C3.ai, Inc. (AI) Earnings Call Transcript & Summary

[Audio Gap]

[Audio Gap] Silicon Valley, and I'm in the computer software business. And -- very pleased to talk with you this morning, kind of net-net, I've been in the software business now going on 4 decades. I'm a computer scientist from the University of Illinois. Has my graduate working relationship with database theory. I was involved in a start-up company called Oracle Corporation, turned out to be a pretty good idea. Left there, about a decade later. When I went to Oracle, it was about a $2 million business, and we grew it quite a bit larger than it. [indiscernible] one of the guys who ran that business, started a company called Siebel Systems in 1993. That was about the application of information technology, communication technology to sales, marketing and customer service. We invented this market [indiscernible] CRM. I hired Mark Benioff out of USC and in Oracle and '96 to '97. He studied business administration at USC. I assume he got into the front door, but I don't know, and Mark is obviously a talented guy. Anyhow in '93, we started a company called Siebel Systems. By 2000, we're doing about $2 billion in revenue, and had 8,000 employees in 29 countries. I think even today, that's the fastest growing enterprise application software company in history. We ran that company too during the dot-com era when nobody cared about profits and nobody cared about cash flow. We ran a consistently profitable business, consistent with cash positive business. Some of you will remember that there was a pretty significant economic downturn, first when the dot.com thing blew up and then when those guys flew a couple of airplanes into the World Trade Center. And so 2000 -- 2001, 2002 were a little rocky in the equity markets, more than that. And in that 3-year period, we generated, I think, $2 billion in free cash flow. So we understand how to run a cash positive business. At the C3.ai, we're addressing a market called enterprise AI. So if you can think about -- so there's -- this market of enterprise AI is kind of very confusing because you have all these bright shiny objects out there that all appear to be the same thing, right? Alteryx and Dynamo and Databricks and data stacks and data log and data IQ and Datadog and all these d*** things, okay? And they all do useful things like auto ML or data persistence or machine learning services. Well, we spent -- we went about it in a little bit different way. And so where we spent, say, a decade and invested about $1 billion building a software stack, that includes basically the union of all of those capabilities that you're aware of, that allows us to take these enterprise applications that we started building in the '80s. So let's think about what was that all about you guys -- some of you tracked that. So in the early '80s, companies like Oracle and SAP and others, we took relation database technology. On top of that, we built application development tools, we -- things that did forms and reports and screens and whatnot. We used those application development tools to build applications like ERP and CRM and manufacturing or what have you. It turned out to be a pretty good idea, right? And so that's about a $0.5 billion software market today for enterprise application software. And what do we do with those systems? We go into Boeing or [ CAT ] or Verizon or whatever it might be, and we tell them with perfect 2020 hindsight, what their inventory levels were for all the components in the build materials. Say a Boeing or a 777, 767, 787, 737, 757, 777 alone has got a million components in the build materials. So that's a supply chain that goes from South Carolina back to Shenzhen and they need to be able to report exactly, like how many units were in each band every like 30 days. Or what their customer churn was or what their cash balances were or what their failure rate was. Now what we do in the way that we think about enterprise AI is a little bit different than all of these other guys. Well, the business that we're in is taking on these enterprise applications, putting our stack next to it and making them predictive in nature rather than descriptive. So rather than looking with perfect 2020 hindsight, you see how many parts you had in each band or each resistor or [ plap actuator ], whatever the h*** it was, in between South Carolina and Shenzhen, okay, 90 days ago, we would tell you exactly how many parts you need in each band in the next 180 days to meet the demand function so that you can deliver on time and full. Rather than tell you at Verizon, Bank of America, whatever it might be, what your customer churn was 30 days or 60 days or 90 days ago, we can take these existing applications and by making them predictive we can tell the bank or Verizon or whoever it might be, what customers by name are going to leave in the next 180 days. And we can do that with very high levels of precision. So now we can change the future, right? We can -- if we're Bank of America, and we see that Apple's about to like close their cash balances with us, we're getting to light up a G5, go out to [ Keith ] [indiscernible] go talk to Tim Cook and do what you need to do to save the deal. So you could -- and at Verizon you do it in mass. Let's talk about fraud. We have opportunity. Fraud is a big deal in the utility sectors in insurance, okay, in health, in banking, and we have this thing called AML that you guys all read about. We every day, you pick up the Wall Street Journal, you can see that Deutsche Bank is being indited yet again for laundering cash, for God knows who. They used to do it for the Nazis and now they do it for the Russians. True, true, true. Okay? Okay? Okay. Now they're true stories. But these guys make Bernie Madoff look like our girl scout, anyway. So let me -- I cannot lose that pen, okay? So don't let me get off here, Andre, I know you guys have figured one of these things out. These remarkable deals, if you figured about these note-takers, oh my God, they're unbelievable, anyway, I will get a pitch for that some other time. Okay. So now if we don't report at Bank of America or whatever the financial institution, what our AML activities were in each reporting period, we're in a world of [indiscernible]. CEO has got to rewrite his resume. We've got billions of dollars of problems. And so this is what AML is all about. Well, by making it predictive we can identify these fraudulent events as they're happening and prevent the fraudulent event from happening. So fast forward, what we do is we just sit on top of the enterprise application software market, as you know it, be it CRM manufacturing supply chain, ERP, and we make these applications predictive, so that people can rather than simply look at -- with perfect 2020 hindsight, what happened 6 months ago, with perfect foresight, you can change the future. And this promises to be an, believe it or not, enterprise application software, AI, applications promises to be a $600 billion market, okay? We -- I believe we are the largest company in the space. I will argue that 2, 3, 4, 5 years from now, nobody will be satisfied with perfect 2020 hindsight, and everybody is going to know what customers are going to leave next or identify fraud in real time or understand how to deliver their products or TIF or be able to avoid the airplanes flying out of the sky, predictive maintenance, okay, rather than simply report and how many did fall out of the sky. Now so rough numbers, we -- so this is a 600 -- analysts suggest that enterprise AI space is a $600 billion addressable market. We're the largest company in the space. We're roughly growing at a -- roughly $0.25 billion enterprise application software company growing at roughly a 40% compound annual growth rate. We have roughly $1 billion cash in the bank, and we currently have an enterprise value of about $2 billion. So it looks like an opportunity to me. What question. What questions could -- Andre, you can go after me or will you guys what questions do you have?

Yes. I guess I'll start with one. So you gave some great examples of how you provide value for clients. But could you go into a little bit detail about how the interoperability and flexibility of your platform is really a unique differentiator in the market? And because there's hundreds of different AI players, and I think one thing that I've always thought about like what's the difference between A, B and C. But you have a truly unique solution. And so in terms of being able to plug-and-play with any single system, can you talk a little bit about how that is a differentiator?

Let's talk about a couple of our let's say Shell AI. Shell's a fifth largest company in the world, roughly a $300 billion company. And they standardized on C3.ai for upstream, downstream, midstream integration of renewables. Okay, Shell stood up in front of a Bank of America conference, okay, in front of financial analysts, and said they're getting $2 billion in economic benefit from what they call Shell AI this year. $2 billion, up from $1 billion last year. Okay, their journey from us has gone from over a period of about, I suspect 7 years. I think the first deal we did with them was a trial for maybe $250,000, then another trial for $0.5 million. And then a $9 million deal and then a $50 million deal. And now it's like probably close -- it's roughly $100 million. I think $90 million we've done with these guys. And it keeps growing and growing and growing. Now at Enel, Enel is the largest utility in the free world. They're based in Rome. They have 60 million meters in 40 countries. The problem we solve there is smart grid analytics, which is a difficult problem because the grid is the largest, most complex machine ever built. And we use predictive analytics to basically optimize the whole grid infrastructure. This is everything to do with distributing energy resource. It's about supply safer, more reliable, cleaner, lower-cost energy. And if we look at the annual report for Enel through these efforts, Francesco Starace, the CEO, will claim to his shareholders that he's realizing EUR 5 billion a year in recurring economic benefit from what they do for us. At the United States Airlines and that -- and this is an account that's grown over the years from our first transaction, whether it was probably a couple of hundred thousand dollars and now it's definitely in the scores of millions. I can't tell you -- I mean I just don't know the actual number, but it's big. The United States Air Force, we're doing predictive maintenance for the United States Air Force. This is the largest fleet of aircraft in the world. It's about 5,000 aircraft, where the average deployment -- and some of these things cost like $100 million a copy, right? So like it's a pretty big investment for these guys. And the average deployment rate of many of these aircraft is like 50%. So pilot goes in, in Zero Dark Thirty, pushes the start button, and some might live test [ officers ], you're not flying because some system failed. Well, by taking all of the information that they have about these aircraft, [ sorties ], build materials, flight history, weather, okay, maintenance histories, all the telemetry off of the devices, there's a lot of information. Say about a B-1 Bomber or an F-35 joint strike fighter, we can build machine learning models that will tell them what device is going to fail, what system is going to fail next. You say in the next 50 or 100 flight hours, be it propulsion, flight controls, [ ocular ], power unit, whatever it might be, so they can replace that system before it fails and increase the availability of aircraft by 10%, 20%, 30%. Well, at the scale of the United States Air Force is operating, this is billions of dollars. And today, into production, we have the B-1 Bomber, KC-135, F-15, F-16, F-18, F-35, joint strike fighter. I think we have 18 platforms in production today. And we will have 22 platforms in production by sometime this, I think, September. So that's the scale at which we're operating. So there's -- if we look at the other -- this AI space, it's very confusing with the Datadogs, the Alteryxs and Dynamos and Cassandras that all appear to be equivalent to what we do, they're completely different animals. These are the problems that we saw. And I don't know how many companies you're aware of where the company stood up on stage like Shell did, like LyondellBasell did, okay, like Coke Industries did, okay, at our [ user ] conference and say that they're getting $1 billion, $2 billion, $3 billion of economic benefit from the use of their solutions. But I can tell you, it has never happened in an SAP conference. It has never happened at an Oracle conference. It has never happened in a Salesforce conference and it never happened in the Siebel Systems conference. Not that those aren't all fine companies because I think they are, it's just that it's an unusual amount of economic benefit.

Great. Well, so I think one thing that really highlights the value proposition is just your ability to grow within your client base. So you gave an example when you started out with I think a $250,000 contract and expanded that materially, how do we think about -- like how large is an average contract initially? And then what does that typically grow to? And I guess, how penetrated do you think you are currently within your client base? And what's that long term [ ARPU ]?

Our average, this is a big example of a large integrated energy company in Europe. Okay, where the first contract with us was EUR 300,000, then EUR 4 million, then EUR 35 million, EUR 36 million, EUR 54 million, EUR 90 million, EUR 91 million, EUR 120 million. This is an example of a large chemical company in the United States. Their first contract was just for $9 million, and then $14 million and then they expanded to $59 million. These people are expecting to get $1 billion in economic benefit this year. This is one of the government agencies. First contract was $1 million then $6 million, then $17 million. Now they now have a $0.5 billion authority to operate, to expand this application to the federal government. This is a large industrial manufacturing company, where the first contract was $1 million. So this is the value of customer service. This is not selling. This is about keeping customers happy. Our average transaction last quarter If I'm not mistaken was $2.9 million. It used to be about $15 million. So it's really come down, down, down. What was the other part? So I think that our average customer is probably 5% to 10% penetrated. And the -- that's a swag. But the value of keeping customers satisfied is high.

So still very early innings. So I guess when you think about that how do you think about the growth algorithm going forward? What's the emphasis of expanding that, continuing to service these clients you have over 200 large enterprise and government clients from -- versus going to get new logos? And on that topic, just if you could touch on the importance of the partner network and your strategy to expand that and to drive further sales and scale through the partner network.

Yes. If I'm not mistaken, in the last year, we went from about 153 customers to 215, Juho, is that right?

[indiscernible]

Thank you, 151, 223. Okay. And the customer would be a business -- it's not -- so it's hard to call it Shell or Department of Defense is one customer, right? Because we have all these independent decisions, independent budgets, and they all are separate customers. When you think about customer account, honestly, we're more focused on penetration than we are in getting 100 new logos next quarter. If you look at a Snowflake, you look at an Alteryx, and what they want to do 200 new customers at $20,000 each. I'm just not in that business. So we're in a little bit different business. Partner Ecosystem. Partner Ecosystem is a very important component of this. So we, some years ago, started to get -- put a partner ecosystem together. We did a strategic business partnership with Baker Hughes to go to market with them in oil and gas. We go to market with Google Cloud in a big way across many verticals, with Satya and Judson and Microsoft Azure in a big way. Raytheon, in defense intelligence, FIS in financial services, more partners to come. 64% of my business last year was closed with a partner ecosystem. It's hugely important to us in closing business and is a main, main -- a very, very important part of the strategy, particularly as we move down market into lower-priced products. Lower-priced products like we have a product Ex Machina, which is kind of 20% reversion of Alteryx. We have a really hot product in AI-enabled CRM, which is a relatively low-priced product where we apply AI, so we don't replace any of the existing CRM systems like Dynamo or Siebel or Salesforce. We sit on top of it and make it predictive, so we increase the precision of the revenue forecast or the product forecast from like 50% or like 5% [indiscernible]. So particularly as we get into higher-volume mass market products, these channels like the Azure Marketplace and the Google Marketplace becoming increasingly important.

Great. And so following on with that point, I think one thing that really stood out to me was how successful the Baker Hughes partnership has been and in a pretty quick time frame, too. How do you think -- how replicable can that be? And what additional verticals or industry you're potentially looking at expanding into beyond what you're currently partnered with or targeting from a vertical [indiscernible]?

We are focused on establishing and maintaining a global market leadership position in enterprise AI. Now that sounds as ridiculous today as it did when Larry Ellison said it in 1983, right? Okay? Well, he did it, okay. And it sounded ridiculous when I said it in 1993 and CRM, well when I sold Siebel to Oracle, we had 80% market share at CRM. We were doing roughly $2 billion in revenue and Mark Benioff was doing about $180 million. So that wasn't -- he wasn't even an [indiscernible] by the way, I will say after that, he did a pretty good job. The -- so we're -- okay, so we're focused geographically on building major account units, enterprise sales units and middle market units in APAC, in EMEA, okay, and in North America. Then we have vertical market units for utilities, oil and gas, chemicals, aerospace, financial services, what have you. Our strategy there is in each of these vertical markets to form a partnership with a Baker Hughes, with a Raytheon in telco, partner to follow in financial services, FIS. And so Baker Hughes, for example, gives us access to 12,000 salespeople around the world. So we've been walking to the Board room at Aramco, I couldn't get into the Board room at Aramco without Baker Hughes. Nobody gets in. On horizontal market partners include Google Cloud, Azure, AWS, [ ADF ], so we will grow out the partner ecosystem, and we're focused at the large enterprise market, the middle and the low end of the market. And I think we're -- I think it looks to me like the next couple of years could be a little bit rocky out here in the markets. We like -- rerun the block a couple of times like you guys have and I don't remember a time, this is like the perfect storm. I'm like why is this the Four Horsemen of the Apocalypse, right? It all happened at the same time. We have pestilence, famine, war and death, I think we got it all. Recession in Europe, recession in the United States. I don't know what -- I just know -- see what I read in the newspapers from Jamie Dimon and others, right. Inflation, the idea of inflation is 8%. I think that's laughable, the 8%. I mean, how do you have fuel prices go up 100%, groceries go up 22% and rent go up 30% and get 8% out of that? Like who's lying to who? So okay. And we have pestilence, we had COVID, we have war in Europe that I don't think is likely to stop soon. We have a recession going on in the United States and likely -- and I think likely in Europe. And so this is -- and we have equity markets. We have market expansion that looks to me, and again, you guys are expert of this. I'm not an investor, I just like to build software companies. But it looks to me like this market expansion were about 70 years too long. And so I think this looks like a zinger. And our position is as we go into this, we have $1 billion cash in the bank, okay? So this is an ongoing concern. I mean, with all due respect, I mean, it's not that I don't have respect for equity markets and investors because I do, okay? I do. And I think to myself every day that I'll do what we do for a living, because I think it would be really hard. And the -- particularly if you're investing other people's money. That being said, I think last quarter, I consumed we're at about $15 million cash, just negative free cash flow. I have $1 billion cash in the bank, we have a -- we're growing, we have a business where we have 80% gross margins. Come on, guys. It's not that hard to run a profitable business of 80% gross margins. My current cost structure looks like this. I'll show you this actually. I don't have to rely on my memory. Okay, there's the current cost structure on the left row et cetera, we've cost of revenue at 21%, okay? Marketing programs, I'm invested massively in branding, okay? You saw the NPR ads, and you see the ads at CNBC. We've been building a brand of Enterprise AI. We did it. We invested in it. It was a good investment, okay? Cost of sales, I think it says 23% there. Research and development, 44%. Nobody spends 44% on R&D, right? I mean we have been investing in innovation. G&A at 15%. So now we bring these down to a reasonable level, okay, where marketing -- cut marketing for rationalized -- our cost of revenue will be about 20%. Slash marketing programs down to 11%. Grow -- continued investments in sales and partner capacity a little greater than the rate of revenue growth, so that gets to about 26%. We have economies of scale from research and development, which gets us down to 28%, which is still high. I mean, compare that to other enterprise application software companies, we're still investing in innovation. This is a cash positive business. I think in focus, a steady state. We're looking at a -- this, we said that we could do in 8 to 12 quarters. Honestly, I can do this in 90 days. Now, it wouldn't be rational, okay? But I could do it in 90 days, okay? But I'm not going to do that, okay? That is not in the best interest of the shareholders, or the best interest of my employees, or their families or of their customers. But we'll do this over 8 to 12 quarters. We'll be operating a business to consider that with a non-GAAP operating income of 20%, just like we did at Siebel, we know how to do this. Again, we have roughly $1 billion cash in the bank. I think we burned through $15 million last quarter. Fast math, I think it would continue to burn cash at that rate, we'll run out of cash at $67 million a quarter. So come h*** or high water, this is an ongoing concern. So we're here to take care of our employees, take care of our customers and grow the business.

Awesome, following up on those 2 points. There was a lot to digest there, but how do you think about demand for your services and your growth rate if we do have what you call a zinger recession or environment? And what is currently kind of baked into your outlook? And then we only have a couple of minutes, but also if you could talk on -- I mean that's an 8- to 12-month target, but what do you view your longer-term operating margin targets is what...?

20%.

20%. How long do you think it would take to get there? And then how do you think about free cash flow-through at that point? And also what kind of the long-term growth rate would we be thinking about?

I mean I think non-GAAP operating profit and free cash flow are kind of the same thing. Okay, it is just a better timing. Okay. Yes. So it's basically the same things, it is just timing differences there. And so that should be the objective I mean with $1 billion cash in the bank, I think we're pretty well set. So I did my job as CEO. What's my primary job? To make sure it's an ongoing concern. Okay, this is an ongoing concern. What was the other question you've list in there?

The first question was about just how do you think the business would hold up in a zinger recession, and what the playbook would be because obviously you have a...

A lot of what we do is supply network risk, okay, which is very -- and let me tell you, that's a huge issue, okay? The supply chain management at the Shells, at the Cokes, okay, at the Air Forces, I mean it's a big issue. The new supply chains globally are completely s****** up, okay? And so we allow people to characterize their supply chain all the way from North Carolina to Shenzhen, times and product lines, and identify where the hot spots are, where the risks are and mitigate that risk so they can ship on time in full. So that looks like a tailwind. We are in the cost savings business, so I think as people tighten their belts, I think the economic benefits that people realize on what we do happens pretty fast. These are economic returns if they get in 6 to 12 months. And so whether we're talking about scarce optimization of supply chain, demand forecasting, AI-based predictive maintenance, all about saving costs, it's all about generating economic benefit. So I think that we should -- if this turns out to be a zinger, I think we should weather it quite well. We would like to have opportunities [indiscernible].

Great. And you're not seeing anything currently?

Not yet.

Okay. Well, I guess, I think we have time for one more quick question. So you talked about $1 billion cash in the bank, 6 to 7 quarters until you go out of cash. What are you thinking about from a share repurchase perspective? Because I know you guys issued an authorization of $100 million. So how do you feel about share repurchases currently in this environment?

I think my #1 objective, okay, my #1 obligation, fiduciary obligation, okay, to the investors, okay, to the customer, okay, and to the employees is to make sure come h*** or high water, this is an ongoing concern, okay? My #1 objective is not to go to manipulate the stock price this week so that we squeeze some short -- or somebody makes too much on their stock, but not my job. Okay. This is just not what I do. I think that this could get ugly. Okay, what's the sequence? Come on guys. You saw this in 1980. You saw it in 1989. You saw it in 2000. You saw it in 2002, you saw in 2008, okay, we have seen the beginning of it. Okay, the markets get all nervous and twitchy to get all this volatility. Then people started announcing, okay, hiring freezes, okay, then people started announcing layoffs. See Facebook, see Amazon, see Salesforce, see a Twitter, Tesla, okay? What happens next, okay? After we hit the labs, how many of these companies that are not generating cash that no longer have excess capital go out of business? I mean there's not a day when I don't get an e-mail about another AI company that's for sale. By the way, it takes me about 20 milliseconds to hit the delete button, okay? And not a day when some bank isn't investing one of those things that were up there. So how many companies went out of business from 2000 -- software companies from 2000 to 2003? Maybe 2,000? Literally 2,000. For those of you who took a write up 101 in Silicon Valley, I mean all those buildings are transparent, right? So then companies go out of business and all these people who went home to work in their PJs and get paid in Bitcoin, all the while their stock is worth like 10% of what it used to be, okay? The rent has gone up 30%. The groceries have gone up 25%. And my dear, I need to go get a job. And so things are going to change. And so we have participated and bought some stock. We may again in the future, we have the authority to do that. But I'm not again, as far as I'm concerned, the market could close for 5 years, okay? Just turn out the lights in the market. You turn the lights on in 5 years, I think that customers will be happy, employees would be happy shareholders will be happy.

Well, unfortunately, we're out of time, but I just want to thank you so much for your perspective and for coming. Everyone, please join me in thanking both Tom and Juho.

Thanks guys.