Latest / Machine Dreams / Fortune 500 Companies Are Deploying AI Agents Faster Than They Can Secure Them
Transcript
- speaker-0: Hi everyone and welcome again to Machine Dreams, the show where we talk honestly to the people at the very front lines of the AI industry about what's working now and what the future might look like ⁓ some years from now. My name is Colaule Samuel Adibayo. I'm a technology analyst covering AI, cybersecurity and enterprise technology. As always, I'm joined by the fabulous one and only, Very well, I miss you too. It's been a couple of weeks. Right? It's good to see you again. speaker-2: But in your. Pleasure to be here. Thank you for having me. speaker-0: Great. Thank you so much once again, Affair, for joining us. So as Lea said, we had a great time chatting with Garl the last time. He made a case for, a technical case for AI agents, AI agent security, and know, AI SaaS sprawl. And so I wanted to get a sense of what the conversation might look like from your vantage point as a CEO. What are you seeing in the boardrooms? How is the conversation shaping up right now? How is the issue of AI agents and agentic security shaping the decisions that people are making in boardrooms today? speaker-2: Yeah, so it always starts from the business. took a while for the AI to actually provide significant value that the board ⁓ will care. Once it provided, then they started using agents, ⁓ just a pilot and then it's expanded. Now the value is very significant across the board. And once one company, for example, a big telco is starting to use it, the CEO is public about it, all the telcos are running to do it. So it always starts from the value. How can I move faster? more effectively, more efficiently, and do things I never ⁓ did before. Once you do that, then what we call the ⁓ shit moment comes. ⁓ we're using so much more than we expected. We're using it across the board. What do we do with it? And then we step in. And the main goal for Rico is, not to say, okay, let's block everything and go back to the Stone Age. How can we enable the business much faster than ever before while keeping the visibility, governance and security? for these AI agents. Yeah. So, so I can mention the, ⁓ mention a few, a few companies, for example, like, ⁓ waste management and Nestle and one of the top three telcos and a bunch of banks, ⁓ insurance companies like TIA, ⁓ and others. And the, and the motive is pretty much the same, which means our business is running more and more on AI. How can you help? And it used to be, let's say 10 years ago, 15 years ago, we're going to block and they're going to beg us. in order to approve apps or any system. Now it's the opposite. The business is running so much faster. Anyone can deploy new agents. Even in banks, you'll be surprised how many unsanctioned apps and agents we find. Which means if you're a marketing person and you were told to run faster because the company publicly announced one of the top 10 banks, they're going to save $1 billion over the next two years. What would you do? You have to use AI, have to use those agents. You're not going to wait for the security to keep up. And now the tension between security and IT and the business is something that needs to be managed. speaker-0: Great. know, just on the back of the question that Leah asked, you've made quite number of hires this year. If I'm not mistaken, the number stands currently at 40. I don't know if you still have plans in stock right now, but I know that the numbers, if I'm not mistaken, stands at 40. And you know, those hires include the VPR sales and business development, VPR marketing, VPR product management. And what that points to me is that this is not just, ⁓ you know, some product hires that you're making, but also skill hires. And so I'm curious to know what that says about where your company is headed and what that also teaches us about how to scale ⁓ in the AI space. speaker-2: Yeah. So 40 is in the last two and a half months. It's not the entire year. So we're growing faster than that. ⁓ the answer is we do that because we get a huge market pool and market pool. can measure in multiple ways. For example, when we have, ⁓ close to 200 % a attainment, which means they're doing double than what they should do continuously month after month, you say, okay, I need to hire faster. When the customers tell you. I never heard your name, your company name, but this is the best solution in the market. Again and again, you need to move to move faster on marketing. So on all of these fronts, because of the adoption in the market, we understood we need to move faster and basically go in and take the market. So we hired across go-to-market, as you mentioned, from sales to customer success, which is extremely important for us. and obviously marketing and then the technology as well. with multiple AI, PhDs, deep learning, machine learning, security researchers, because the technology is very deep on one hand, but with a few clicks and a few hours you get the value. Yeah, I think the backbone of what you do with the IDF on after every story before and after you basically debrief. And when you think about what we should do, how should we do it exactly to the details? And then you go and do it and go and debrief. Technically, everything is perfect. I call it awesome, awesome mentality, which is not what you're going to use in order to grow your company. So we are extremely transparent and extremely direct. And that's how we are with our customers and vice versa. Which means if this is great, let's do it more. If this is not good enough, why, how do we fix it? And let's go and fix. It's never blaming, by the way, with the relationship with customer and, and the, us as a partner, it's the same thing. They know they can tell us what is great and what is not good enough and to get the response immediately. speaker-0: You know, on the heels of that, I know that you're expanding into some new markets like in Texas and in the UK. And Texas for me is quite interesting because, you know, it's not the first market I think that you would most security companies will consider when they're trying to expand. So tell me what's your logic here? speaker-2: On, on Texas. So Texas has some of the largest, ⁓ fortune 500 companies and one of the highest concentration of fortune 500 companies, ⁓ in the U S and we found amazing partners and we found a great technology and more importantly thought leaders as CISOs, CIOs that know they need to step up, step up and take their business forward. ⁓ as you know, speaker-0: Yes. speaker-2: they're all talk to each other. And actually now we just signed with one of the top three telcos. And even behind our back, they had a call for another fortune 500 CISO and asking, that real? What do you think of this company? What do you think of the team? What do you think of the technology of the support? And when they start to talk and they say, yes, you can trust Rico, you can trust the team because I'm using it and this is how, then you can start accelerating. Yeah. So we have Boots on the Ground in the UK as well. We have customers in Europe who are fully certified in Europe and GDPR and everything you need to work with companies. And now we started to expand, including with Boots on the Ground, UK, Ireland, and other countries in Europe as well. And the reason is Europe is usually one to two years behind the U.S. in adopting such technologies because of privacy, because of other regulations. Now it hits Europe as well. the adopting apps, agents, AI faster than ever. We're talking to some of the largest organizations that told us a year ago, I wouldn't even consider, but the board came to us yesterday and said, you have to use AI, thank you. Yeah. So AI is important in the sense of you cannot say I didn't know and it just happened. So number one, you need to know if you have AI sprawl, people are using AI, people are using agents. Number one, you have to know. Number two, it's not an excuse saying the business is running so fast. I didn't have time to govern it. You have to govern it to a level of degree. And number three, then it comes to the level of the security. I'll give you an example. If I'm a marketing person, I can create more an agent that is connected to all of the backbone of the company, to Salesforce, to HubSpot, to all of the most sensitive information. We call those the okay messages. Can I connect to yes, okay, okay, okay. Boom, you created a monster. You're not a technology person. You couldn't care less about security. You couldn't care less about the privacy. You care about better marketing. And this is how, you know, security and a business evolve. And basically security now needs to move at the state of the business. And the business is dramatically fast. speaker-0: Over, it's really great that you just spoke about security ⁓ because I know that that's a major concern when it comes to, you know, AI agent, autonomous systems that can take decisions without you ⁓ being behind the scene. And you sit in the rooms where this decisions like boardroom decisions are made. And so I'm really curious as to how people in the C-suite or people making decisions about agentic systems they want to build or deploy are making those decisions and how much of ⁓ prioritization should be given to security ⁓ when you're considering deploying or building an AI agent. speaker-2: Yeah. So again, let's talk about the reason why they build their agents because my business, I'm the CEO of let's say 10,000 people insurance company. My business is going to run faster. My customers are going to get better experience and it's going to be more cost effective for me, which means let's do it. Then the second derivative of that, how can we make sure that we're not getting breached because we're a regulated company? Okay. So the boardroom starts with how can we expand the business and make it faster and more effectively? And then. Here it is, here is the money, AI security budget in order to go on security. So I can tell you that mythos and the mythos like models are now top of mind in every board discussion because what effectively happened, the bad guys now have the weapon to be 10 times faster than ever before. And if you do more of the same as security and IT, you will get breached for sure. The secretary of treasure. ⁓ treasury secretary ⁓ in the u s basically called the c o's of the big banks and told them one of you is going to get breached if you continue this way you have to wake up and the same thing And you mentioned that the most important word is change. You see companies today that adopt change in our space and companies that don't. If you don't adopt change in the level of weekly, you will basically die. Which means what is working, what is not working, what is new this week, what is new this day, how does it impact us? Mythos is a great example. We stepped in and changed and adopted, we call it the war room for Mythos to help our customers. It didn't happen, ⁓ it didn't exist a week ago, which means you have to continuously change as you mentioned. You have to continuously listen to customers, what they are doing and be two steps ahead of everybody else. So the company DNA that we built is always ask, ask why and not just what, because what is too slow. And if you really understand why and you have the best people and the best technology, you can move the fastest. speaker-0: Ofer, you've just talked about mythos and every time I hear about mythos, I get excited. I don't know why, but I just get excited. And that's because, you know, mythos sort of like puts a sophisticated tool in the hands of people who may not have technical knowledge and they can now launch malicious campaigns at scale, even without technical knowledge. And so this is a question I'm always curious about. What are you seeing? Who is winning the war? Is it on the offense side or the defense side? What types of breaches are we going to be seeing? or the next couple years and you know, where are we really going from here? speaker-2: Yeah, it's a great question. And it's always like a ping pong back and forth between the good guys and the bad guys. And by the way, it's not only mythos. Open AI has a model that is very close. Microsoft just released a model that is even better in some ways. So the fact that the bad guys will have a tool that allows them to reverse engineer every piece of software to penetrate in 11 minutes versus 27 days means that if you are a CISO of a big company, you need to be proactive versus reactive. It changes the entire game for the defenders. I need to understand what is the blast radius. I need to understand what apps, who is using them, what is the level of permissions, configuring them. All of those steps that before the Mythos days, it was fine to do it next week and next next week. So it basically changes the game for proactive approach in security across everything AI versus the periodic reactive. It is going to be way too late. ⁓ Yeah, I just mentioned a few of them. So I'm using heavily using a Claude and co-pilot and OpenAI and Gemini, each one for different things. ⁓ Mainly for text and the presentations and to summary, but also to take actions. Like the CEO dashboard that we have is now agentic. And by the way, it's not cost effective in the short term. We are paying a lot of money, probably too much for AI tools. and GenAI and vibe coding and all of that. But we are moving, I can tell you that last quarter we finished the quarterly plan in five weeks. This quarter, if we took the same quarterly plan, it would have been two weeks. We didn't write one line of code in the backbone ⁓ in the last six months, which means we have agents and skills and agents talking to agents and only then a human being supervision because we have to do by regulation. So the world has changed significantly across every aspect of the company, including myself. speaker-0: That's amazing, Luffa. And this is going to be my last question. You've talked a lot about human in the loop for agentic AI decisions, right? Decisions that are made by this AI agent. But of course, the question of responsibility and liability always comes to the fore. And you're very knowledgeable about this space. So I wanted to ask that question, because the question that keeps coming up in all of conversations that I listen to. about who takes the liability when an AI agent makes a decision that then, you know, crashes everything. speaker-2: Yeah. So it's good that the auditor and the governing agents don't care if it's AI or not. It's your responsibility as a company to make sure it's right. Because you have customers, because you have people that are putting their money in your bank or whatever the reason is. So on that front, it didn't change. What does change is that the baseline, the basic stuff and actions could be done by agents and AI. And then you have the level of how do you check it and make sure that it's good enough. Until now, the AI is not reliable enough to just to take the actions by themselves across everything because of hallucinations, because of mistakes. And even because you cannot explain to the auditor, Oh, I just pressed the button and something happened. I don't know. It's your problem. No, it's your responsibility as a company because you're ensuring, I don't know, 2 million people and you need to make sure that it's right. So the man in the loop still stays along the way. And obviously at the end point, we call it decision support system. Some of the actions are taken by themselves, but at the end of the day, you take decision faster. You take decision faster and ⁓ the AI is helping you, but it's not fully autonomous. By the way, unfortunately for us as human being, it's not long from now that AI will be more autonomous. So more and more positions. For example, Meta are going to lay off thousands of people tomorrow because more and more positions could be, ⁓ you know, automated and autonomous and make it fun. speaker-0: unfortunately for us as humans. speaker-2: Yeah. ⁓ Again, unfortunately or fortunately, depending on which side you take. If you endorse change, you are excited, you move faster, you can do things you could never imagine. And for us, it's by far the most exciting time in the company's history. speaker-0: I think that's going to be a great takeaway for many of our viewers who are concerned about AI replacing them. Just embrace change, right? Thank you so much for joining us on Machine Dreams today. This was quite exciting, even, and very insightful. Appreciate it. speaker-2: Totally. It was a pleasure to be here. Thank you for the conversation.