C3.ai, Inc. Q3 FY2023 Earnings Call

Thank you, and good afternoon, and welcome to C3AI's Earnings Call for the third quarter of fiscal year 2023, which ended January 31, 2023. My name is Reuben Gallegos, and I am the Vice President of Investor Relations. With me on the call today is Tom Siebel, Chairman and Chief Executive Officer; Ed Abbo, our Chief Technology Officer; Juho Parkkinen, our Chief Financial Officer. After the market closed today, we issued a press release with details regarding our third quarter results as well as the supplemental to our results, both of which can be accessed through our Investor Relations section of our website at ir.c3.ai. This call is being webcast, and a replay will be available on our IR website following the conclusion of this call. During today's call, we will make statements related to our business that may be considered forward-looking under federal securities laws. These statements reflect our views only as of today and should not be considered representative of our views as of any subsequent date. We disclaim any obligation to update any forward-looking statements or outlook. These statements are subject to a variety of risks and uncertainties that could cause actual results to differ materially from expectations. Before a further discussion of the material risks and other important factors that could affect our actual results, please refer to our filings with the SEC. Today, all figures will be discussed on a non-GAAP basis unless otherwise noted. And also after the course of today's call, we will refer to certain non-GAAP financial measures and the reconciliation of GAAP to non-GAAP is included in our press release. Finally, at times in our prepared remarks in response to your questions, we may discuss the metrics that are incremental to our usual presentation to get greater insight into the dynamics of our business or our quarterly results. Please be advised that we may or may not continue to provide this additional detail in the future. And with that, let me turn the call over to Tom.

Good afternoon, everyone, and thank you for joining our call today. Two quarters ago, I mentioned economic challenges and extended sales cycles as our customers anticipated a recession. In mid-2022, we experienced a significant downturn in the business environment, impacting decision-making timelines, and I warned that the market downturn could be substantial. Now, as we move into our fourth quarter, we are seeing positive trends with increased business optimism and growing interest in our C3 AI solutions across various industries. This marks a notable shift from our experiences in mid-2022. There is real enthusiasm in the marketplace for our offerings, and overall business sentiment seems to be improving. During the quarter, we confirmed our shift to a consumption-based pricing model, broadened our partner ecosystem, expanded our business pipeline, and achieved leading product innovation in enterprise AI. Importantly, we remain on track to be cash positive and non-GAAP profitable by the end of fiscal year '24. In our third quarter, we reported strong results with total revenue of $66.7 million, surpassing our guidance. Current RPO rose to $176.3 million, and we now have 236 customers. By the end of the quarter, we had nearly $790 million in cash. As we approach Q4, we believe C3 AI is well-equipped to continue investing in growth through enterprise AI innovation and sales expansion while maintaining our path to profitability. The reception to our consumption-based pricing model from partners and prospects has been overwhelmingly positive. Currently, we have more than 290 qualified pilot opportunities in our pipeline, exceeding expectations, and our pilot to production conversion rate is on target. The revenue conversion model we provided last quarter indicates substantial revenue growth rates in fiscal year '24 and beyond. We made significant headway with our partner ecosystem in the third quarter, establishing and expanding our go-to-market partnerships. With Google Cloud, we secured eight new customer deals and expanded our pipeline. Our combined teams are pursuing 291 enterprise opportunities for joint solutions, with over 100 in active licensing discussions. I held a joint meeting with Google Cloud CEO Thomas Kurian and several clients and partners in the U.S. federal region, ensuring all C3 products function optimally in the Google cloud environment. We also expanded our partnership with Google, allowing customers to purchase any C3 AI software solution on the Google Cloud Marketplace. Additionally, we renewed our partnership with AWS, which funded enhancements to our C3 AI law enforcement application for better performance on AWS. C3 and AWS are currently pursuing 75 new opportunities, with 41 highly qualified, and we closed six agreements this quarter. Together with Azure, we secured a deal with a major U.S. energy company and a European technology firm, and we've successfully piloted solutions for a large U.S. defense agency. We formed a strategic alliance with Booz Allen to provide solutions to government defense intelligence sectors, collectively marketing the C3 AI platform and prebuilt solutions. We've already closed our first engagement with the Chief Digital Artificial Intelligence Office and are collaborating with Accenture to drive product innovation and strategic support. We've also partnered with EY to address healthcare needs in the U.K. and entered a partnership with Peraton to modernize the Veterans Administration. We substantially extended our partnership with Baker Hughes this quarter, resulting in an additional $32.5 million in C3 bookings and accelerated payment terms. Our collaboration in the oil and gas sector, aided by partnerships, has led to 87 contracts totaling over $650 million in bookings, with over $350 million in recognized revenue through the third quarter of fiscal year '23. Regarding our ESG solutions, we have made notable advancements with our ESG application, part of our sustainability suite, which includes the C3 AI energy management product. This solution measures and mitigates the energy and greenhouse gas footprint for over six million residences and businesses. We released C3 AI ESG in September 2022, providing comprehensive ESG reporting compliant with various standards. C3 AI ESG features predictive analytics to help organizations track materiality gaps and suggests measures to meet ESG objectives. According to Verdantix, the ESG market will reach $16 billion by 2027, and our product has received positive feedback from initial clients like EY, Shell, and Baker Hughes. Now, turning to our intellectual property, C3 AI has invested significantly in technology innovation, with 26 patents awarded and 90 pending. One key innovation is our model-driven architecture for enterprise AI applications that supports rapid development and deployment of these applications. This architecture also enables integration with ongoing innovations in both the open-source and proprietary domains. We are witnessing a significant surge in the availability of large language models and generative pre-trained transformers that work seamlessly with our platform, enhancing its utility. Recently, there has been much discussion about C3 generative AI. We have developed a novel approach that improves human-computer interaction for enterprise applications by combining our platform's capabilities with generative AI technology. This development is not merely about chat; it focuses on enhancing enterprise search, representing a breakthrough that simplifies usage and explainability of enterprise AI applications. This new capability is being deployed as a critical feature within the C3 AI platform and is being tested at Koch Industries and Baker Hughes. To safeguard our intellectual property, we have multiple patents pending worldwide, and I encourage you to visit our website for more details. Now, regarding guidance, I want to note that this marks our eighth consecutive quarter as a public company where we exceeded our revenue forecasts. We expect Q4 2023 revenue to be between $70 million and $72 million, and for fiscal year '23, we anticipate revenue ranging from $264 million to $266 million. In summary, Q3 was strong, our consumption-based pricing model has been validated, the market potential is vast, business is robust, customer satisfaction is high, our workforce is effective, and we have a promising future ahead. I will now turn the call over to my colleague, Juho Parkkinen, for more financial details.

Thank you, Tom. I will now provide a recap of our financial results, add some color to the drivers of our financials, discuss our expected path to non-GAAP operating profitability by the end of fiscal '24, and I will conclude with some additional color related to the consumption-based revenue model we introduced two quarters ago. All figures will be discussed on a non-GAAP basis, unless otherwise noted. As Tom mentioned, we ended the quarter with revenue of $66.7 million, of which subscription revenue was 85.6%. Gross profit was $51 million and gross margin was 76%. As I mentioned during the last quarter's update, we have a short-term pressure on our gross margins due to a higher mix of clients that carry a higher cost of revenue during the pilot phase of our customer life cycle. Operating loss of a negative $15 million improved year-over-year and was significantly above our guidance due to improved vendor expense management and timing of payment. Operating loss margin was flat at negative 23% as compared to the same period in the prior year. However, on a sequential basis, our operating loss margin improved. Our customer count increased 8% to 236, and we closed 27 deals during the quarter, 17 of which were pilot deals under the consumption model. Now turning to RPO and bookings. We reported GAAP RPO of $403 million, which is down 14% from last year. This was expected as we transition to consumption-based deals. Trade GAAP RPO of $176.3 million is up 3% from last year and 7% on a sequential basis. We continue to see positive trends in pilot bookings diversity as we have increased to nine industry segments in Q3 compared to six in Q2. Regarding our cash flow, free cash flow improved to an outflow of $71.7 million compared to $77 million in the prior quarter. Breaking this down, $19.4 million was related to the build-out of our new headquarters, which we moved into in February. Normalizing for this payment, our adjusted free cash flow improved to an outflow of $52.3 million compared to $54.3 million last quarter. We continue to expand our headquarters and will have additional cash outflow in the following quarters as we take over additional spend. During the quarter, we expanded the Baker Hughes partnership. As Tom mentioned, the changes are designed to provide increased flexibility to Baker Hughes to offer the BH C3 AI solutions to the market. This resulted in the elimination of the variable consideration, which increased the transaction price by $32.5 million. Regarding outlook. As Tom highlighted, we are able to narrow our range as we have more visibility as we enter Q4. As such, we're guiding Q4 to $70 million to $72 million. And for the year, we're tightening the guide from $264 million to $266 million. For Q4 '23, we expect our non-GAAP loss from operations in the range of $24 million to $28 million, that is negative. And for the full year, we expect the non-GAAP loss from operations of negative $69 million to negative $73 million. In accordance with our plan, we do expect gross margin percentage to be negatively impacted by the number of pilots active in the fourth quarter. Please recall that during the two-quarter pilot period, customers will have unlimited run time and premium support resources sufficient to be successful with their target engagement and gain value from the C3 AI software. During this pilot period, our subscription cost of revenue will be elevated until the conversion to consumption-based pricing occurs. Over time, we expect gross margin percentage to revert to historical ranges as a higher percentage of customers on a pilot move to consumption revenue. Our operating margin guidance reflects the fact that we have C3 transform the world's premier enterprise AR conference and our major customer events happening next week, and we will have related marketing expenses for this quarter's costs. In addition, as we have discussed previously, we're increasing our sales headcount to meet the demand we are seeing for our consumption-based pricing sales and expect increased expense in the fourth quarter as a result. It is important to note that we expect that our cash balance will continue to decline into next fiscal year as we complete the build-out of our headquarters. We expect our cash and investments to be at its lowest at around $700 million during fiscal '24. Broadly speaking, as a result of the introduction to consumption-based pricing, we expect RPO to trend down over the coming quarters with some exceptions relating to renewals and existing customer expansions. Also, before I end, I'd like to take the opportunity to shed some light on the details in our financial plan and the progress we're making with our consumption-based pricing. As Tom mentioned in his comments, we are on track with our plan to achieve non-GAAP profitability by the end of fiscal year '24. On a quarterly basis, we assess the business landscape and adjust our operating plan based on what we're experiencing in the market. During Q1, as Tom mentioned in his remarks, back then, we experienced headwinds and had to caution that the market downturn could be significant. We and the management team adjusted our operating plan and built out a careful track to become operating profitable by the end of FY '24. We are on that plan. In order to ensure that we monitor that plan, we have prepared detailed department budgets. We track variances, and we hold managers responsible. If there are variances, we correct them immediately. In our investor supplement, we have shared our current projected path to operating profit on a relative expense basis. Obviously, as I mentioned, we review our plan each quarter, but I wanted to share the detailed path with you guys at this time. As it relates to the model assumptions that we provided two quarters ago for our consumption-based pricing business, our preliminary analysis of the actual results suggest that we are at or better than that model. Therefore, to summarize, quarter results were above expectations and guidance is clear; a well-understood plan for our path to profitability is in action and our consumption-based model assumptions are on track. With these remarks, I would like to open this call up for questions.

And our first question comes from Mike Cikos with Needham & Company. Your line is open. Please go ahead.

I wanted to begin with some comments on customer dynamics to ensure I understand how we view the situation. If I'm counting correctly, the total customer count remained flat quarter-over-quarter. However, I recall Juho mentioning in his prepared remarks that C3 had closed 27 deals, with 17 of those being pilots under the consumption model. Could you clarify how you prioritize which clients fall into that category and when they would be considered customers?

Of course, yes. Thanks, Mike, for the question. So if you take a look at our quarterly filings, we've provided a detailed definition of what our customer is, but I'll provide it here as well. So, within a customer entity, which you can think about as the ultimate parent of, let's say, Coke Industries, there could be several subsidiaries, groups, departments, or various budget owners or various groups that use either incrementally or a tool or use it in different use cases. All of those individual groups we would characterize as a customer.

I guess what I'm getting at is if you closed 27 deals and 17 were pilots, right? If I think about the 27 deals that you guys closed, the assumption is then that they were all with existing customers. Is that fair? Or I guess there's movement under the hood in that customer cap that you're providing. Again, I'm just looking at the customer count being flat quarter-to-quarter.

Yes, totally understand. So let me clarify. We are still broadly speaking in our transition phase to the pure pilot model. If you recall in some of our prior conversations, some of the old trial arrangements, which are a customer in the period when their trials end, we engage in negotiations to turn them into production customers. During that phase, that customer is not considered a customer. It effectively falls off the calculation. And then when they would enter into a production deal, they come back into the calculation. So those dynamics are still at play because obviously, we still have a tail of the former trial model during the quarter. Now the 17 new pilot arrangements are also a mixture of new customers and existing customers who want to do a trial project with us, so that could be part of the bridge as well.

I understand. One more question: I know you highlighted improved profitability compared to expectations. There are two main factors to consider. First, we saw a significant increase in pro services revenue and its gross margin. Can you explain what contributed to this strong performance in pro services and the gross margin, which was over 90%, significantly higher than your usual levels? Was there any particular item that helped boost that gross margin?

So we have a highly skilled professional services organization, and we do various projects where we are able to command a high margin. Our standard kind of historical implementation services would be at a lower margin, but then we also do certain consulting-based or more ad-hoc projects that carry a very high gross margin.

Okay. And then the other thing that benefited the margins obviously was the much lower than expected sales and marketing expense. What was it that drove that benefit? And I guess the follow-up question on sales and marketing is I know that you guys have previously outlined your assumptions for sales and marketing headcount to grow 40% to 60% year-to-year. Are we tracking to that? Where did the sales and marketing headcount shake out for the quarter?

Okay. Thanks, Mike. That's another great question on that. So if you take a look at the investor supplement, we broke it out again for your benefit between sales and marketing, so you can see the relative proportion there and how we plan on how our current plan is to track into operating profit. To answer kind of quickly to your question, from a marketing perspective, as discussed in prior calls, we believe we have a very high brand value, and we no longer have to spend the historical amounts to gain or increase that value. Secondly, on the sales side, we have very aggressive hiring targets, but we are slightly behind those targets because we look for the best type of candidates to sell our products and services.

Thank you, and one moment for our next. Our next question comes from the line of Gil Luria with D.A. Davidson. Your line is open. Please go ahead.

It looks like the general AI conversation, you were ahead of. Can you help us a little bit with the context for that? When did you start the work on that? When did the pilot start? You talked about a couple of implementations maybe a preview of what you're going to talk next week about at your conference. It seems like there's a little confusion out there about what came first, the interest in generative AI, or your work on it. So would you mind walking us through that timeline?

Let me try that. This is Tom. We have been working with generative AI models since 2020. The recent boost in creativity was driven by a request from the Department of Defense, one of our larger customers. About four or five months ago, I received a message stating that they needed us to become the Google for the DoD. At first, I didn't understand that request, but after speaking with some colleagues like General Hyten, former Vice Chair of the Joint Chiefs, and General Cardon, Chair of the U.S. Army Cyber Command, they explained the process. They told me that during Joint Chiefs meetings, a question would arise regarding topics like diversity goals or satellite coverage, which would then get passed down through various ranks to be addressed. Our brainstorming sessions on the concept of Google for DoD led us to consider how the Google search interface, specifically the search bar, could be utilized for our applications across different sectors, such as DoD, manufacturing, and healthcare. We designed a system where users could simply type in questions, like satellite coverage, and generative AI would provide suggestions, followed by normal AI delivering the answers and interactive chat capabilities. The chat interface would contain only the specific information from the enterprise where it is implemented. Users receive logs of related content, including various document types. We combined enterprise search with indexing, generative AI, and reinforcement learning, leading us to create a solution that the Joint Chiefs could access at Wright-Patterson Air Force Base. We've shared a link in our news release to showcase this application, and I encourage everyone to check it out. We anticipate that our C3 AI generative search feature will be integrated into the C3 platform and made available to our customers this spring. It's already being deployed at Koch Industries and Baker Hughes as a core capability. To protect this intellectual property, we have several patents pending worldwide, and I encourage you to visit our website to learn more.

My follow-up is it sounds like you're talking about two things. One is how generative makes your current products better, and another how you can apply it in an enterprise level beyond your product set to other data sets within the enterprise. Which one is the bigger commercial opportunity?

That's a really good question. There is a big opportunity to do this in enterprises that do not use C3 and do not plan to use C3, but they want a unified view of their data. The honest answer is we have not figured out how to monetize that yet. We haven't set a price on it, but there is potentially a very large market there. It's not included in any of our operating plans yet, but it will be. This is a very non-obvious use of these open AI technologies. This is not about Chat GPT and chat, which is somewhat entertaining. Someday, I think it will be useful, but that's not what we're doing here. We're doing the opposite of chat. We're using these large language models to essentially explore the enterprise. But you raised a very good question. There is a monetization opportunity today that we have not yet figured out.

Thank you and one moment for our next. Our next question comes from the line of Michael Turits with KeyBanc. Your line is open. Please go ahead.

Tom, just to continue on the product front. On the product on generative today, I think it makes a lot of sense to the combination and the story with DoD. I guess the question is you're basically enterprise search that you're talking about, which is the name of the product. But that's not unless I'm wrong, was that a product you had before because that's obviously a market unto itself, and there's a lot of existing players in that market. So what's the history of having developed that?

What is an example of an enterprise product in the market?

Even what Elastic does for many people, right? And then how are you innovating to interact with other people's generative AI models?

The innovation is the way that we have combined those core technologies in a new and novel approach for a non-obvious application. We're taking the enterprise search UI, NLP, generative AI, reinforcement learning, and predictive analytics and combining those in a non-obvious way to solve a problem of utility.

Yes. So as Tom said, basically, we're using more modern techniques than, say, Elastic to interpret the questions using large language models, and then retrieving information from across the enterprise information systems, documents, and BI dashboards, etc. This is a novel approach to be able to search the entire corpus of an enterprise and leverages generative AI, utilizes these large language models and all the capability we have developed over the past decade to integrate and unify data across systems, sensor networks, images, text, etc. So this is different than it's new and is much more effective than traditional approaches for indexing.

Got it. That makes sense. Regarding Baker Hughes, I believe we had projected around another $270 million in contracts, though I'm not sure if that's accurate. Should I just add $35 million to that figure? Does this affect the amount you're forecasting for each quarter? Also, what was the amount received from them this quarter?

Okay. Good question. I think we need to look at your model effort. The short answer is that we disclose both in our release and the Q and the supplement, the Baker Hughes RPO, which is $188.5 million. We may need to check offline to clarify what that means for your model. Regarding the results in the quarter, the related party amount for Baker Hughes this quarter was $28.9 million.

Okay. Great. And does it extend the amount of time? I think you said it extended the dollar amount, maybe I missed it, but is it extending the period over which it's being paid also so adding through the same length?

Actually, one of the really exciting parts of this extended agreement is that we accelerated the payment schedules. So we're actually collecting cash faster from Baker Hughes. And then as it relates to the transaction price, so the accounting transaction price increased by $32.5 million because the variable consideration was eliminated as part of the expansion deal.

I think there might be some confusion regarding this, Michael. They have the option to extend the agreement if they choose, but that doesn't refer to extending the payment terms. Instead, it means adding more years to the agreement for previously agreed amounts of cash.

Thank you and one moment for our next question. And our next question comes from the line of Sanjit Singh with Morgan Stanley. Your line is open. Please go ahead.

So, if I take the comments from Tom in his script and Tom's like notable that you are speaking to a better demand environment and I think the rest of the software is still pretty gloomy. And then I sort of Juho's comments around the consumption actual versus predicted sort of coming in line with what you thought, I try to put these pieces together. Revenue is declining year-over-year right now, obviously because the business model is going through some transition. And so I guess the big question is, where does revenue growth go to? I know you guys still have another quarter to go before your Q4, but The Street is sort of expecting 20% revenue growth. Is that something that seems achievable from your line of sight and more optimistic view? Is it something better than that or below that? I'm just trying to understand where the business is.

Juho will handle this, Sanjit, but just for the record, if I'm not mistaken, our revenue for fiscal year '23 will be greater than fiscal year '22, not less. Okay? I just want there to be any misunderstanding about the listing. Juho, why don't you take the rest of it?

Yes, of course, yes. Thanks, Sanjit. Let me try to recap your question. First of all, as Tom mentioned, the annual guide would be a year-over-year increase on the total annual results. On a quarterly basis, yes, we're seeing very promising signs that our model assumptions are good. We're seeing actual results that are at the model or even better. Previously, we have provided some really early onset outlook that 24% would be around 30% growth. The Street expects around 20% growth; I'd say that it's achievable at least to what the Street is saying, and we certainly are interested; we certainly are targeting a higher growth, so our model actuates as we plan it to actually.

No, that's super helpful. I really appreciate the thoughts. And I guess, Thomas, a follow-up on just your more optimistic view versus, let's say, a year ago, where is that coming from? I mean, is there a way you can sort of talk to it from a vertical perspective. A lot of the companies that are struggling now are selling to other tech companies, which kind of explains a lot of their weakness. But what are you seeing in your customer base?

That's a great question. Although you may not be able to display the slide, if you could show the diversity in the industry, it would highlight that oil and gas is a very strong segment for us and one of the healthiest in the economy. We are fortunate to be in a good position with the right partner, and it's benefitting us significantly. Juho, could you elaborate on the industry diversification we observed in the pilot projects, as it spans across various segments of the economy?

Yes, thanks, Tom. Sanjit, on the supplemental or the supplement that we have on our website, if you check out Slide 24, we show the diversity in total bookings, and then we also show the diversity of the pilot bookings. So we're very excited about the nine industries that we have pilot deals during the quarter. Just to rattle them off, we've got Fed accounting services, consumer packaged goods, manufacturing, oil and gas, financial services, high tech, telecom, and then we have state and local. So we've got a diverse group here.

Yes. I guess the point would be is that...

The number remains a very small slice of our business, unlike a lot of software companies.

Yes, that's correct. The main point being made is that it's not just about diversification. The trend in the software industry over the past few quarters has been a general caution regarding investments, especially in cloud-related areas, as many have already spent significantly in the last two years. It seems you are indicating that the C3 customer base is not showing this hesitation and is actively investing. Would that be an accurate way to describe the situation?

From my perspective, and I think maybe Tom could shed some light or Ed, but I certainly believe and have seen from our customer base that our products provide value to customers. There's efficiencies, there's more productivity. When you have whatever market, whatever industry you have, everybody is interested in cutting costs and becoming more profitable, we are right there. Our stuff is helping them to get business value very quickly. Maybe, Ed, do you want to shed some light on that?

In the qualified pilot opportunities that we're working on, it was Google alone; we're in the process of trying to close deals with how many. It's a big number. We are engaged, okay? We have 291. We have over 100 opportunities with Google Cloud, and we're currently engaged in licensing discussions. I mean, something changed Sanjit, from July of 2022.

Thank you and one moment for our next question. The next question comes from the line of Arvind Ramnani with Paper Stanley. Your line is open. Please go ahead.

I wanted to ask about your expanded partnership with Google and also AWS. When you look at these partnerships, how are you sort of measuring the commercial success or the ROI of these relationships? You have a lot of good relationships and partnerships we have established and have continued to expand. But sort of looking forward and saying like, these are the partnerships that are resulting in the greatest commercial success, and this is where we need to continue investing in. What’s the approach in figuring out that ROI and investment?

In the short term, we're measuring on pilots closed. Looking in the medium term, we're measuring it on consumption, okay? The reason that AWS and Google and others are partnering with us is because we accelerate consumption. In other words, the customer doesn't have to spend two years building the ML application or whatever it might be; the also can consume it much faster. That's what these guys are interested in. They're interested in CPU cycles and storage hours.

Terrific. Regarding the macro commentary, you mentioned being early in discussing some of the headwinds. You're indicating that those headwinds might shift to tailwinds. Do you have evidence to confidently suggest that things are starting to improve, or are we still in a situation where things could go either way?

Well, you're an expert at this, and we're not, Sanjit. But we were one of the earliest to call that it was really ugly out there. Then everybody else filed on, but then everybody else started doing layoffs, 5%, 10%, 20% at a time, and those aren't over yet. All I can say is what we're seeing now is dramatically different. I didn't say that to pump our stock price; I was just telling you what was really going on in the market. I'm telling you what's really going on in the market. Now I don't know whether this is ephemeral or whether this is going to be sustained. But I think that the way this looks to me is sooner or later, the Fed is going to take its foot off the brakes. When the Fed takes its foot off the brakes, this will be a cash positive rapidly growing business, and I think we're going to be off to the races. Now I don't know when that is, okay? But when that is, that's what we'll be. Whether this thing could take a dip or not, that's beyond my pay grade. But I just know that right now, something has changed.

Thank you and one moment for our next question. Our next question comes from the line of Kingsley Crane with Cannaccord. Your line is open. Please go ahead.

Tom, so I appreciated your comments on the applicability of the generative AI product outside of the typical customer base. But when we think about the core customer base, what do you think the potential is for generative AI to drive new trials? Or is this primarily a product that would work best as sort of a second, third, or fourth product within the basin?

I tell you, go look at the demo. People get awfully excited about it. I just came back from somewhere. I know where; Barcelona. I just got off a plane from Barcelona, where I showed it to a bunch of people, Ed and I. It's a pretty exciting product. It's something brand new. You got to just take a look at it. It's hard to describe. I think it will make our products more attractive. It will make it easier for people to use, easier to handle change management, which is everything in enterprise applications. It will, I think, substantially differentiate us in the market. Take a look at this user interface compared to what SAP, Oracle, Salesforce, and everybody else has in the market, which is just kind of a cheap copy of the Cboe version of the 7 architecture that Ed and I embedded in 2002. This is something very different, and the excitement appears to be palpable.

Wherever growth ends up next year in fiscal '24, I just want to think a little bit more about the balance of existing customer expansion and then the net new customer deals. How much of the ramp in trial conversions are we factoring into growth next year? Okay. Great question. So I would suggest that you take a look at the assumptions we provided and model out based on the actuals we have provided. That should give you a good idea as to where we should be as it relates to the consumption-based business, and I'm happy to chat with you if you have any additional questions on it.

Thank you and one moment for our next question. And our next question comes from the line of Pat Walravens with JMP Securities. Your line is open. Please go ahead.

Great. Tom, one for you first. Can you just sort of give us an overview of how the business with the DoD is going? Maybe as a touchpoint to that, in December 2021, you guys signed the $500 million production other transaction agreements. How much of that is the DoD actually taken down? How much of the $500 million do they spend?

That's a good question, Pat, and I don't know the answer. We have two agreements like that; one for $500 million and one for $100 million. Our business with the DoD appears to be very promising. Recently, we hosted a significant number of CIOs from the DoD, including representatives from the Army, Navy, Air Force, Cyber Command, Spacecom, Marines, and National Guard. They were all in our office, and there were extensive discussions about some very important projects. So the DoD seems to be looking favorable. However, I can't specify how much of that deal we have already utilized. We have a proposal in front of one of the agencies that will use a substantial part of it, and we'll see how that develops.

Hopefully, that gets done. All right. Juho, I wanted to point out something simple. If I look at your Slide 20, by multiplying your average TCV by the number of deals you have for each quarter on the chart, I can calculate your total bookings, right?

Yes, I think, Pat, that sounds about the right math, yes.

Right. That seems reasonable. So you had 27 deals in Q3 at $1.9 million. So that's $51 million in bookings. And you had 20 deals basically a year ago at $5.6 million each, $112 million. So bookings have been cut by more than half. Is that a fair assessment of what's really going on?

Let's also highlight that Juho will provide further details, but I understand that a year ago, the deals typically ranged from $10 million to $50 million. This quarter, due to consumption-based pricing, we're seeing smaller $0.5 million trials. As Juho explained, we are shifting to a completely different type of deal. In the upcoming quarters, you will notice an increase in the number of deals we are closing. However, we are no longer pursuing our previous model of $10 million to $50 million deals. We're now focusing on $0.5 million agreements. Juho?

Yes. Thanks, Tom. Yes, Pat, I think Tom summarized it perfectly. This is directly expected, and this is a direct result of the consumption-based pricing shift two quarters ago.

Right? So, I mean, it's great transparency, right? Most companies don't give us their total bookings. So as long as you guys keep getting these slides, we're going to be able to keep doing the math. When do you think we start seeing that go up?

If you have the chart open, please turn to Slide 18. On that slide, you will find the same chart we provided for the last three quarters, which compares the expected ramp-up in consumption revenue against the ramp-up in subscriptions. This aligns with our expectations, and as I've mentioned, we believe we are following the model we've shared with you. From this point, we can determine when we can anticipate the ramp-up to occur.

Thank you and one moment for our next question. Our next question comes from the line of Arsenije Matovic with Wolfe Research. Your line is open. Please go ahead.

This is Arsenije on for Gal. It seems like this is now two quarters of coming in above expectations that were initially communicated for progress in the pilot initiative. Should we think about these pilots coming in faster than your initial communication suggested, given the more positive macro outlook you're seeing relative to the first quarter? And then I had one brief follow-up.

I think, Arsenije, like I said, the original assumptions provided are still valid. We're seeing great results. We're very excited about the traction with our partners and the number of deals that we're tracking, but we're quite confident with our model at this time.

Got it. That's helpful. And then what caused the large services contribution from Baker Hughes in the quarter? It looks like it was $8.6 million. Did this contribute to the above 90% services gross margin? And can you remind us how COGS associated with Baker Hughes service revenue is accounted for?

Perfect. So I think you missed the initial call or initial question for Mike who asked this, and I explained the gross margin impact. Just as a quick recap, we have various types of professional services, and we have a highly skilled workforce; we're able to command a premium on those services, and Baker Hughes is one of our many customers whom we provide professional services.

Operator, we'll just take one last quick call again.

And our last question is a follow-up question from the line of Michael Turits with KeyBanc. Your line is open. Please go ahead.

Thanks, gentlemen, for letting me get a follow-up. The guide for next quarter of the $70 million to $72 million, you have that big bump in services, which seem to have come back faster than I think we might have expected. Any sense you can give us for what the trends are for both subscription and services into next quarter, the subscription trend down, flat services too high?

I think we've historically provided that our target is between 10% and 20%, and I think we're going to be in that range quite nicely. Awesome. So thanks, everybody, for joining for our third quarter conference call, and we really look forward to chatting with you guys all in the future, and thank you for your time.

This concludes today's conference call. Thank you for participating. You may now disconnect.