This morning, there's a little bit of history here. Thank you. Let's try to close that door. Gary, I think you can wait a few minutes. Thank you. I will play you in tennis. Here is the thing. Many people do not know the origin of LD and why we started conferences. Our next presenter does. It was December of 2007, we were at a conference with my Analyst, and someone said something very derogatory to him, and it really pissed me off. The last thing that they said is, if you think our conference sucks so much, why don't you have one of your own? You never want to say that to a Persian guy that has a lot of disposable income and no hobbies. We did it, and not surprisingly, they don't do a conference anymore.
It was really hard to find people to come to the conference when we were doing it back in April of 2008. There was one company that it didn't matter what date, where it was, they gave me their full and firm commitment that they were going to show up. This company, 8x8, EGHT, Bryan Martin, Joan, Hugh, these guys were the first company to sign up for LD Micro. I have to tell you guys, we just hit 2,100 companies. All it takes is one. Thank you for being there. Bryan, I'm sorry I wasted three minutes of your presentation time, but in this aspect, I would love everyone to give a warm round of applause to the first company to ever sign for LD Micro, EGHT, Bryan Martin, Chief Technology Officer. Thank you. Thanks for being here.
Thank you.
Thank you.
It's really hard to present when you're in tears, so thank you, Chris.
Thank you.
Hugh Rees was my head of sales, also still in tears. Joan Citelli was my in-house head of IR and PR. We had a wonderful dinner last night, which was really amazing. I don't think Dan's. Yes, I'm here. Dan. Dan was the CFO that turned us cash flow positive in 2010? In the June quarter of 2007. June quarter of 2007. He's the sharpest CFO. He's now a venture capitalist. I think he has an open fund. If you have money available to invest, you should invest it with Dan. Because the only reason we're in business in 2026 is because of Dan Weirich. Carbide Ventures is the name of his fund. I'm joined by my current in-house head of investor relations, Kate Patterson, who's over here in the green.
We're here all day, so if you would like to meet with either of us, we are here. Excuse me. I will try to pull my emotions together. We love LD Micro, by the way. Forward-looking safe harbor. All the numbers I'm going to show you are June quarter numbers. Our earnings for the September quarter is going to be November 4th, which is the day after the election, so I'm pretty sure no one will be looking at our earnings. Hopefully, some of you will remember us. Ticker symbol is EGHT. It's EIGHT without the I. We've been in business, I've been with the company 36 years. I've done almost every job. I take the garbage out. I'm currently on my second tour of duty as Chief Technology Officer. It's what I love doing. But I did serve as CEO for 11 and a half years.
That's when I met Chris. We have posted 22 consecutive quarters of operating cash flow. The CEO between me and our current CEO, Sam Wilson, who's been in the chair just coming up on five years in January. I hate debt. I didn't have any debt when I ran the company with Dan. Our current CEO hates debt as well. There were two CEOs between me and our current CEO who loved debt and ran up a ton of debt, and we'll talk about that in a minute. But we have got just over, I think it's 43% now of our debt off the books. And we are committed to getting to a zero debt state again. We have spent about 20 years building a global telecommunications network. I know the word carrier has negative connotations to it. Low margin, infrastructure, all these negative things.
But in a world of AI, you can't have artificial intelligence that doesn't communicate with the outside world. So AI has to send text messages. We call them SMS messages. It has to make phone calls to people. It has to send video messages. And 8x8's got an incredible global network. Even as in North America, people like AT&T want to shut their network down and get out of the business. We're investing in our network. We have more than 400 patents of how we do that. And we're also doing AI as well. We have a great product we introduced this year called 8x8 AI Studio. In my one-on-ones, I'm showing examples of what we can build with it, how easy it is to build with it, and I'll show you some examples of that coming up. Last fiscal year, we did $736 million in revenue.
We are not small. We're trying to grow that to $1 billion. If you look at the breakdown of that revenue, 50% of that revenue approximately is what we call unified communications. It's business telephony. When we started the business, we sold to consumer. Everything that we do now is B2B. So it's all businesses that we sell to. They're small businesses, they're large businesses. Some of our businesses are people like Mediterranean Shipping Company, PetSmart, very large companies that you've all heard of. But I have a lot of small companies, too. So that's 50% of the revenue pie. Approximately 25% is contact centers. These are call centers that we power. And the other 25%, which is the fastest-growing part of my business, is messaging. So it's sending text messages. There's a new messaging standard called RCS, Rich Communication Services.
It's kind of like the WhatsApp over your phone. So you're getting some of these. If you have an iPhone, it'll actually say RCS. You're getting these RCS messages today. But we hook up with WhatsApp, with Viber, all the things that my kids like to send as well. This is our last quarter. This is the June quarter, $190 million in the quarter. We're growing kind of 5%-6% is our growth rate. We acquired in 2022 a company called Fuze. It used to be known as ThinkingPhones. It took us about three years. It was kind of a rocky road to move that base onto the 8x8 network. We finished that at the end of last year. Had a lot of churn and some headwinds of moving those customers over. So we've been reporting excluding that base that churned off.
Our growth rate without Fuze churn would have been 8%, so a little faster growth if you take that away. You can see 10% operating margins there. I already mentioned the cash flow that we've been throwing off. We have $92 million in the bank at the end of June. This is what the annual numbers look like. Again, without Fuze, 6% growth rate over the fiscal years. I talked about the carrier infrastructure. We're also applying AI to these customers. We capture, think about it, not just the voice, but every message between employees that they're communicating with each other. Every Teams chat goes across our network, every text message, emails go through our contact center platform. Any message that generates a lookup in a CRM like a Salesforce, if you run your business on Salesforce or HubSpot, we integrate with those platforms.
The data that gets pulled up out of those platforms is also stored in a data repository that we maintain. Then we make that available to the AI. All of this data is being fed into AI engines, and I'll talk about that. We're using that in our AI Studio product to help these businesses make decisions about how to run their business better. The vision we have for where we're going is to be a platform that drives all of those decision-making capabilities. You think about a business, the most valuable data that you have out there is your communications. Contact center worlds, it's the business talking to its end customers, and that's where AI started, but it's just as important to know what employees are saying to each other. Let's say I have one of my customers calling in with a billing issue.
Great. We've got a billing issue. Maybe there's a problem with the billing platform, maybe a tax rate didn't get updated. All my customers in the state of Illinois have an error in how their taxes are being rated on our platform. What I want to know is not just that I have that issue. I want to know how my billing team starts communicating internally to 8x8. How are we resolving the issue? How did the tax rate update not happen properly in the first place? That's going on in email conversations. It's going on in Teams chat conversations. I want to know how my employees are communicating. If you're not the system of record inside the company employee to employee, you're only seeing part of the problem.
You're seeing how your customers are complaining about an issue, but you're not seeing how your employees are communicating with each other to run your business better so that the problem doesn't happen again. That's what 8x8 brings to that interaction. That's how we see the problem. Most of our customers come to us with disjointed communication systems. They have Microsoft Teams in this part of the business. They might have Slack over here. They've got this Avaya system over here. They acquired a company that was running on Cisco CallManager over here. They've got this mismatch of data all over the place. We unify that, so it's one data store. By the way, they get one bill at the end of the month. Globally, they've got a corporate directory that's all unified. If I want to find Fred in accounting, I can find Fred.
We've got AI on that lookup. When I search for Fred, I get a different list of Freds than when someone else searches for Fred based on the Freds they're most likely looking for. There's AI across everything that we do. We believe the conversations become the system of record. In the old days before modern AI, I would go to a class, and I'm listening to Bryan up here. He's teaching this class, and I'm kind of half paying attention. But every now and then, Bryan says something important, and you might be taking notes, right? Oh, Bryan says the conversation's most valuable data on the record. I'm going to write that down in my notebook. At the end of the quarter, I look back on my notes, try to remember what the professor said. Oh, I got this exam coming up.
I got to study my notes. Ah, that's some of the important things he said. That's how CRM systems used to work. I visit a customer. I take some notes. Customer seems interested in the products. Next visit, we should probably present this demo, do this follow-up, maybe do a quote. It's all notes in a CRM system. In the world of modern AI, what's happening is we actually take the complete recording of my presentation to you. Word for word, exactly what I said. We take the slides, the presentation. We take the questions that you're going to ask me. It's all recorded. Not only is it recorded, it's signed. We know it's authentic. It wasn't tampered with. No one modified it. We even take the consent. We have the consent.
I probably clicked something when Lahiji. When I agreed to come to this conference, I probably gave him permission to record it, to use my likeness. Someone took a photo of me. I gave him permission to use that photo in his advertising for next year's conference. All of that gets stored in a record, which is a perfect record of exactly what happened at this meeting. Which do you think is more accurate of what I presented today? The little notes that I scribbled or didn't scribble, or the exact record of what I did? Obviously. But I have a tool now that AI can immediately go into this exact record, all this data that I have on what was presented, and give me a real-time summary of what I need to know on the fly.
These CRM reports were built years ago when no one knew what the query was going to be. But I'm going to go into this customer, I'm going to do a sales call, and I need to know exactly what do I need to know in this presentation I'm walking into. If I have the exact conversation stored, signed, exactly what happened, what was presented the last time someone walked into this customer, that's what's valuable. That's what 8x8's trying to deliver. We call it Conversation OS. In the world that we live in, the most difficult channel to deliver that on is speech. I love this photo. This photo also tears me up every time I see it, because speech is the most difficult channel to capture right. Because it's got background noise. It's hard to record a perfect record. The mic might be having problems.
I'm on a cell phone that's breaking up. I got traffic noise with ambulances going by on Sunset down there. Speech is a very, very difficult channel to capture correctly. Moreover, the people out there building speech bots typically will do everything. There's thousands if not tens of thousands of companies doing everything on the left side of this slide. They can all build you a bot that's agentic. They can all build you application logic that processes the speech, but very few of them have what I started with, which is that carrier infrastructure to connect you to a global voice network, to send a text message, to connect you to WhatsApp and Viber, to basically do all the regulatory and compliance work, make sure it's legal, make sure it's secure, make sure the data's not going to be stolen by somebody.
Most companies that are trying to deliver automation via AI have to stitch together everything on the right. I can tell you, 8x8 spent decades, literally decades, Chris told you how long ago I've been working on this, putting together solutions that solve the right side of this slide. You as investors may think, "Ugh, carrier's infrastructure. I don't want to invest in that. That's old school." Well, guess what? The right side of this slide is what's going to keep me employed for the rest of my life because this stuff is hard. You go over to Europe, you try selling to a German automaker, and I've got 12 of them that are big customers of 8x8.
You go through six months of security audits explaining how you're going to deliver the right side of this slide, and they're not going to deploy a voice bot until you convince that auditor that you can deliver on the right side of this slide. The right side of this slide is AI durable. You're not going to go to Claude and say, "Oh, Claude, can you deliver me a toll-free number in Singapore or Berlin?" Ask Claude to do that for you. It'll spin for a while, and you won't get one. Ask my AI bot on my platform to deliver you a toll-free number in Berlin, you will get one. That's the difference of having the right side of this slide. It's very important stuff when you're delivering communications automation. The other thing I want to point out, Kate actually drew this diagram for me.
Actually, we've been trying to draw a picture of our platform for a long time, and there's really bad drawings of our platform, including some that I've drawn. I really like this version of it. What I want to stress is the AI layer down. Read from the bottom because the bottom is that very boring to investors, the carrier infrastructure is the bottom of the slide. That's where the AI lives. The AI is the box second from the bottom. When you have the AI living down in the network, the advantage I have is the AI can operate before the data even hits the applications. If I'm doing fraud prevention, I want to stop calls that are coming into the business that are fraudulent. I don't have to wait till it gets up into the applications. It's too late.
If it's already ringing phones in the business and then I decide that it's spam, it's too late. It's already disturbing the business. It's already got fraudsters in the business. We're stopping that before it gets into the automaker in Germany. We stop it down at the network layer. No one in our space is doing that other than 8x8. It's a very unique way to think about where you deploy the AI. If you don't own those network assets, if you don't own your global network, you can't do that. You can't outsource to British Telecom and then go to British Telecom and say, "Oh, do you mind if I put my AI assets in your network?" They won't answer the phone. That's not going to happen. That's kind of cool.
The other thing I want to hit on is, I've spent my entire career trying to convince the world to move into cloud communications. There's many companies, especially in France. France is kind of, I don't know, 25% moved to cloud. They're still running old Mitel, Avaya, on-prem stuff. It's taken forever to get in the crossing the chasm vernacular. It's like the late adopters, the laggards. That end of the technology spectrum. It's taken forever to get people to move to cloud in some countries. The U.S., I think is like 60%, 70% moved. Think of banks, law firms, very conservative businesses, German automakers. We're still trying to get them to move to the cloud. Every single one of those industries or those types of companies are all early adopters of AI. This revolution of AI is going to be a totally different playbook.
Totally different. Something totally new. We're seeing the numbers in our AI deployments. Absolutely, the early days, when Dan and I were running the business, we had a spreadsheet, we called it double data, because month-over-month, our business would double, our numbers would double. I have a similar spreadsheet with our AI numbers these days. Our AI deployments, our AI results that we're seeing businesses as they deploy this stuff are just through the roof. These are year-over-year numbers, not month-over-month numbers, but you can see what's going on in our business. We have 3,200 agents in production doing production workloads on our AI Studio product. We launched this in April. I've only had two viral products in my career. The first was our video meetings product in COVID. We had every Italian classroom running on our meetings product.
It wasn't fun because it did scale, but it scaled not smoothly. That was interesting. This thing is reminding me of those days. It's running pretty well. There's a couple customer examples. This is a woman named Janelle who runs a wholesale supplier to companies like Home Depot. They sell nuts and bolts and screws, and she's been standing up bots in her business, and she can't keep up with the results that she's getting. She's on our website if you want to see some of her testimonials. This is a power company in Ireland that's doing the same thing. They've got about 800 bots running right now. They built them all their own. You build them in English. You don't have to go to 8x8 University and get certified in code and things that you don't understand. You can build these bots.
Kate, my IR person, she has an IR bot on her IR line that she built in 5 minutes. It is fun to build. The other thing I want to touch on real quick, Microsoft Teams. We are a partner of Microsoft since about 2020. I have about 1 million seats of Microsoft Teams that run on my network. I want to take that to a much larger number. We acquired a company in January that is called Callroute. You can go to callroute.com and see what they do.
That software we are certifying with Microsoft right now. Most people use Microsoft Teams, especially larger enterprises, but they use it for video calls, or they use it for team chat. They do not use it to make phone calls. 8x8 makes it, in a couple clicks, you can actually power your Microsoft Teams to call your spouse and make phone calls.
Most people do not understand that, and they do not know how to do that. We make it super easy to do that. Once you do that, you can either use the AI I have been talking about, or you can actually use Copilot. We are very friendly with Microsoft, so we are happy to send that data to Microsoft. A lot of my direct competitors want to use that data for themselves and keep it in their own AI. One of the reasons I am friendly with Microsoft is we are very open to sharing that data with Microsoft. Happy to talk about that. There are a lot of products we can add on to the Microsoft deployment once we integrate with them. Microsoft is not a big believer in things like fax. So they are happy to have us do things. Microsoft deploys in 35 countries.
8x8 has Microsoft Teams support in 59 countries, so almost double what the Microsoft footprint is. They cannot walk into all countries, especially in Europe, without triggering antitrust and things like that. Nobody cares about 8x8. We are too small to have that sort of thing. I talked about the debt at the very beginning. So this is, year-over-year, how we are reducing the debt. So there is about little over $300 million of debt on the balance sheet. Was not my fault, but we are getting rid of it. So, we are on a good trajectory there. We throw off about $20 million a quarter in operating cash flow.
That is how we are paying this down. So I will close. I will leave a couple of minutes for questions. I will leave this slide up so you can read it. Who has questions? We have just under 4 minutes left.
Mr. Cannon, go ahead.
You have a business where you are charging per seat for these call centers. As these call centers use AI and make them agentic and move those seats away.
Yes.
Are you actually making money by this AI integration, or are you losing money because you are losing per seat?
It is a great question. My year-over-year seat count of human contact center agents is actually up. It is not going down. I think it is a huge false investor thesis that AI is eliminating humans in the contact center. I blame Benioff, because he, December 2024, said Agentforce 2.0 is going to get rid of all the humans. I think he has retraced that statement. He says, "Oh, the AI is there to augment the humans, not replace them." His stock has suffered as well since that time. Number one, we do not see it. Number two, all this AI application sits on top of that seat revenue and the number of seats is increasing, not decreasing.
Okay.
Great question.
Tom.
You mentioned one of the slides that paid AI adoption is in the early innings. Is that for you guys or the industry? I am just trying to figure out how to quantify that for your overall business.
Yeah, also a great question. The reason I think it is in the early innings, the deployments of AI are still very sloppy. When you deploy an automation in a customer service scenario, there is two things you have to do right. You have to fail fast, which means if you do not exactly know what the end customer is asking, you got to get the interaction to a human immediately. Do not waste minutes of my life with your bot that does not know what I want, so get me to a human. When you transfer to a human, there is two things you have to do. Number one, they have to answer quick, less than 20 seconds. Number two, they better answer knowing who you are, what the bot has tried, and what you are asking.
If you start re-authenticating who are you, what is the last four of your social, what is your birthday, I am ticked off because I just wasted two minutes of my life with your stupid bot, and now you are asking me to start over. If you do what I said, you answer in 20 seconds, you know who I am, "I am really sorry my bot screwed you up. Let me get you the answer to your question," your CSAT will go through the roof.
So many people do that wrong, and I do not know why. If you do it right, your customers will love you because the bot answers 24/7 immediately. If you get 60%, 70% of the bot answering the right question, and the things it does not answer, you get a human to do it immediately, your CSAT will skyrocket. People do it wrong. My favorite is Comcast. I do not know who your cable company is. Their goal in life is to get you to not talk to a human. If you do get a human, guess what they do? "Who are you? What is the last four of your social? Where do you live?" Ticks me off. Who else?
Bryan, I think you and I are Comcast customers.
Yes, we are.
Yes.
Well, they have the best data service on the planet.
No doubt. So, question. I entered the stock in early 2004.
Yes, sir.
You were profitable, you were growing, your churn was less than 1%, you had a bunch of cash on the balance sheet.
Yes.
My logic in that report that I wrote was that your intellectual property was worth more than the market cap at that time, which was approximately $80 million-$90 million.
Yep.
Are there any synergies that you can make from what happened in 2004 to where you currently stand on that podium today?
In IP or a bigger sense?
Just bigger, just as a collective whole.
Well, the AI thing is bigger than the internet, so yeah, I think so. The AI revolution we are about to go through, I think we are in 1996, 1997, except this is a bigger thing in my opinion. This revolution will be 10x the size of the internet.
Wow. Okay. With that, thank you so much.
Good to see you. Thank you, my friend.
Everyone, thank you.