% to 387.

Our retention metrics also remain strong. The dollar-caliber gross retention rate (DBGR) is still at the best level in the industry, at 97%; the dollar-caliber net retention rate (NRR) is 109%, an increase of 2 percentage points since this year, indicating that the entire business is stabilizing. Excluding exchange rate effects, the net retention rate for the US dollar was 108%.

Back to this quarter. Remaining performance obligations (RPO) increased 14% year over year to $1,378 billion. Excluding exchange headwinds of around $19 million, RPO increased 16%. The current RPO has grown to $901 million, up 14% year over year.

Next, let's talk about fees. Overall gross margin for the second quarter was 82%, and software gross margin was 90%. Operating expenses for the second quarter were $247 million. GAAP operating profit was $32 million. This is our fourth consecutive quarter of GAAP profit, higher than GAAP operating loss of $20 million in the same period last year. GAAP operating profit includes $45 million in stock compensation expenses, compared to $78 million in the same period last year, down 42% year over year. Stock compensation accounted for 11% of revenue, a year-on-year decline of more than 1,000 basis points.

Non-GAAP operating profit for the second quarter was $89 million, and the operating margin was 22%, an increase of more than 400 basis points over the previous year, thanks to our continued focus on operational efficiency. Non-GAAP adjusted free cash flow for the second quarter was $31 million compared to $45 million in the same period last year, mainly affected by tax-related payment schedules.

At the end of the quarter, we had $1.4 billion in cash, cash equivalents, and marketable securities, a healthy balance sheet, and no debt. During the second quarter, we repurchased 2.4 million shares at an average price of $9.63 per share.

Now, I'd like to hand over the phone to Hitesh and ask him to introduce the performance guidelines.

Hitesh Ramani — Deputy Chief Financial Officer and Chief Accounting Officer

Thank you Ashim for your cooperation and guidance over the years. I am happy to take on this responsibility and continue to build on the solid foundation we have built.

Next, let's talk about guidelines. Our philosophy in this regard has not changed. We provide guidance based on what we are seeing and maintain a cautious outlook. We are pleased with the team's execution performance in an ever-changing macroeconomic environment.

Before introducing specific guidelines, I would like to explain that starting this quarter, we will also disclose the incremental and year-over-year effects of exchange rates compared to previous guidance. As Ashim mentioned earlier, our performance is affected by changes in various currencies such as the Euro, Japanese yen, Indian rupee, and Romanian leu.

Next, let's talk about guidelines. For the third quarter of fiscal year 2027, we expect revenue to be between US$440 million and US$445 million. This does not include additional exchange rate effects since the last guidance, but includes the $10 million year-on-year exchange rate headwind. ARR is expected to be between $19.92 billion and $19.97 billion. These include the $1 million exchange rate headwind added since the last guidance, and the $4 million year-on-year exchange rate headwind. Non-GAAP operating profit is expected to be approximately $100 million. We expect the basic number of shares in the third quarter to be around 523 million shares.

For the full fiscal year 2027, we expect revenue to be between $1,789 million and $1,794 million. These include the $1 million exchange rate headwind added since the last guidance, and the $20 million year-on-year exchange headwind, including the $2 million headwind already achieved in the first half of the year, and the $18 million headwind expected to occur in the second half of the year. ARR is expected to be between $2,065 billion and $2,070 million. These include the additional $1 million exchange rate headwind added since the previous guidance, and the $5 million year-on-year exchange rate downturn, including the $10 million smooth wind already achieved in the first half of the year, partially offset by the exchange rate headwind expected for the second half of the year. Non-GAAP operating profit is expected to be approximately $445 million. Finally, we continue to expect non-GAAP adjusted free cash flow of approximately $425 million and non-GAAP gross margin of approximately 84% for the full year of fiscal year 2027.

II. Q&A session

1. Michael Turrin - Wells Fargo Analyst

I'd like to ask a quick question about full-time employees (FTE). It sounds like coding agents are drastically reducing the FTE required for implementation, so how much has the ability to deploy each FTE increased? As customer demand expands, will this change your recruitment plans?

Daniel Dines

Yes, we are still in the verification phase and would like to know how much incremental value the combination of coding agents and FTE can bring. The initial results we have achieved so far are very encouraging, and I believe we are on a positive trajectory. I don't think it has much to do with how many FTEs we plan to hire, but rather to how quickly our customers can realize value. This is also an important technology for our partners, as many of our customers use implementation services provided by partners. We'll keep everyone updated on the progress. This is a very important focus for us in the future, and one of the important tasks of the entire P&E organization is to continuously improve the performance of coding agents on our platform.

2. Bryan Bergin——TD Cowen Analyst

I'd like to know about your current commercialization and monetization methods for intelligent and AI solutions. How did the customer side of the discussion evolve? Also, can you talk about how model costs and tokenomics (tokenomics) influence clients' willingness to sign large-scale transactions for intelligent and deterministic automation?

Daniel Dines

Yes, we continue to see a growing demand from customers for a platform that combines intelligence with accuracy. Our platform is at the best level in the world in terms of process orchestration, task automation, and file handling.

We maintain a fair degree of neutrality with the best intelligent frameworks in the world, such as LangChain, Claude Agent SDK, and CodeX Harness, and others; at the same time, we are also model agnostic. I think this combination is extremely appealing to customers. Essentially, we provide a “track” for business operations, and customers can choose the type of intelligence they need.

Bryan Bergin

OK. My question is, can you talk about the improvements in net ARR added in the second quarter? Obviously, we're trying to figure out how much of this comes from AI-related products. Can you break it down in some way: how much did existing customers contribute to the penetration of new intelligent AI product deployments, and how much did new customers acquired through the full product suite contribute? Of the top 20 deals you just announced, 18 involved AI, and this is certainly encouraging. But if we continue to look closely at the new ARR, do you have anything special to say about the results for the second half of the year implied in your guidelines?

Ashim Gupta

OK, I'll leave this question to Hitesh and let him answer the guide section. Currently, as we discuss regularly, we report ARR by product, Bryan. But the data you mentioned is really encouraging. And we're even more encouraged when we hear customer calls, sales team feedback, and engage with customer executives. The reality is that they are making transactions have a higher ROI, which will result in larger transaction amounts.

Another encouraging point is that we are actually solving bigger, more complex problems. As the world continues to change, I think this will increase the stickiness between us and our customers. So, it actually has two aspects: on the one hand, it allows us to generate more upfront revenue; on the other, it makes us more strategic within our customers. We are very satisfied with the entire product platform and our ability to deliver these results.

Hitesh, if you want, I can talk about the guidelines section.

Hitesh Ramani

Of course, Ashim. As I said, in terms of guidance, our philosophy hasn't changed. We provide guidance based on what we are seeing and take a prudent approach. Regarding the platform, as Ashim said, our platform positioning is very well recognized by our customers. I myself spoke with three clients last week and received a very positive response to every conversation. As we mentioned, 20 of our biggest deals this quarter included AI in 18. We also took this factor into account when considering guidance for the third and fourth quarters.

3. Scott Berg - Needham & Company Analyst

I'd like to start by talking about entering the market and some sales success you seem to be having. You've been talking about improved sales execution over the past few quarters, but it looks like the terminal market you are facing is also showing an improvement in demand. What stage of this cycle do you think you're currently in? Has your current sales execution returned to the level you were hoping to achieve, which is close to 100%? Or do you still think there's still a long way to go before actually being at your best?

Daniel Dines

I think we're currently actually working simultaneously on both ends of the entire spectrum. On the product side, we're taking what I think is the most innovative product initiative in the company's history. We're ready to announce our new ideas on how businesses can adopt AI, orchestration, and automation at a major FUSION event. On the sales side, given market dynamics, I think we're starting to better understand how customers view AI adoption. To some extent, among existing customers, we're seeing less confusion surrounding AI. They're already understanding — I think they have a better understanding of when it's best to use AI, when it's best to use automation, and how the two coexist.

But I can't quite say that about the customer market as a whole. Especially when we're trying to get new customers, the conversation can be different. Overall, we're also seeing an increase in market interest in outcome-oriented trading. This is an interesting area for us. Currently, these transactions are scattered all over the world, but it is likely to become a larger trend. We are watching closely to see how we should work in both directions simultaneously.

Scott Berg

Understood, very helpful. Then, Ashim, when I look at your net revenue retention metrics, they've been pretty stable for the past six quarters. But — and maybe you'll talk about this at the upcoming Analyst Day — how should we look at net revenue retention for some time to come? You still have tons of products to sell to your customers, right? It sounds like your demand environment has definitely improved. My guess is that compared to what we saw a few years ago, customer expansion businesses will begin to recover. But is it likely that this figure will rise above 110% again in a longer period of time? Or is it that for the near future, 100% high ranges such as 108% and 109% are the correct way for us to understand NRR?

Ashim Gupta

No, I mean, that's right, that's exactly our goal. I think the progress we've made is actually excellent. We finished at 106% at the end of last year, so we've increased 3 percentage points as we move towards this goal. Therefore, I would say that the current trajectory is steady upward, which I think is much better than fluctuating up and down. So we feel really good about it.

As far as the point you just mentioned, we have more products that are promoting large-scale applications to customers, as Daniel said. As I mentioned, I think sales execution continues to improve. Frankly speaking, our focus on consumption is also critical in this discussion. So we're actually very happy with this development trajectory. We will discuss this further.

Obviously, we won't be making long-term predictions for these key metrics, but the trend is positive. I would also like to point out that this upward movement and stability is taking place on an increasingly high scale, which indicates that the expansion business in dollar terms is expanding. Those are some of the situations I can provide here.

4. Sanjit Singh -- Morgan Stanley Analyst

I have two questions. The first one might ask Ashim. As we look ahead to third-quarter federal government operations, their fiscal year will end at the end of September. So, I'd like to know about sales pipeline opportunities for federal government businesses and how they are progressing so far. Then a question for Daniel. I think you and I have been discussing which strategies and usage scenarios are currently most popular. I remember you specifically mentioned that software testing is very popular in the market. Are there any other use cases that are starting to gain popularity in Q2, whether industry-specific or cross-industry?

Ashim Gupta

Yes. I think our federal business has performed very well. Joe Perrino is in charge there. I think he and his team were very impressed with us and the entire team, particularly with their strong relationships with customers and government agencies. They're working with some incredible partners who are doing transformative work in the Ministry of War and beyond. At the same time, they also applied and applied some of the experience we have accumulated in the healthcare business to healthcare processes within the government. All of these factors have enabled the sales pipeline to develop very well. We've also done an excellent job in understanding and influencing this market environment. So, we're actually very happy with the way the federal business is developing.

Daniel Dines

Yes. Regarding use cases, we are very excited about the use case sales and vertical solution approach. In addition to testing, we're seeing increasing demand in the revenue cycle management sector, and of course in the area of financial crime, where we're seeing good sales pipelines forming. The CFO office is also an area where we have traditionally been very strong. Additionally, we recently launched a loan origination solution in the financial services sector.

Overall, this is becoming our focus area. Because we believe vertical sales and solution sales can drive the entire product platform. Traditionally, our business model has relied heavily on “land and expand” (land and expand), and this approach can help us continue to advance this model.

5. Jacob Zerbib — William Blair Analyst

You just mentioned that the market's confusion about AI is dwindling, which is very reassuring. Can you talk about how your sales team is adapting to this new environment, especially when it comes to getting big new customers?

Daniel Dines

I think we are doing a lot of educational work in the market to help people understand what kind of boundary exists between what AI is best at and what it is best at accurately executing. As I said in my previous answer, we're changing the way we sell and focus more on sales based on usage scenarios. We started this trend in our US business a few years ago, and have further refined it. We plan to further extend this approach to the entire market entry (GTM) organization.

6. Raimo Lenschow - Barclays Analyst

First of all, Ashim, I wish you all the best. Then I have two questions. Daniel, one of the questions that came up today — which is probably why you need to share, and why we need to share [after the market reacts] — it looks like a new AI model from one of the major cutting-edge AI companies is about to be released, and it is said that it will be much more capable of getting the job done and executing the workflow. I don't want you to answer about this specific model. But how do you view this issue in your conversations with customers? Obviously AI will get more powerful, but you're more in a world of certainty. What do you think of the differences between the workflows you're currently executing and those that you want to hand over to AI or that AI should execute? I know this is another slightly basic question, but this one has come up again, so it would be helpful if I could talk about it again. Then I have another question for Ashim.

Daniel Dines

Listen, I've had a lot of discussions with our clients over the past few months. If you look at AI, it's clear that it's getting more powerful every day. However, AI has an interesting limitation, and I want to point this out in particular: AI cannot learn during work. When you hire an employee, you expect it — you won't give him a manual that says “this is how our business works.” No company can have such a complete manual. Employees learn by reading some documents, but they also learn from others, attend meetings, and talk to customers. It's a continuous learning process. They will be changed by these experiences.

That's not the case with AI. You use the same model for all businesses. Every time you ask an AI a question, you essentially have to provide it with the operation mode of the entire enterprise, that is, the complete operation mechanism of the enterprise. So if you consider this limitation, it becomes increasingly clear that businesses must create what I call a “map of work” (map of work). You have to describe how the business works in a very specific way. At the same time, you must also invest as much resources as possible to build a framework that provides a track for the operation of the enterprise.

In my opinion, everything that can be done through automation and orchestration should be left to automation and orchestration. Because it's accurate, reliable, token-free, and less expensive. And AI actually runs around this corporate framework.

To some extent, you can think of our platform as an enterprise-grade harness. It can control AI and provide AI with all the information it needs to run an enterprise. But all of the customers I talk to want these workflows to be theirs, not the model. The entire “manual” I am referring to is also the customer's own, not the asset of the model company.

So for me, the best combination is actually building an enterprise framework that provides orchestration and automation capabilities, and making that framework a harness around the model. This will provide businesses with the best options.

Raimo Lenschow

OK, OK. Very perfect. Yes, that makes a lot of sense. Then Ashim, if I consider ARR and revenue — or the subscription revenue you're reporting — clearly there's some kind of relationship between the two. Revenue growth seems to have been ahead of the ARR growth we've seen over the past few years. What do you think of this relationship? Especially if you look ahead from now on, how do you think this relationship will develop? I wish you all the best.

Ashim Gupta

Thanks, Raimo. I'm still here, but thank you, I'm really looking forward to working with Hitesh and Daniel. Raimo, from—remember, we have a phenomenon brought about by Accounting Standard 606. As a result, as we pre-sell more complete platforms, the composition of licenses and software changes, especially if our platforms are sold in certain bundles. We'll discuss this further on Investor Day. There is still a slight SaaS headwind here, but depending on the deal mix and where we sell more platforms, this could lead to structural changes between subscription service revenue and license revenue. That's actually it.

So when you look at the overall ARR, as we pointed out in Net New ARR, we're actually happy with the acceleration we're seeing now. I want to emphasize that again. For everyone, our year-over-year performance actually declined last year due to 606 and other factors. We were mostly in the first half of this year — the first quarter was quite stable. And you can see the acceleration from the second quarter results. This really shows you what metrics we think today better reflect the state of the business and its development trajectory.

7. Terrell Tillman - Truist Securities Analyst

Hitesh, congratulations on expanding your CFO role. I have two questions. The first question is about 18 of the 20 largest deals involving some kind of AI product. But I wonder if the situation is basically similar in terms of the initial landing or impact of these deals? Does this involve outcome-oriented monetization? Then I have another question for Ashim.

Hitesh Ramani

I mean, again, 18 of the biggest 20 deals included AI, which basically reflects how excited we see customers being about the platform. We have seen that as long as AI or platforms are part of the transaction structure, the scale of the transaction will naturally be much larger than in other situations. As a result, this development trajectory continues.

Ashim Gupta

Do you have questions for me?

Terrell Tillman

Of course I have it. Yes, I have a harder question for you, Ashim. I'm kidding. You talked about strengthening execution and leading strategic priorities. I'm assuming you have a whole set of things at hand, some of which are easier to achieve and can be done in the short term. Then as you get to the end of the year and continue to develop the product, maybe next year there will be something larger on a larger scale. At this early stage, can you share some areas that you think can have a quick impact and are exciting for you?

Ashim Gupta

I think we're already making an impact fast. I've seen a very significant improvement and transformation in our team's execution, especially in terms of how to quickly launch after sales have been completed. These turnaround times now occur even before the deal is officially closed because our team is moving faster on delivery. The second aspect is coordination among partners, service teams, and FTE teams when we carry out complex implementations. I think these are all areas where we can continue to improve, but we've already seen some “low hanging fruit” being addressed. I'd also like to tell you that I'm very excited about the delivery team and the collaboration between us and the product team. Our CTO Raghu Malpani is very focused on the front line. So as the link between product, delivery, and market entry continues to strengthen, I think this really qualifies us to win when faced with larger, more complex customer issues.

8. Evercore Analyst

You mentioned improvements in sales execution. Can you talk about some of the specific factors driving this improvement? Also, in terms of motivating sales representatives to let customers try out more AI products, have you made any adjustments to the remuneration method?

Ashim Gupta

I think the first thing is actually the frontline team. We have outstanding leaders in every market sector, such as US financial services, US healthcare, public sector, manufacturing, and what we call Summit—that is, industrial and manufacturing companies—and the rest of the world. Many of them have been working in these positions for quite some time. So I think it all comes from their focus in the first place, right? Compared to the role of Daniel, me, and the top leadership team, it is more important to deploy the expertise on the front line and the “customer first” philosophy. We've been trying to cut down on bureaucracy for the past two years, and it's clear that we can still do more. So that's the first point.

Second, I do think cross-functional collaboration between product, sales, and marketing continues to grow. Very fast paced. So how can we deliver better content to sales teams faster and more comprehensively? These are all areas of focus for us, and are driven together by multiple leaders within the company. As for compensation, we will of course use sales compensation as a tool to drive this goal. The reality is that for many customers, there is a need for a wider range of platforms.

Frankly speaking, combining probabilistic automation with deterministic automation is part of what we offer. As we launch new products, we will of course try to provide incentives, whether it's short-term incentive plans (STIPs) or additional incentives when completing sales quotas. We take these steps selectively, and we are very happy with the results. But as the market environment and our product portfolio changes, we still need to continue to do so.

9. Sanika Merchant——RBC Capital Markets Analyst

Congratulations on your results for this quarter. You mentioned the positive developments currently being made in intelligent products. Can you talk about how you are considering pricing for the company's intelligent products in the future, especially after customer adoption of these products begins to expand?

Daniel Dines

We are currently still experimenting with different pricing models for intelligent products. We recently introduced a transaction-based pricing model, which includes all necessary calls required to complete an exchange in our process orchestration. I think we may move more and more towards outcome-oriented pricing, which will include the tokens needed to complete the exchange.

Sanika Merchant

As a question, you talked about the acceleration of ARR, as well as reaching the $2 billion ARR milestone. What do you think are the most important factors that will drive you to reach the upper end of the expected ARR range for FY2027? Also, are there any positives or negatives we should pay special attention to?

Ashim Gupta

I mean, once again, as I mentioned before, we saw a high degree of consistency with our customers, and the platform strategy resonated very strongly with them. In particular, the combination of deterministic ability and intellectualization ability. This is helping us scale up our deals. So this is one of the key elements we're excited about, and it's already included in our guidelines as we consider Q3 and Q4.

10. Keith Bachman -- BMO Capital Markets Analyst

Daniel, I want to ask you this question directly. You mentioned the growing importance of governance and orchestration as customers put AI projects into production. So what I'm asking is where do you see the most pressing needs when you engage with customers today? Do these discussions usually begin with governance and control requirements or with broader orchestration plans?

Daniel Dines

I'd say there is a growing demand from customers to access the breadth of our platform. I think, to some extent, our platform fits very well with what is called “business orchestration and automation technology” (business orchestration and automation technology) in Gartner's Magic Quadrant (Magic Quadrant). So, I don't think customers are bound to wake up and think “I want to buy orchestration.” But I think customers do wake up and think: What's the best platform to help me achieve results, run processes faster, reduce human error, and introduce AI while protecting my intellectual property in some way? I think it's a combination of these factors that is driving current platform demand.