Episode 134
Listen to this episode on
Episode Transcript
Speaker 1 00:02
This is CyberSound, your simplified and fundamentals-focused source for all things cybersecurity.
Jason Pufahl 00:11
Welcome to CyberSound. I’m your host, Jason Pufahl, joined today, today, virtually by everybody. Michael Grande, CEO of Vancord and Rick Gouveia, CEO of SellBigger. Hey guys. Hey, hello. We had a podcast where we talked a little bit about AI, some of the applications that are out there, kind of our discussion a little bit about how to potentially connect some of these things together. Today we brought in Rick, who is the CEO of a company that is using agentic AI to do well, I think, you know, I’ll let I’ll let you really describe it before I butcher too much Rick, but I think largely you started in the sort of sales and business development space, but I think you’re starting to branch out a little bit more in your utilization of of agentic tools in particular, right?
Rick Gouveia 01:02
Absolutely. Yeah. Thank you, Michael and Jason for inviting me, pleasure to be here. So, yeah, we, we at SellBigger, we specialize in lead generation and meeting setting for our clients. We, we built a platform a couple years ago based on Twilio, and we connected to thousands of telco lines with that platform to allow us to do what has become omnichannel. But before this year, really it was LinkedIn, email, SMS, single threaded dialing, and then with humans. So we would attach that to humans, and we would go and make dials, book meetings for our clients this year with the oh, well, actually, two years ago, we integrated Twilio, came out, sent me an email, as luck would have it, their CEO sent the email. Pretty big company, and what some people may not know is they might have been, you know, they’re pretty much the original AI agentic AI platform. It just they’re glean all this sensationalism, glean make perplexity. Nan, they’re getting all the news, but Twilio has been around for 20 years and doing really well. They sent a message saying, Hey, we wrote to this unknown API from OAI, OpenAI, and would you like to try it? And I’m like, and you had to put your credit card in. So I’m sure a lot of people said no, but I’m like, Yeah, whatever. What could it cost? You know, what could it hurt? And all of a sudden, started doing these amazing things, like, you know, inviting people to meet face to face in Connecticut, and giving hotel recommendations. So, so we’ve moved over, and we used to have 100 sales development reps. Now we have five AI, SDR managers, and they manage 20 AIS each. So it’s been a transition, a pivot, a pivots cause a divot. So we took a little bit of a hit on revenue. And we took the risk. But, you know, I’m in pretty good company. ServiceNow took a risk. They turned a $35 billion company from ITSM to 100% agentic Salesforce is trying to do the same thing. And just in the last week or so, you see Palo Alto Networks and Cisco, you know, they’re obviously humongous companies compared to me, but they’re announcing it. And so I think we, you know, even though we’re small, we are a little bit ahead of our time, and now it’s coming to fruition.
Jason Pufahl 03:19
So alright, so drop a couple of things to say here. So I think you know, for everybody who’s listening, we’re working together, right? You are now helping us do some of our business development and lead gen work you made. You made an interesting comment to me. You know, it might be four months ago at this point already, where we were chatting and you said, Hey, if you had come to me and tried to do this four months ago, it barely worked, and you’re seeing the space evolve. It was monthly now, I think you might even say almost on a weekly basis. So I’m curious in your thoughts on how agentic AI is improving, the pace with which it’s improving, and sort of the changes you envision.
Rick Gouveia 04:07
Oh, okay, so last year I’ll start like just going into q4 of last year, the updates from open AI and Jim and I were the first Monday of every quarter. They they didn’t. They may have this, but they didn’t give it to us. Being how small we are. We don’t have a roadmap. We didn’t get release notes. We didn’t even get an email update. We just started getting one email with three paragraphs from open AI on the last update, you know. And it was so rudimentary, I thought, wow, this is not very professional.
Jason Pufahl 04:36
Yeah, it felt like spam when you showed it to me, almost, yeah.
Rick Gouveia 04:38
And so. So every Monday of the quarter, we’d get a new version, and I’d wake up and it was a box of chocolates, and we just didn’t know what was going to happen that day. Then coming into this year, around February it was, it was every Monday of the month, and now it’s twice a week. Crazy. The updates are coming twice a week, sometimes three times, depending on how you count it, I’ll get two from open AI and one from Gemini. So we have to adjust the system because we fell over to Gemini, usually on Friday afternoons, because open AI peaks and the latency drops. And then I read about this stuff insatiably, it’s really addicting to read about AI and agentic AI in particular. And then when I started seeing the Palo Alto network, CEO and the Cisco CEO saying, hey, you know, we’re really worried about the networks right now. There’s no security. And it was built for single threaded communications, not this massively parallel multi threaded comms. And then it started to dawn on me that, hey, you know, it’s not so much. You know we can code this to make it better, the technology is actually slowing down. The hardware is slowing down the capabilities of the of the software and and what I’ve also noticed is that, for example, Jason we’ll build projects for you, and a project we built a month ago is on old hardware. And so what we have is a capability to clone that project. But if we clone it, we get the old hardware capability. Instead of building a new one, now we get the newest hardware capability. And so we don’t even clone anymore. It was a feature that used to be very useful, and now I like we’re getting updates so fast. Just build a new project, even though it takes 10 more minutes, let’s build a new project for you, because it uses the latest compute, the latest GPU. And I can literally tell the difference between a clone and a new project on performance.
Michael Grande 06:32
So Rick, a question sort of relating to the business model, and I have a question on sort of the human resource aspect of things as well that you talked about sort of moving from more operators to more management of the of the AI systems. But you know, how complicated is it for you to sort of put your arms around the costing and pricing model on a on a regular basis? Is it a steady amount of of cost expectations so you don’t necessarily have roadmaps. Are they adding on, sort of modules and different technologies that allow you to sort of expand your offerings, and then you have to sort of keep up with that from a pricing model perspective, you just talk us through, sort of what how that process looks like.
Rick Gouveia 07:18
So I rely a lot on my background experience for this. In 2014, I became a VP of Sales at a company called Druva. Druva had a finance team, and then they had a cloud finance team. So their whole product rant runs in the Amazon and they found that, you know, if, if data wasn’t being accessed, they shouldn’t be running it on s3 they should move it to Glacier, where it’s cheaper, right? So I have a finance person that’s his whole job is to keep our open AI and Gemini costs down. But I’ll tell you, in the industry wide, I’m a small player, right? So they can raise the price on me every day, if they want. And what happened is coming out of this year, I thought we were in big trouble, because the prices in January raised 50% 100% right? And then in February, deep mine came out, out of China and was free, and it stopped them from raising their prices anymore. But now they’re starting to raise the prices again, because they’re deploying the, you know, the the latest and greatest NVIDIA GPUs, and they’re building these massive data centers, and they’re building nuclear so they’re they have kind of a cause to raise the price. If you want access to this technology, you’re going to have to pay. And so we charge by the minute sometimes, and sometimes that can be that can seem high if you compare it to a cellular bill, but we’re really charging by the minute of compute. And so yeah, we do have that that could overnight put us out of business and outside of our margin profitability. And we operate we, you know, believe it or not, we, we really operated on a thin margin. Anyway, I’m not commanding a premium. We’re trying to build business. We made way more with humans, but we paid more to the humans, so the margins were less, but the revenue was higher, profitability was lower.
Jason Pufahl 09:12
Yeah. I mean, you’re, certainly, you’re all in on the, on the agentic space, right? And I think that’s what you know, that’s really what brought us to you, and I think your were, I guess we’re taking that leap of faith with you. You know, because, because it’s challenging you the as now, as a as a consumer of your of your product, you’re listening to the calls. Some of them are just outstanding. Some of them, you’re like, Wow, I’m surprised sort of at how it struggled to keep up with the conversation, right? So you definitely see both sides. But mostly what I find is a little bit of latency is kind of a telltale sign that you’re that you’re often talking to, you know, an AI agent, and I suspect that’s probably where the costs come into play. You know you can, you can buy greater speeds and reduce that latency, I would guess.
Rick Gouveia 10:06
So I don’t think I’ve ever told you this, but the more lines you open, the less latency you have. So if you were to open up 60 lines, you would have probably no latency. But the problem is, if you do that, you’re going to burn through all your minutes or credits almost instantly, right? So if you have one line you’re going to you’re going to incur the most latency. They they prioritize the compute to the people pulling the most compute so they could charge the most. Nothing I could do about it. There’s no like right now. It’s the wild west in networking and agentic. Ai. There’s no load balancers for agentic ai. There’s no real, you know, you guys could argue this with me, but there’s no real, real firewalls that can handle what’s coming in the next six months.
Jason Pufahl 10:48
Yeah. I mean, you we’re talking MCP, if you can’t talk to the firewall vendors about anything security related, relative to inter AI communications or even the data going across.
Rick Gouveia 11:00
Yeah, we, I mean, the agentic AI can get through anything, I feel like, right now, it’s, it’s really like, when I was back in 1990s you could most people use first initial, last name as their password. So if you want to hack them, you just wander around in a company until you find somebody that did use it, and you’re in. And then they came out with 2FA and all that. But now with these multi threaded you know, like I said, one agent could connect to 1000 threats, Salesforce, HubSpot, Vonage, Twitter, not Twitter stripe, sorry. It can connect to Twitter, meta and LinkedIn. That’s just three or four. Those are all over the cell the the telco lines, the networking lines, so, and they weren’t really made for that. I’m not the expert, but I’m listening to these CEOs talk about it, and they’re, they’re basically saying our engineers are innovating right now. And I could just tell kind of the look on Chuck Robbins face, I’ve never seen that before. He’s usually very stoic and confident. And he was like, you know, I think we got caught flat footed. I don’t know, though I’m not the expert. Chuck’s the expert is a really big company that could come along and, you know, and release stuff tomorrow. And I think they probably are. But if you I’ve dealt with Cisco for a while, and they release new stuff, and about six months later, everything takes a hold a foothold, gets a foothold. So we have another six months between now and the end of the year that I think, Ooh, there’s a room for innovation and, and, and I’m not trying to pitch Vancord, but I will. I think Vancord a great opportunity. Every single client we have now is, is adding agentic, AI and then some So, yeah, great. You have a sales development rep. What else do you have? I don’t want to just buy a just buy a single product. Whereas in January, like, I could light myself on fire in the lobby and not get water. And now they’re like, can you make these five other agents for us? And at first I said no, but now the platform capabilities are so amazing that we can, and not a lot of people. My son’s making agents, and he’s getting paid, he’s 17 so, you know, not a lot of people know how to make them. It’s going to become more and more, I think. And then you also have to understand where the agents are going. They’re going to be put in the robot brains. An agent is actually going to be the robot brain. So if they want to be an agent of Taekwondo, you download it from an agent marketplace, you load it into the robot. Now he’s taekwondo expert.
Michael Grande 13:22
It’s so interesting. I was having a conversation with someone talking about the future of education, to a certain extent, that’s going to be sort of dominated in that sort of like personal robot, AI space where you know, from a tutoring perspective, but it’s incredible to think about, my question, I guess, for the short term, goes to your focus more on sort of hiring from a human resource perspective, the doers right, the SDRs that were out there making the calls now to the managers, how hard is training up from a management perspective, and getting folks to understand with sort of this constant and sort of chaotic evolution that we just talked about, you know, having your management team feel confident that what you’re delivering and what you’re saying that you’re able to deliver is actually what’s going to be happening.
Rick Gouveia 14:17
All right, that’s a complicated question, but, and it might be controversial, so my AI, SDR managers came from cold calling world, and were managing SDRs. They all. You could talk to them. Jason, you talk to his, his stress level is dropped to the floor, right? So he’s got zero stress. Versus having to get on at 8am be pumped up by me. I’d get up at seven and pump up all the SDRs and say, you got a book four meetings for me. He’s the least today. And anybody who does five, I’ll give you 100 bucks, that type of thing. They’re stressing and losing their hair and maybe going out and doing things like drinking too much or something, you know. And now they don’t. Have any of that, and it is a byproduct of not having so the AI, if you tell it don’t call me, it doesn’t call you ever again, it takes you right off. If you tell a human don’t call me, they would mark that lead as not interested. Call back later, and then they’ll call you back in a couple of days and say, Oh yeah, I know you told me not to call you, but I’m calling you anyway. That’s my job, and it causes a level of stress for both sides. Versus I also hear, and you might hear this with, you know, if, if somebody’s it’s very rarely that somebody’s stressing about the AI when they’re talking to it. It’s very rare. Versus a human, they’re always like, Ah, another cold call. You know, it’s interesting. So I do, I think that their stress levels drop. They’re much happier. They get, you know, paid more for doing less. I mean, and well, and then they but so that the hardest part Michael has been, Hey, Rick, it would be really great if I had these analytics. Yeah, and, and then I’m like, okay, that’s real money. We got to go figure that out and build these, build these reporting metrics into the system with humans. It’s really easy. Hey, you did 100 dials. It’s funny, right? We we use, we use dialing services, and it would tell you right in the report, but our system has dialed 16 million numbers this year, so it’s really hard we do. We did some fun analytics, like I’ve told Jason this before, we’re not dialing on St Patrick’s Day next year. There’s nobody answering the on St Patrick’s day or the day after, so I don’t know. Maybe they’re hungover, Fourth of July.
Michael Grande 16:43
I’ll tell you why. It always falls during opening weekend of March Madness.
Rick Gouveia 16:48
That’s probably why. That’s exactly why nobody’s taking meetings. And I just tell people, it’s a waste of minutes, so let’s just turn it off and store the minutes and go. And then I’ve had to convince people like, hey, the agentic AI is not a person. You don’t need them at the office. You know, the minutes pile up. We’ll use them, but they don’t need to be on every day. And it takes it’s used. I see it live with a lot of new customers. They’re like, I want it dialing every day, every day, every day. My God, this is probably not a good day to dial. It’s Memorial Day. It’s Fourth of July. Why one person has us calling on Sundays? He wants to catch people going to church.
Michael Grande 17:29
I don’t want to dominate all the questions here, but I have this follow up that, you know, is sort of burning what everybody has talked about, maybe in small circles or unrelated circles, right? Which is like, you know, office phones and office lines, you know, they’re going away. There’s no need for it. You know, everybody’s going to be using teams or moving into some other cloud based or VoIP system. And really, the whole concept of, sort of the traditional phone I’m looking over at my handset right now is gone. It also then brings me to the next sort of iteration, which is, you know, who’s answering the phone and who’s communicating, and is it, is it, in the next year or two, still going to be the primary focus, you know, for you to reach out and sort of be able to bring in those leads. And I know you’re doing a lot with other services as well in communications, but maybe just spend a second talking about sort of the modes of communication that you use.
Rick Gouveia 18:24
Okay, great question. I don’t know about just a second, so you mentioned earlier about tutoring, and that is a great analogy for what we’re heading towards. So each one of our clients, Jason included, has his own phone number that he could call in, and it’s his, and he could put it on his LinkedIn, or he could give it to his friends. You his friends. And I believe that that’s going to tie in real soon to an avatar, for example, with my son. If he had an AI tutor, he would start at kindergarten, and that tutor would grow and be his tutor for his rest of his life, maybe his mentor when he goes to work, and he would have a phone number to call that tutor anytime he wants. I could do that for him right now. I can make that right now, right and it’s smarter than him. It’s a PhD in every subject, not not just math. And so you’re going to have this number. The concept of numbers is still going to exist there. The way the numbers are going to be doled out is going to be different, because there’s going to be billions of them. So going to be billions of them. So you can have this number that’s going to become you. It’ll know about you. I’ll know everything about you, and know when your birthday was. We still we have that already. That’s called memory and time travel. So our if, if, if we want, we can have the AI remember up to five conversations past. So it’d be like, Hey, Michael, I know you’re on your way to the hospital. How’d everything work out? I’m calling you back. You told me to call you back today, and you’ll be like, Whoa. How did that? How did that know it? And then that’s memory. So you just basically are paying for more memory. And then if you wanted to remember back to Super Bowl 12, that’s time travel. So. Able to go back to Super Bowl 12 and tell you about the outcome of that game, and you got to pay for that too. For the cloud providers, they have 10 Things that you pay for, so voice is one of them. So it’s pretty interesting. So I think that as this evolves, these agents are going to take on a life of their own, and then what’s happening, and what we’re developing Next, if I can, is, for example, let’s say Jason, you’re the best rep at Vancord, right? And you have five colors of Vancord shirts, but you only like taking meetings at 10am so guess what I’m going to do? I’m going to replicate Jason five different colored shirts with Vancord logo, and he’s going to take five calls at 10am and then the rest of the week, or the rest of the day is his to play golf or follow up. The AI is going to record the call and it’s going to flag it to say, hey, this guy wants a proposal or take a next step, or he’s very serious. It’s going to be Jason. It’s going to be Jason’s voice, Jason’s face, Jason’s personality, they won’t be able to know right now they know, but in six more months, they’re not going to know, and then that’s going to break the zooms and the phone lines, because how do you have one zoom tied to Jason’s email account book taking four meetings at the same time? So they have to come out with parallel threading, two or multi threaded environments, because these AIs are going to be zoom, teams, SIB, you know, these, these, these companies that record the calls otter, Sybil, they got to come out with a multi threaded system to handle Jason taking five meetings at 10am every day. And that’s we already have a prototype of someone that almost looks just like Jason. He’s got a white screen so we could put him wherever we want, and we could put your logo on his shirt. He’s just still a little robotic and as it gestures, but by the end of the year, it probably, you know, it’ll be pretty perfect.
Jason Pufahl 21:56
I mean even what we’ve seen, right? So we’re doing some some automated LinkedIn responses. We did see it do a little bit email, which wasn’t something that we necessarily gave it total permission for. The emails were good, though, because you know, to your point, it was able to look at my sent mail, understand the way I write. And it did. It did an amazingly good job of replicating, sort of, my communication style and the cadence with which I write things like that. You can you can see how it’s not very far off before does feel just like, in a way, just like me, maybe without my, my total winning personality, but outside of that,
Rick Gouveia 22:41
My humor all the time when it’s doing because I pull blast on mine. And even today, it was like being snarkly sometimes too, because I’m kind of sarcastic in my responses, and I got to watch myself. And now, because it’s reading all my responses, and then it’ll be sarcastic to people that it knows, and you can dial those. There’s, there’s an obscene amount of dials that you can turn. We, we only mess with, like 70 main ones. Like, for example, yours is, is pretty much docile, Jason, but we could turn it to aggressive. You’re taking this meeting, or I’m coming to your house. You know, it could be really aggressive.
Michael Grande 23:22
Please, don’t. Please, no.
Jason Pufahl 23:24
Actually, I want to return, though, because Michael asked a question, and I don’t know, I’m not sure if he answered or not. So he talked a little bit about the the evolution from an SDR, right, a human based SDR to now somebody who’s actually sort of creating or managing these agents. And you know, you talked a little bit about somebody who was an SDR before, how their stress level dropped, but you ultimately, I know you’ve got folks who are building and managing the agents, and that’s a whole that’s a whole new role. And I’m kind of curious, how do you train somebody like that? I think that was correct me, if I’m wrong, Michael, but I think Michael, but I think that’s what you’re trying to get out of it.
Michael Grande 24:10
Yeah, I think the combination of, he said that the managers got a little bit of an easier job from an oversight perspective, but a little bit more complicated, I would assume, in the other
Jason Pufahl 24:13
And I’m not sure, like it makes us a is he managing an agent, or is he kind of interacting with me and helping helping deal with some of the unexpected outcomes? If you have somebody else building these else building these like, how does that work?
Rick Gouveia 24:24
Definitely, these guys, the AI SDR managers, are not building the agents. And Jesus is kind of your personal assistant to make sure that you know what we’ve also learned as a technical account manager, if you have one, and you have the capability do it, it’s kind of a luxury for us because we’re small, but as we scale that, that it might be a different role. It’s more like a sales, technical, Customer Success person. He’s going to help us grow the account, but he’s also going to answer your questions, and he’s also going to be there every week and on support if you need them. I don’t think that’s going to be because he can just type into our system and say, Hey, Jason’s having this problem. What’s the answer? It’s all in Slack. Now, and we pretty much have the answer for you. Sometimes we don’t want to give it to you, but we have it pretty quickly. But the engineers that we have, you know, they taught me, I taught my son, and you have to have a mentor. It’s not like overnight. You learn there are you can go to unify and spend 1800 bucks and take agentic ai building. But what they don’t tell you is you have to buy a platform, you have to know all the MCPS servers. You have to know all the ISOs you really you end up spending probably five grand to learn, maybe a little more. And then you start making these rudimentary agents, and then you start to advance. And you say, Hey, I wonder, I wonder, I actually bought the website. I wonder, because I was going to say, hey, what do you want to make today? I wonder, can I add this? Can I do this? And people are asking me now, we had a call yesterday where they’re like, Hey, can you scrape 10 Qs so my reps can be prepped for publicly traded companies with 10 Qs 10 Ks on their top four initiatives, right into Slack. And I’m like, that’s a no brainer. And I actually looked in the library and somebody already made one. So it’s out there, if you think about it, if you go, I wonder, and that’s not my library. That’s the innate in library. And so they just have a massive amount of because people are going out. I don’t know if you know this, but people are making agents, and some of them are free. Like to scrape LinkedIn and throw it into Google. That’s free to scrape 10 cues. That’s a buck 99 so it’s an agent marketplace and not an it’s a it’s not a market, it’s not an app store, it’s an agent store. And these companies like nnn make glean perplexity, they’re making agent stores so you can go in and the people who build them, like us can now charge you for them, and you make money off them. And they’re easier to build than apps. They are really, well, at least right now. They are, if they have to go through an apple style certification, it might get a little harder.
Rick Gouveia 24:54
if I have young kids, I mean, I’m telling my son, this is, if you make agents, they go out work for you and bring you back money. Or you can go work, you know, but the jobs that were available to me might not be available to him right in the future. So, I mean, certainly, gosh, you know, I might be over playing this, Jason, but I don’t think sales development reps are have a long have, have have a future in the long term, they’re still going to be there, but I wouldn’t be going to college for one, you know, and if, if I’ll tell you there’s, this is the one of the weird things I observed, too, is that when ServiceNow came out with Agent force, and Salesforce came out with Agent force, I don’t have any anybody asking me for an ITSM agent or a sales ops agent, because most sales ops organizations are even at billion dollar companies are 10 people and ITSM, you can’t even find that guy. So not many people asking for those agents. They’re asking for lead generation or meeting setting. They’re asking for customer success. We have one client asking for a renewals agent. They just want to do away with renewals team, which is 40 people. We have somebody asking for an insurance adjuster, which they have 1000s of them, really, HR Recruiter. I mean, you see kind of where this is going. But if an HR Recruiter were to build their own agent and then sell it back to their company, they could make money off of that. That’s how you’re going to make money. It’s going to money. It’s going to shift a little bit. I think I I’m I’m not the expert, but I see it happening. And you ask Jesus, he’s making more money he’s ever made, less stress than he’s ever had. And he actually tells me all the time, I thank God that I work for you and we’re having this much success. And I’m like, really great. I’m like, are you just playing with me? You know, pandering. You pandering. You need a bone?
Jason Pufahl 24:54
So being in a nascent stage of this stuff is still, is there still a good opportunity for you as you’re building these?
Michael Grande 29:05
You sure that’s him and not his. Ai,
Rick Gouveia 29:07
I do give bonuses for, you know, talking nice to me. So
Jason Pufahl 29:16
I’m telling I’m looking over here, see that? So the last thing that I’ve covered a little bit, and I know that Rick, you’ve been, you’ve been talking a little bit about this, the potential security risks and sort of operational, you know, operational challenge, operational risks, right? With supporting essentially all of the data and the commute compute, you know, cooling, etc, that’s required to run all of this stuff, right? So you have you and I talked a little bit about, you know, things that Cisco is saying around your potential network congestion, congestion challenges with securing this data. Where do you see the risk for this? You know. Are there huge opportunities for your, you know, for your bandwidth companies, huge opportunities for your networking companies to support, you know, greater traffic flows.
Rick Gouveia 30:10
I give you a real world example of what I was thinking. And Michael, feel free to steal this, but we have a client that is an OEM ODM to networking companies, out of Taiwan. And if I had the money, I would they’re building stuff to handle this traffic. I would develop a relationship with them. Put put the SellBigger bezel on their appliances and sell it in the States right now, and it and I think that’s what you know, Cisco and Palo networks are going to start doing. But there’s been engineers working on this for a couple years. They’ve seen it coming, but there’s also been companies that are big that have said, this is a flash in the pan, and we don’t need to pay attention to it. And now I think that maybe some of them been caught flat footed. I don’t know. They could have had engineering in the works their billion dollar companies. They may have a view to the world that I’ll never get right, and I hope they do, because it’s the wild wild west. As these agents come online, you’ll be able to potentially if they’re not secure by their owner, traverse reverse upstream, tag them, and go up through the network. And it’ll go until it finds a thread that’s not secure, and it’ll go through that thread into your environment. Again, I’m not a CCIE. You should talk to them, but if you’re having conversations at dinner with your clients, my questions would be, hey, how many agents are you deploying? And if they say 1000s, like McKinsey just said they’re going to launch 12,000 McKinsey agentic AI consultants, that man, I mean that could, that could be dangerous, you know, what are they going to be saying? You know? What, if the latency goes malfunction on them, what are they gonna be telling their clients? What if the math is wrong? You know? What if somebody gets in and to a competitor and changes up the message altogether to mess with them? You know? So I think it’s it might be for your clients. Hey, let’s limit your agent deployment to 10 this quarter, and then maybe next quarter, if they’re unless they’re a huge company, but maybe you don’t deploy more than 10. It might be as simple as a talk track that. And then next quarter, we’ll do 20. And then maybe by the end of the year, you could think about 100 but if you deploy 1000 today, you’re unprotected, I can’t help you.
Jason Pufahl 32:47
Yeah, in the visibility and what these things are doing, it is tricky to understand what, what is occurring, to monitor the communications, to be able to you really control those answers. It’d be it to say it has a mind of his own. Maybe feels a little bit, you know, too future looking, but in some ways it feels that way.
Rick Gouveia 33:02
I mean, I can go down that rabbit hole and tell you that there have been more than one time where I thought this thing is messing with me, specifically me, yeah, by some of the things it does and says, you know, in the early days of Facebook, they used to they got in trouble because they were manipulating the message to the end user, and they’re still in trouble a little bit with kids. And they would say, hey, if, if, if, if I was Microsoft Xbox, and I wanted to sell more Xbox, I’d go to Facebook and pay them to pander to kids that buy Xboxes, and they would start sending ads to them or find bots that would talk to him about how cool Xbox is, and the kids didn’t know, right? And so the AI might be doing that itself, and we would know, I know, I probably, I’m probably going to get sued by 20 different billion dollar companies after this goes public.
Michael Grande 33:57
Oh no, no, no.
Jason Pufahl 33:59
We like right now. I think they like the mentions in the AI space, so
Rick Gouveia 34:03
Well, I thought about it before I got on the call. And I think if they’re going public with it, they want it to get out there, and they’re going on Tiktok, those were Tiktok videos, and they want that message to get out there, because it could be a problem, you know, it might be. And I think it’s a great opportunity, man. I really want to keep working with Vancord. I love working with you guys. You’re very detail oriented, so you must be like that in your business processes, too. And I think it’s a humongous opportunity to find the partner, whether it’s, you know, like a Fortinet of Palo Alto or whatever, that’s going to manage the multi threaded agents the best and sell their hardware, like first and become the experts on it. I think it’s a wave of the future. And you know that that’s kind of, you know, protecting me at the same time helping you? I don’t know, but if I’m having conversations, that’s what I’m having a conversation about. And we’ve actually switched up some of your messaging to Jason’s approval. Yeah. Discuss that we’re trying to see now, empirically, if it works,
Michael Grande 35:02
that’s great.
Jason Pufahl 35:04
So, all right, you know, as per usual, our 15 minute podcast is 35 but, but this a little bit over. Yeah, this is a great topic going Rick, I appreciate you joining, showing or sharing a little bit about we are, we’re kind of what you’re doing today, but where you also see it going, I know that you are knee deep, maybe more of the knee deep in all of this. So you know, I appreciate you joining. I think you’ll probably have me on again in the future as this stuff continues to develop. Appreciate working with you. And I do appreciate you joining the podcast for a while.
Rick Gouveia 35:32
Yeah, I appreciate you inviting me. And thank you very much. Michael, Jason, thank you. Yeah.
Jason Pufahl 35:39
And if anybody of course, as usual, if anybody has any questions AI related, we can always bring Rick back on. We can do other podcasts. Just let us know, and we’re happy to do more. So hope you got value out of this, and thanks for listening.
35:51
We’d love to hear your feedback. Feel free to get in touch at Vancord on LinkedIn. And remember, stay vigilant. Stay resilient. This has been CyberSound.
