All right. Fantastic. Welcome everyone. It's really awesome to have you joining us this morning. Welcome to the Citizens of Technology conference. I think this is the 24th year that we've had this conference under different names. 23 of them in this hotel, if you can believe that. Yeah. Yeah. Stephen, we're just delighted to have you join us today. Thank you for coming. Sitting here on stage with me is Stephen Ehikian, the CEO of C3.ai. My plan is we're gonna talk about Stephen's background for five minutes and 'cause a lot of people I think haven't had the chance to meet you, and your background is super relevant. We'll talk about what's going on with C3.ai, what it was like when you landed, you know.
How many months has it been?
Six months.
What the last 6 months have been like. It's been tough, right?
Yeah.
What the plan is to right the ship, we're gonna talk a little bit about current events because your background for what's going on sort of with Anthropic and the Department of Defense and what's going on in the Middle East, I think is super helpful. Then we'll open up for questions.
Great.
Sound good?
That sounds great.
All right. Cool. Where are you from?
born in Stanford Hospital.
Oh, really?
Born and raised in the Peninsula, yes.
Yeah.
Grew up here and then, East Coast for about a decade, and then came back out here for grad school.
What was grad school?
Stanford.
You worked at, you did a little investment banking, right?
My background was an engineer in college-
Yeah
that graduated in 2004 when tech really wasn't recruiting on the East Coast. Wall Street was gearing up. I played sports in college. I played football, and a lot of my teammates went to Wall Street. I knew nothing about business, knew nothing about banking. Found myself at Morgan Stanley doing investment banking for two years, and then followed the herd into the private equity, then hedge fund space. Born and raised in California, Silicon Valley, my role models growing up were always people in the technology space. It was individuals like a Bill Campbell that was my middle school football coach, the guy who got me introduced to computers at a young age.
always had this desire to be part of the Silicon Valley lore, and my opportunity was to, when I came back to grad school in 2009 at Stanford, and one of my classmates who worked at a company called Palantir, which I'm sure you're familiar with. we had an idea of leveraging the methodology of Palantir to bring it to sales teams. What was the pain point we're trying to solve? sales teams were logging a lot of data points, emails-
This is RelateIQ?
It was RelateIQ.
Okay. 2011.
2011.
I never heard the Palantir angle. That's interesting.
So, uh-
What was the part of it? Was it the ontology? What was it about Palantir?
No, it was the passive capture digital communication signals.
Okay.
What they were doing with the government, we were trying to do with sales teams.
Oh, interesting.
you know, sales teams are logging emails and phone calls all day long, and so we said, "Can we automate the data entry component? Once you have your data into this database, can we derive insights off of it?" That became a company called RelateIQ. It was a next gen CRM. We were fortunate, Salesforce recognized the value and purchased us, and that became the foundation of their Einstein platform. Spent 3 and a half years there. We spent a lot of time.
All right. That's 2014.
2014, sold that company. 2014 to 2017, built out, worked a lot in the customer service space within Service Cloud, within Salesforce. Saw the incredible rise of customer support and at the same time rise of automation, with like RPA systems. We said, "The intersection of automation and support, let's start a company around that," really focused on the call center. We started building a solution around automating customer support and making it easier to build those applications.
Wait, what year are we in now? Are we in?
This is 2018.
Are we in Airkit? This is after you've already... Okay.
Oh, sorry. Yeah. Let me go through the steps.
Okay. You left Salesforce.
Left Salesforce.
You lasted three years at Salesforce.
three and a half years. Left Salesforce.
Was that longer than your earn-out, or did that match your earn-out?
No, it was actually, it was the right time.
Yeah.
Independent of the earn-out, it was an incredible opportunity.
Right.
I mean, starting a company zero to one was great, but then getting in the hands of Salesforce and have that be a core foundation of their you know, AI platform, Einstein, was incredible, and learning how to operate at a much larger scale.
Yeah.
Around that time, we also saw a whole new opportunity. Mark, the CEO, was fortunate to help fund us initially to leave and build this company called Airkit.
Oh, Benioff funded Airkit?
He, Salesforce was the initial investor.
Yeah. Oh, sort of like Salesforce Ventures. Okay.
That's right.
Oh, cool.
We were fortunate to have the backing of them and Accel, and we started a company called Airkit, and the goal there was to use technology to automate the long tail of customer service issues. You know, how do you respond to tickets? How do you resolve what a agent would do, and how can you put technology in that place? We stumbled into the LLMs, GPT-3 in 2022.
Mm.
We built a lot of technology to make it easier to construct customer support applications. All of a sudden, we had this big beautiful brain that we inserted into our builder that all of a sudden you can construct these experiences with just natural language prompts. We pivoted from customer support automation to autonomous service agents, and that became one of the early leaders in that space. We launched this before ChatGPT dropped in fall of 2023. When we emerged, we were kind of an early leader in the space. Salesforce recognized that. They acquired us, and that became a core part of their Agentforce platform.
Airkit would've been like Sierra and Decagon today.
A precursor to that.
That space.
Yes.
Both of whom are presenting later, by the way.
Oh, great.
Yeah. Yeah.
It's a great space.
Okay. Yeah, great space
love those companies.
Yep. okay, Salesforce buys Airkit too, right, in November?
That's right. In November of 20-
This time, did you go to Salesforce again?
We went to Salesforce again.
Yeah
...the entire team. We had an opportunity to now take our technology and be a kinda key part of their Agentforce platform. What's funny, we're actually building on top of the legacy code base we sold them initially with RelateIQ. It was an incredible journey, very fortunate, and again, to be able to be part of Salesforce and building out the Agentforce platform was incredible. Again, for me, I was about to have a baby boy in April of 2024, and so after over a decade building companies, I decided to take a step back and enjoy some time with the family. I was doing that. I was dad of the year for about six months, and then I get a call before the election saying, "Would you be interested in being part of a, you know...
helping reduce fraud, waste, and abuse in the federal government? I raised my hand. I always wanted to go into government. I never knew. Growing out here, I never knew what that meant. Government was not something that was actually promoted or even talked about. I had an ability to go into the federal government. I joined the GSA, the General Services Administration. This was with the goal of reducing costs. GSA is actually a perfect vehicle 'cause the intersection-
Okay. Just so everyone's clear, this is DOGE, right?
no. This is the General Services Administration.
Yeah, yeah. You came in during the.
We work closely with DOGE.
Oh, you work closely with DOGE. Okay, good. Okay.
This is really the tip of the spear, right? I mean, if you look at federal spending...
This is tip of the spear.
This is, yeah.
In terms of agency?
Yeah.
Yes. Do you guys know the General Services Administration? Some people don't know. It's the intersection of a lot of the spend in the government, all procurement for common goods and services, all real estate, the largest real estate portfolio, fleet management, IT shared services. All this flows through the GSA. For me, running the GSA was this perfect spot to figure out how do you reduce fraud, waste, and abuse. I joined the administration, started January 20th, and it was a sprint. It was an incredible experience. A couple things kind of coming out of that. We had GSA was tasked to help implement the AI Action Plan. I helped draft that, helped implement that. A big part was the procurement, but also the adoption.
We had spent the first six months reducing the size of the federal government, reducing people, reducing contracts, coming out of that was the build back phase. This actually leads me to C3.ai, 'cause we're looking at AI platforms around procurement automation. You can imagine procurement, 40 different systems around contract writing. RFI process is super manual. RFP process, super manual. How can you streamline this? By the way, every agency was tasked to think about this. I came across C3.ai. The demo jumped off the page.
I'm like, "What is this?" Around this time also, I, you know, a big push for the federal government was to win this AI race, and a big part of that was to drive actual adoption usage. The other big piece of this was national security. We're re-industrializing in the maritime industrial base. We're trying to bring back manufacturing. That was gonna be a big opportunity for AI. The third part for me and why I joined the government was just the debt levels of the government was just. We all know this. It's, it's unsustainable. It's increasing. Despite cutting some costs, we increased the spending bill. I think the only way we actually reduce costs and grow and get out of this is to grow.
I can tell you, it's really hard to cut costs in the government. I think AI, automation, and robotics are the only way to drive productivity and top-line growth in this country. Those three are the big macro factors. I had an opportunity to join C3.ai, and it's primarily because the mission to go accomplish all three of those things.
Yeah. You joined in September of 2025, right?
yes, sir.
You'd run across it when you were at the government. What caused the phone to ring this time to say, you know, "Stephen, would you like to come run this company?
I was transitioning out of government.
Yeah
... and I was connected to Tom through a mutual friend.
Okay.
Tom was looking for a CEO.
Yeah.
They had a CEO search.
Yeah.
I was put in touch. I had multiple conversations, and we hit it off, and he offered me the opportunity to run C3.ai, and I jumped at the chance to do so.
Okay, awesome. When you landed C3, six months ago, what did you find?
Well, I spent a lot of time initially talking to our customers.
Yeah
... employees and prospects. A couple learnings. One, the problems we were solving were some of the biggest problems at the biggest stage for the biggest organizations. This wasn't just my experience in the past deploying AI for customer support. This was solving some of the largest mission-critical operations for Shell to manage 10,000 critical infrastructure assets, managing the uptime for the B-1 bomber for the Air Force, global supply chains for the largest berry producers. So these are massive, you know, operational challenges that we're able to use our technology to solve. That was very. For me, that's very much a mission orientation 'cause I feel like if these companies don't adopt AI, they're gonna be disrupted.
The second thing I found was the density of talent inside of C3.ai is some of the highest I've ever seen.
Mm-hmm.
Some of the smartest people, they also care. They deeply care. It's also an office, it's an in-office culture, 5 days a week, which is something I was looking for. The combination of density, talent, desire to care, and solving some of the biggest challenges in the world was very appealing. The other thing I found, honestly, we're an ambitious company. Tom Siebel, Edward Abbo, who started the company, 15 years ago, we're wildly ambitious. We're doing a lot of things. An opportunity I found immediately was, well, let's, you know, the noise, there's a lot going on in the AI market. There's a lot of noise we can discuss. How do we be focused on what we do exceptionally well, where we're best in class?
Those are like the industrial assets, industrial applications, that we've kind of pre-built apps. We do this across the value chain for manufacturing, healthcare, energy, oil and gas, and public sector, especially federal. Ability to double-click on that was gonna be an opportunity for me. Then also realized just the market itself was growing so fast. The innovation happening, it's probably the We always say innovation in the last six months has gone faster than maybe the last 30 years. Just as a point of contact, point of reference, how many people have bought a Mac Mini in the last 30 days?
Hmm.
Has anybody bought a Mac mini? Oh, you haven't bought a okay. You guys know what I'm referencing. The ability now-
Like I said.
Have you bought a Mac mini?
No, but okay.
Oh my gosh.
We use, you know, We're a bank, right?
Yes.
Like things are locked down, right? You can definitely not run Claude Cowork on a Citizens' machine.
Yep.
Right? I was like, "I cannot come to this conference without having tried Claude Cowork." I got a Mac.
Did you try it?
Right. Which, yeah, I tried it last night. It was mind-blowing. Like, you can be like me, where you literally are just I mean, I talk about this stuff on TV. I hadn't actually used it yet, right? It was insane, right?
How was it?
I had to use it. I did it with my wife's calendar.
Okay.
because my calendar is all, you know, locked down. We just, we picked the very first thing, which was, you know, can you optimize this calendar for me?
Yep.
You guys know how this works? It's crazy. It launches a browser. It goes through, it takes screenshots of your calendar, it takes screenshots of your appointments. It then, you know, consumes what's going on, and then it's like, "Hey, you know, I see that you have," you know, true story. This was my wife's calendar. It's like, "I see you have a faith sharing that you're hosting at your house at 11:30 on Tuesdays, and you have another one that you're hosting at 11:45. Are these really different things or are they the same one?" Click. It's the same one. "Do you want me to fix your calendar going forward so you don't have two things?" It just keeps going.
Yeah.
It just organizes your entire week. Yeah. Anyways.
Can I ask you, how hard was that to set up?
It's so easy.
easy.
Easy. Now, that's not true with, I don't know if Nick is in here. The lights are blank. Nick right there spent a lot of time trying to set up Claude Bot, now OpenClaw.
Yep.
Right? Did you ever get it running?
No.
No.
Yep.
All those TikToks where people are like, "Oh yeah, I'm not a developer." That's hard.
That's hard.
That one's hard. Okay, we gotta cut to the chase. Q3 results were brutal, right?
Yes.
Brutal. Like, I mean, revenue dropped, I forget how much. I could look.
Okay.
You know? What is going on in your business?
Yeah. yeah. Very disappointing Q3 results across, especially North America and EMEA. Full ownership. Like, I've been here 6 months, that's on me. couple things I wanna highlight. Again, I've spent six months talking to customers. Let me highlight what I've learned.
Let me just throw the numbers out for people.
Please.
Revenue came in at $53 million. The consensus was at $76 million.
Yep.
Right? That's like. We don't see misses like that very often.
No.
in software. yeah. What was going on?
Very disappointing. Just flat out miss. North America, EMEA, hugely disappointing. Just full ownership of that, full stop. Let me unpack this though. We have a technology, when we focus on the right customers and solve the right problems, it's incredibly successful. Like, when I stick to industrial applications for those solving the across the business value chain of manufacturing, oil and gas, energy, and public sector, we are phenomenal. We're doing too much. When I come my six months in, you know, I think that there's so much going on in the market today that unless you have a very laser-focused message of what you do and how you do it that's unique, you're gonna get lost in that noise. I think we're victims of that.
You know, we were pioneers in enterprise AI a decade ago, That was novel. That is not novel anymore. You hear this from OpenAI, Anthropic, Databricks. This Enterprise AI is a massive market, You can't be a general purpose tool in this. For us, the opportunity and the actions I'm taking right now are literally to focus on where we're best in class. What have I done? I've observed for six months. I've talked to customers, partners, and our employees. I'm saying immediately let's cut costs, restructure the cost structure, reduce the cash burn, and focus on long-term growth. Number 2 is flatten the organization. Over time, I saw this in government, over time you get layers of management. You get layers upon layers, and that removes me and the executive team from the customers.
I took action in Q2, where I actually eliminated those layers. The sales team reported to me, and you saw those results, a 130% growth year-over-year. Why that's the case? We were able to get much closer to customer, demonstrate value in days to weeks versus months, and that drove faster conversions. We weren't doing that in North America and EMEA. My process now is to flatten the organization, eliminate a lot of those layers of management. Those sales teams report to me now.
In North America and EMEA, you know. What I did in Q2, I did Q2 with Federal. I'm doing this now in North America and EMEA. I'm gonna get much more involved in these deals. If you think of I've done startups, I've done government, kind of like in the middle. I'm gonna get much more into the startup side in terms of speed of execution. That's immediate. The flattening and the focus. When we focus on the right types of customers, right? Industrial applications, public sector, just do that. What does that mean? asset reliability, predictive maintenance, supply chain optimization. In the Federal space, we will win. I know this. We just can't do this long tail of everything we've, you know, in the past I've done 130 applications.
That's great when AI was novel and new and you're trying it out. It's not great when there's so many general horizontal platforms you can choose from. Unless you have laser focus with a clear value proposition, you're getting lost in the noise.
Okay, say that again. I think that was really important.
Exactly that way. Well, let me highlight. The market is competitive. There's a lot of horizontal players. You have everything from an OpenAI Anthropic, to a Databricks, to the hyperscalers offering AI solutions. The market's confused. Our own team, everybody in this room, if I asked you, "What is an AI agent? What is an LLM? What is AI? What does machine learning?" You're gonna give me different responses. Now imagine you're a CEO and a CFO trying to make these choices. Unless you have a very laser-focused message of where you deserve the right to win and you have references which actually show proof points, and the number one proof point, have you done this that demonstrates economic value? That term, economic value, is gonna be the number one thing you're gonna hear every single company talk about.
'Cause unless you can show results, co-CEOs aren't gonna wanna do pilots. They don't wanna do proofs of concepts anymore. They're exhausted. They wanna go faster and bigger. Maybe full circle, we are focusing on fewer applications, right? The asset reliability, supply chain optimization, and the federal space. That is where we have clear customer success stories, lighthouse accounts, and we're gonna double-click into that and double down on that. That's the focus piece. On the sales side, I'm gonna highlight, customers do not wanna keep piloting software and AI. They wanna do broad AI transformation. That is full stop what I'm hearing from every CEO, and they're not trying to buy a one-point solution. They're trying to buy a platform as they think about transforming their supply chain. I just think about their asset maintenance and predictability.
We have the ability to get started fast with pre-built domain specific applications in those spaces, where we have deserve the right to win. We have a methodology and a playbook to go from one app to transform the entire supply chain. All I wanna do is focus on those initial applications in the federal space. I'm gonna keep coming back to that. Focus, focus.
Industrial and federal. Got it.
That's it.
Yeah.
Larger AI transformation. This is why Tom Siebel and Ed Abos started C3. It was the ability to leverage all the data organizations have, apply AI and automation to transform the business value chain for select industries. I'm gonna get back to our roots here.
What I'm hearing is that the way the market has changed is basically two vectors. One, the horizontal players obviously have a much higher profile now, right? Everyone's talking to Anthropic and OpenAI.
Right
... and Databricks, who's presenting here. Customers wanna do broad transformations, not pilots, whereas a year ago, it was, you know, we don't trust this technology throughout our organization. We wanna find, you know...
That's right.
specific areas and just solve that first.
A year ago, we were talking about hallucination risks.
Yeah.
Can we trust, you know, data inside these models? Now there's not even a question like can you trust or if. It is, how fast can you get this deployed?
Wow.
If you don't do it, your competitor is gonna do it.
Yeah.
Some individual in, over a weekend is gonna string together autonomous agents to re-revolutionize your department. Now, you know, the misconception, I think, and people are getting wrong, this idea that the LLMs are gonna cannibalize or commoditize enterprise AI is fundamentally misunderstanding the difference between predicting the next word or the next line of code and predicting financial or physical outcomes for business. you know, we don't compete against LLMs. They're our partners. We help orchestrate those LLMs. I think that's a big distinction because the problems we're solving, supply chain, asset performance, national security, contested logistics for the Navy, these aren't probabilistic problems. You need deterministic accuracy, and that's exactly what C3.ai has built.
All the tools, all the layers on top of the LLMs to not just understand the world, which is their phenomenon, but to actually operationalize them at scale. I think that to operationalize requires what C3 has built, which is the data integration layers, the data ontology layer, the governed agentic execution layer, and then the interaction layer. I think this is what's so exciting right now, is that we can actually transform a business. You know, these aren't like businesses that started two years ago. These are some of them are 50-year, 100-year-old businesses that are looking to completely transform themselves. That's the opportunity we have ahead of us.
Okay, one more from me, and then we'll have a couple minutes for questions. There are all these articles about C3 potentially merging with a private company that I think had everyone sort of scratching their head. Can you comment on that, and can you comment more broadly about how you think about sort of strategic outcomes for C3?
Yeah. Can't comment on any market rumors.
Yep, fair enough.
I would say, any acquisitionary, you know, M&A would always be evaluated by the board. I can tell from my perspective, I'm focused on organic growth at this company.
Mm-hmm.
We will evaluate M&A opportunities as they make sense in the future for stakeholder and our shareholders. For me, I'm 100% focused on organic growth.
Okay. All right. Any questions from the audience? I do have to warn you, we are really blinded up here. Yeah, go ahead.
I'm curious, what percent of, like, lifetime R&D spend do you think is kind of relevant for the direction of the business that you wanna take it, I guess, if that makes any sense?
The second part on the sales and marketing side, what percent sort of sales and marketing spend is kind of relevant for the categories that you're going to focus on in the industrial and government side?
Yeah. first question.
Just repeat the question.
Oh. Oh. First question is, how much have you built is actually relevant going forward? Is that more or less what that is?
Yeah.
Yes. I would say, highly relevant, but also transforming. What have we built up to this point? We've built the data integrations. Like, you think it's the evolution of C3 Energy, C3 IoT, all integrations into SAP as an example, right? Inventory management systems, production schedules. Those are highly relevant. The ontology layer, the data semantic layer, understanding the relationship of that data across your business, highly relevant. The third, what's fairly new, is the agentic execution layer. This is how do you keep the human on the loop versus, like, drowning the data. That's very relevant. That's leveraging the policies of the enterprise and the guardrails. The last is where actually We actually started in, which is the application side, so the interaction layer.
I would say the interaction layer, what's so exciting is how you build all this. Historically, that's been coding. That's been a web GUI. You're seeing how fast putting now AI inside of this to reimagine the ability to take a mission-critical idea to build a mission-critical app. That time is being condensed dramatically. We'll have a lot more to share on this, but I think that's where this is going. I'm very excited for that. In terms of marketing, sales and marketing spend, I mean, there's the focus, so we're gonna see much more of us speaking thought leadership around these exact use cases and industries I'm talking about. I don't think it's a throwaway, but it's probably a refining of that message. Enterprise is AI, a phenomenal message, great category to be in.
Where exactly do you deserve the right to win? That's gonna be the double click that I'm focused on right now. Highly relevant. That's why I joined, was like, they have all the components. Now they have this beautiful brain. Like, how can we reimagine how to build enterprise applications a new way?
If we have time for one more quick one. Yeah, please.
Where do you see in terms of, like, market segmentation, small business, inside business enterprise, the future opportunity being for your company? Do you see a lot of opportunity in this market, or do you see generally more adoption in the future in enterprise?
Enterprise.
100% enterprise.
Okay.
Large scale, Fortune 1000, public sector, federal government, that's where we have unique ability to win.
Do you foresee a future where it's more broadly adopted into smaller markets?
Not in the near term. I mean, I think I could see that, but I don't think we need that to win. Like, enterprise AI, large transformations, that's the focus today. Yeah.
All right. Well, Stephen, we could go another 25 minutes easily, but we're out of time. Thank you so much for coming.
Thank you.
We really appreciate having you here.
Thank you very much.
Yeah.
Thank you, everyone.