AI Demystified with Jason Lowe
Our November 28th Telarus Tuesday Call welcomed Telarus AI & CX Solution Architect Jason Lowe, who dove deep into the world of AI. He shared how to get more comfortable with AI in order to sell more AI-centered products to your customers and how and where to look for opportunities. To learn more and access the recording, click here to log into Telarus University today!
Jason Lowe: One of the things that I would really like to go over today is some of the content specifically surrounding the AI lightning trainings we’ve had. So for those that haven’t known about them or been able to attend them, we had 7 or 8 different lightning trainings around the nation over the last few months. What you’re going to get is a decent portion of that presentation that I’m using to train partners on AI, my whole goal there is to get people to feel a little bit more comfortable with AI.
Jason Lowe: I’m gonna cover some of that. And then we’ll transition into areas where you can actually sell AI centered products or work with your partners to obtain AI centered products. So let’s get right to it. Because we do have quite a number of things to get over. So we all know that AI is really in the news quite a bit. There’s a lot of headlines happening everywhere. You know
Jason Lowe: questions about privacy and facial recognition, what the heck is chatGPT. And how does it work? What’s gonna happen with mo robots, and you know, is there moral and ethical issues surrounding AI. How big of a threat is it? Is it going to turn into Skynet, or agent, smith or ultron, or I don’t know scylons or something like that. You know, the government is definitely taking
Jason Lowe: measures to try and safeguard against those things, including the Department of Defense, completely scrapping their AI adoption strategy as recently as the beginning of this month and and incorporating a new strategy. So let’s take a look at some of the most current events in AI. These aren’t all of the most current events. There have actually been things that have happened in the last week that are pretty significant. But just so, you know.
Jason Lowe: these are some of the more prominent ones. So President Biden did for those that hadn’t heard the word issue an executive order on on AI, and how to use it. Just so, you know, it’s toothless.
Jason Lowe: But that doesn’t mean bad things, because we’re trying to put the framework in place to make sure that AI is being used ethically, responsibly and safely, which is very nice, that the government is keeping an eye on it. Is this because it’s needed at this particular moment? Or is it because people are scared of AI. I would say, probably
Jason Lowe: a little bit of both. AI is becoming very, very prominent. There are a number of generative AI engines that have become really big deals. Chat. Gpt. Was released a few years ago, at least in its very early forms, and earlier this year Chat was released, and then later this year, chatGPT. 4 was released and
Jason Lowe: 4.0 had some pretty massive improvements.
Jason Lowe: Just to give you an idea of scale chatGPT. 3 had about 87 billion parameters or factors that were available to it for training. chatGPT4 went up to 1.8 trillion parameters or factors that were available to it for training. And let’s put it in context of the human brain.
Jason Lowe: Human brain has approximately a thousand trillion parameters. So quite a bit. So we’re approaching
Jason Lowe: the capabilities of the human brain. But we’re not quite there yet, but there’s a lot of competition which is good, because, as we all know.
Jason Lowe: competition breeds advancement and technology innovation, which is happening very, very rapidly, in fact, so much so that the Cold War, the cold nuclear war of AI, if you will has resulted in chatGPT, really upping the ante. Recently, they released that they can hear, speak and browse the open Internet using chat. And also they released the chat. GPT. 5 is coming.
I’m sure some of you heard of all the drama
Jason Lowe: of the CEO of of Openai being let go, and then hired as a consultant by Microsoft, and then suddenly rehired
Jason Lowe: chatGPT is still the preeminent engine that is being used the most out there in the market today, and there are other advancements that are going to be made known in chatGPT. 5, including a cool one. If you want to look it up on the Internet called Q. Star.
Jason Lowe: which is really, really neat and a big advancement. So other things that are happening, Tesla is going to incorporate AI in their cars, so much so that their intention is that you can drive it and do everything by voice. Close your garage, change the thermostat in your home. Send a message to people, send an email everything that you would want to do hands free. Possibly you know the Grok AI chat. But Chatbot would be able to help out with that. But this is not just
Jason Lowe: voice or generative. AI, where AI is making advancements. In fact, the medical field is one of the areas where it’s making huge massive strides. This is a big deal. Wouldn’t it be cool if
Jason Lowe: AI could detect and diagnose, or at least lead you to the to the conclusion that you may very well have type 2 diabetes. This can be done just by analyzing a voice recording. Very soon you will see an app
Jason Lowe: on your smartphone that says, Do I have type 2 diabetes you’ll speak into it. It’ll send it to an AI engine which will analyze your voice print and let you know with 89% accuracy, whether or not you have type, 2 diabetes. Now, given the fact that 50% of cases of type 2 diabetes in the United States go undiagnosed.
Jason Lowe: That’s a pretty significant advancement.
Jason Lowe: Other advancements include ways in which a I is interpreting things from the brain. I’m going to show you a couple of those in a few minutes, but one of the big ones is something that Stephen Hawking would have really appreciated
Jason Lowe: putting something in the brain that can interpret your brain signals and tell you what it is that you want to say so. Wouldn’t it be neat if hawking could have put that on and spoken in a more realistic human voice instead of that robot voice? I don’t know if he would have done it, because my understanding is, he really liked it, and considered it part of his personality, but for those that have speech impediments, but can still think words that they want to speak.
Jason Lowe: Look at this, AI is going to lead the way, and that’s going to become something in the next few years that’ll be dealt with very effectively. All of these advancements are leading to
Jason Lowe: a lot of speculation as to how soon are we gonna hit the sky net level, or how soon are we gonna hit the level where we’re close to human capability? Well, just sort of differentiate for you. There are basically 3 levels of artificial intelligence that people are keeping their eye on. One is called artificial narrow intelligence. One is called artificial general intelligence, and one is called artificial super intelligence, artificial, narrow intelligence is everything that we have today.
Jason Lowe: We are on the cusp of of getting close to artificial general intelligence. Now just to differentiate artificial, narrow intelligence is not self aware. It cannot do all of the things that a human being can do. That’s when artificial general intelligence comes into play. In fact, artificial general intelligence is when artificial intelligence can be truly creative
Jason Lowe: and do things on its own, like a human being? Does artificial super. Intelligence is taking that to the next level where the artificial intelligence can be so much more effective than so many people all at the same time.
Jason Lowe: Yes, Jacob, we’re gonna go over what AI products are available in the Channel today during this webinar. So we’ll get there. And Google Deepmind Co founder Shane Leg. He is a preeminent authority, as it relates to AI. He predicts 50% chance of Agi artificial general intelligence by 2028. If you want to ask me. 2 years ago I would have said, no way. It’s not gonna happen for the next 15 or 20 years.
Jason Lowe: Now I’m not so sure I’m kind of wondering if that’s not going to be a little bit of a conservative estimate. Why, well, let’s go a little bit more over some of those advancements. Remember when I said, we talk about brainwave interpretation a little bit later on. Let’s do that now. What I’m going to show you, and this is going to blow. Your mind
Jason Lowe: is on the left. Pictures that were shown to individuals, and on the right will be images that AI predicted the person was viewing.
Jason Lowe: This was done using a functional MRI or functional magnetic resonance, resonance, imaging scan which was taken at the time that the individual was viewing an image. Then, later on, AI was able to interpret the MRI read and say, we think that the person was viewing this image.
Jason Lowe: This is historical, not real time, because the MRI cannot be conducted real time. But this, I think, is going to blow your mind
Jason Lowe: on the left. What was being shown
Jason Lowe: on the right. What a I is predicting that they’re viewing
Jason Lowe: anybody getting a little scared or spooked out. Now. brainwave interpretation
Jason Lowe: pretty big deal.
Jason Lowe: So now people are wondering
Jason Lowe: how close are we to real time on this? Well, our good friends at Meta, who, by the way, on Facebook and a few other prominent companies out in the world. Meta has come up with a method called
Jason Lowe: Let’s call it magneto encephalography. They have these Meg or Meg scanners that are put on the brain. It’s a non invasive method of decoding visual representations. This is a big step in working towards real time, decoding of brain activity. So what you’re going to see
Jason Lowe: is on the left again, an image shown to someone. They’re viewing it for 1 s, and on the right
Jason Lowe: is what AI is using to try and narrow down what that image actually is. It’s going through a bunch of training that it has received and trying to come down to an actual, closest match for the image that somebody viewed for 1 s. Now a big point to show on the right, you’re gonna see it scrolling through images very rapidly, but on the right it’s also being shown at one quarter speed. So the AI is actually executing this 4 times faster than you’re viewing.
Jason Lowe: Let’s watch that.
Jason Lowe: Look at that. It goes from something that didn’t look like a panda to something that had color shading like a panda. We go through something that’s not even a human being to something that had a suit
Jason Lowe: on the right. You’ve got a machine that looks pretty similar.
Jason Lowe: Here we go. Let’s add spots. Let’s get to something that’s flying, that’s small.
Jason Lowe: We went from a dog to a horse or a donkey very quickly. Pictures of tomatoes.
Jason Lowe: cheese, and bread
Jason Lowe: pretty significant. Now that is close to real time.
Jason Lowe: Alright. So one more thing we need to talk about here.
Jason Lowe: What is AI
Jason Lowe: AI? According to AI. Yes. I asked, chatGPT. What artificial intelligence was, and it gave me big, this big, long explanation. So
Jason Lowe: by this definition, we’re going to talk about a few things. AI is something that can acquire knowledge can perceive the environment and make decisions based on what it’s acquired and comparing it to what it’s received. In other words, it’s able to reason and learn and maybe understand different things and make decisions or judgment calls based on comparing it to training points that it’s already had, according to this strict definition. Just so, you know, artificial intelligence encompasses a lot more than we think.
Jason Lowe: I VR on a phone system. or you’re pressing one for sales, 2 for customer service.
Jason Lowe: Yeah, that’s artificial intelligence. Why? Because it’s observing the environment. It’s checking out what it is that you entered on the phone. And it’s taking that information and making a decision. What to do with the call based on what it observed in the environment is an I. VR, a. I. Yes, it is. Is a spam filter, a I.
Jason Lowe: Yes, it is.
Jason Lowe: If sender equals this, then send the email to our spam folder, else wise give it to the person. This is what AI is. interpreting the environment against what it’s learned and making a decision based on that other aspects of AI that people are fascinated with include supervised or unsupervised learning. Let’s put this in context so unsupervised learning.
Jason Lowe: That’s when you were a little kid, you walked into the kitchen, and you put your hand on the hot pot and burned your hand. Ouch! You learned the lesson on your own that’s unsupervised learning. Now, if our mother had done us a solid and told us before we actually did that don’t touch the pot. It’s really hot. It’ll burn your hand
Jason Lowe: that’s supervised learning. If you took it seriously and don’t touch the pot.
Jason Lowe: if you’re like me and you were a kid, and you touched the pot, anyway. Well, it’s a combination of the 2. But really supervised learning is the method that is very preeminent in AI. Today the vast majority of learning by AI is done in a supervised fashion, where we’re feeding information and telling it what it is, so that it can then compare it to other things and determine if it’s similar enough to match them up and say, alright, we think this is what this is based on its similarities to what we learned before
Jason Lowe: machine learning is another aspect of artificial intelligence and deep learning is a deeper, unintended level of artificial intelligence as it relates to learning and training. What’s the difference? Well, machine learning.
Jason Lowe: Okay, don’t ask a data scientist if this is correct, I’m just using this to kind of try and show you the difference between the 2. But machine learning tends to be a little bit more boolean for those that don’t know what Boolean is. We’re talking ones and zeros, black and white. It’s very
Jason Lowe: yes, and no. And so that type of learning is machine learning. It’s a little less broad as it comes to nuances and shades of gray, where deep learning differentiates itself, is being able to take into consideration the probability of shades or gray or nuances. So it’s able to be a little bit more intricate, a little bit more refined in the comparisons that it draws. So to put this in perspective, deep learning actually uses something called a neural network.
Jason Lowe: At least, that’s what’s being used today to learn things. Now, deep learning is made up of. Input and then a whole bunch of layers of consideration. So these different nodes in artificial intelligence, which are the green hidden layers there.
Jason Lowe: AI, in deep learning and machine learning, but deep learning, because we’re looking at a neural net right now is able to interpret things and add decision making points. In the case of machine learning as something a little more Boolean, in the case of deep learning, it could add some of these green dots in there.
Jason Lowe: Now think of the lines coming from the different green dots. Those are maybe probabilities that it could be this or could be that. And then it reaches another layer of something that could make the decision of interpretation, and maybe lend it to go somewhere else, based on what it interpreted on that particular note. So this is where we’re taking into consideration those shades of gray, those nuances that might help it further, refine and better identify things that lead to a specific output or interpretation, or even a decision that is made based on what was observed.
Jason Lowe: Okay, so in summary.
Jason Lowe: artificial intelligence is all of the things where it’s observing the environment and making decisions and helping humans and reasoning based on what it’s learned to do things inside of that you have machine learning, which is the whole set of algorithms where it’s able to learn without actually being explicitly programmed. That’s what I mean by being able to add decision points to those things that’s machine learning. But then, when we get to something much more deep.
Jason Lowe: like a neural network driven ability to reach nuance and shades of grey. That’s deep learning. So now you know the difference between AI machine learning and deep learning.
Jason Lowe: This is a big point. If you get this pat yourself on the back, because you now know more about AI than most people do.
Now another big
Jason Lowe: topic that people like to talk about with AI is generative. AI. Earlier, we talked about Llm. And different large language model engines. So how does that work well, generative AI is really an interesting thing. If you have a conversation with chatGPT or Pi AI, or Bard, or any of these other different things that are easily and readily available to you to play with.
Jason Lowe: It’s kind of funny because it feels like you’re talking to an actual person. If you’re doing these text large language models. In a sense you’re not. What’s happening is that the Llm engine is trying to mathematically predict what it should say next.
Jason Lowe: it’s not actually saying what it thinks. It’s using an algorithm to try and predict what it should say next. And it’s doing this based on what it was trained with.
Jason Lowe: and then it’s generating content by sampling from what it learned previously.
Jason Lowe: So if I at some point fed it a recipe for cookies, and then someone else later on asked it for a cookie recipe.
Jason Lowe: It’s going to feed it what it got earlier. It’s going to take what it was fed earlier, and it’s going to give that to who asked it and said, Sure, here you go. Here’s a recipe for cookies.
Jason Lowe: This is why generative AI is so helpful because it gets trained on
Jason Lowe: billions of parameters or billions of factors that are used to have it give good effective answers, or to conduct conversations. But again, it’s just trying to mathematically predict what it should say next.
Jason Lowe: An important area of this is what’s called the feedback loop. The feedback loop is a principle whereby we say, and that’s right, or Nope, that’s completely wrong. You may notice if you’re talking to chatGPT. There’s a place there for you to give it a thumbs up or thumbs down to indicate whether or not the answer was helpful or accurate, or whether it was not.
Jason Lowe: That is part of the feedback loop. In other words, the human being, right now, when we’re dealing with artificial neural intelligence
Jason Lowe: is really important in helping it learn if it’s coming closer
Jason Lowe: now to demonstrate the feedback loop and the power and the effectiveness of the feedback loop.
Jason Lowe: it’s pretty obvious when you look at what Llm engines are giving you today versus what they did previously. But I think visual graphics are really effective. And so what I’m going to show you is a video that demonstrates the power of the feedback loop. On the left
Jason Lowe: is a video created by an AI based on the prompt that was listed down below
Jason Lowe: in May of this year.
Jason Lowe: using the same generative engine that has gone through 6 months of feedback loops and feedback responses, it’s able to further refine what it creates based on the prompt. And so we took the same prompt and fed it to the same. A. I engine
Jason Lowe: after 6 months of feedback loop. And the video it creates now
Jason Lowe: is based on November 2,003, 23, which is on the right. So 6 months worth of feedback loop. Now you’re going to see the difference. Take a look
Jason Lowe: at how much better the videos are on the right versus the left. It’s going to blow your mind. I’m just going to be quiet while you’re watching this.
Jason Lowe: Anybody freaked out yet.
Jason Lowe: Same prompt for both videos.
Jason Lowe: Pretty good stuff. So that’s the power of the feedback loop, as you could tell, the videos on the right were so much better than the ones on the left.
Jason Lowe: Alright. So another big topic in AI specifically generated AI today is hallucination. It’s being mediated and moderated a little bit. But I still think it’s important enough that we should talk about it. So again, do not take this as a literal scientific graph. This is something I put together just to teach you the principle of hallucination, and why, it happens.
Jason Lowe: if you were to take this argument and present it to a data scientist, they would laugh if you would take this and give it to an AI programmer they would go. I’m not so sure that’s right, but this is a good way to teach the principal.
Jason Lowe: So on the left you have insufficient parameters or insufficient factors. That’s where generative AI is not quite effective. It’s kind of the equivalent of walking up to a 3 year old on the street and saying.
Jason Lowe: Can you teach me the principles of quantum physics? Three-year-old doesn’t know that stuff. There’s not enough stuff in its brain to be able to tell you and answer that question.
Jason Lowe: That’s what happens when you don’t have enough factors to be able to be effective as a generative AI engine that’s on the left.
Jason Lowe: on the far right
Jason Lowe: is when hallucination happens. Maybe it has too many factors. Why is that a problem?
Jason Lowe: Well, maybe it doesn’t know when it’s mathematically trying to predict what it should do next. It’s not totally sure, because it hasn’t been trained by the feedback loop yet. It’s not totally sure which one is actually correct or not.
Jason Lowe: So it goes through this mathematical prediction. capability that it has, and it picks what it thinks is probably the best way to do it.
Jason Lowe: That’s why it gives you a wrong answer. What is hallucination? It’s a generative AI engine giving you a wrong answer. That’s obvious to you, but maybe not to it, because it doesn’t have the ability to really look at what it says and say, does that make sense?
Jason Lowe: You know what I think that makes sense. I’m gonna leave it. No, wait. That doesn’t make sense. I’m gonna try and go with a different answer. It doesn’t have that ability yet that’ll happen at the Ag. I.
Jason Lowe: Or artificial general intelligence level.
Jason Lowe: So there are things being done to moderate that a number of AI providers are actually trying to limit the training set to something that is permanent for the artificial, narrow intelligence paradigm that it’s supposed to use. I can think of a number of different providers in our platform that are doing that today. For example, let’s talk about narrowing things down to specific business verticals or industries
Jason Lowe: if we make it so that that’s the only stuff that we’re training it on. I can ask something that’s a banking, AI, something about finance, and it has a better chance of giving me a correct answer if only it been trained on finance type stuff. And instead of something that’s been trained on everything in the world that way. If I’m saying you know what’s the current interest rate, it’s not gonna say green cats in Greenland. It’s not gonna do that. That’s obvious hallucination.
Jason Lowe: But maybe it’ll give me the current interest rate, which is a little bit more accurate.
Jason Lowe: Now, numbers, by the way, are something that AI has a hard time with. Why? Because it’s just trying to predict what it should say next.
Jason Lowe: There are things being done to try and make that a little bit more accurate as well. Recently there was the Openai Developers Conference, which was, you know, a week or 2 before the CEO was fired, but anyway, one of the big advancements that happened, and that was announced and generally available or used, or at least made publicly available. Or, you know, public knowledge. It’s actually been around for a while. You have a provider in our portfolio that this is the backbone
Jason Lowe: of what it actually uses to make sure that its AI is very accurate and the way that it generates things. But there’s this method called retrieval augmented generation. And what this means is, let’s ask it for something.
Jason Lowe: and then. normally, what it would do is it would try and mathematically predict what it should say back.
Jason Lowe: But this adds an intermediary step. It takes that prompt that you give it, or those words, or that question, and it looks at it from an intent perspective. What is it they’re really after.
Jason Lowe: and then it summarizes what it’s really after. And then it says, All right. Do we have a database or a data source that might actually have accurate information to more adequately or more accurately answer that question.
Jason Lowe: and so then it reaches out to a database and tries to get what it needs to more accurately give a response, and it gets that, and so it gets what it retrieved.
Jason Lowe: plus the original prompt, and then feeds that to an Llm. Engine to generate a response. So that’s why the generation that’s given to that point is augmented by the data it retrieved in that earlier step.
Jason Lowe: This is a big deal because it suddenly made accuracy much higher. Think reduction of hallucination.
Jason Lowe: This is a big deal. There are other things happening that are making this hallucination problem less prevalent. But this is one of those things that’s a little bit more public today. That’s kind of a big deal. Now for those of you that are actually playing around with AI engines right now and using it to generate, copy or to generate. You know an email, or you know
Jason Lowe: everything. There are things that you can do to minimize the possibility of hallucination. Some of these are things like just making sure you’re sure of the answer. Don’t give me an iffy answer
Jason Lowe: telling an AI. That means it won’t give you an iffy answer, it’ll only give you what it’s certain is the correct answer instead of a probability
Jason Lowe: if you’re giving it. Yes, no questions, it’s more likely to give you a correct answer if it says Yes, no, or I don’t know.
Jason Lowe: You can also make sure to have it reference sources which is a good thing to do. Think again, rag or retrieval, augmented generation.
Jason Lowe: If it’s referencing something. That’s a good idea, and then you can ask it. You know what? If you’re totally not sure of what I’m asking and what it is that I’m after. You can ask me. Follow up questions, and then it may, or it may not, depending upon how well it interprets what you asked it. But all of these different methods really reduce that possibility of hallucination. Alright, we’re 31 min in. I promised you we would get to ways that you can actually sell AI and providers that can actually have AI products that you can talk to your customers about.
Jason Lowe: Where is AI being used today in a lot of different areas just about everywhere across the business landscape.
Jason Lowe: You know, one of the more exciting AI stories that I know about right now actually isn’t in one of the categories here. It’s an agriculture of all place. Did you guys know that there is a machine that can scroll down lines of crops?
Jason Lowe: And on this machine are cameras looking at the plants that are in the crops.
Jason Lowe: And there’s also lasers.
Jason Lowe: And if the camera sees a plant that shouldn’t be there.
Jason Lowe: Think, Weed. The laser kills it. That is real and that is happening today.
Jason Lowe: AI is in. It’s freaking, herbicideless
Jason Lowe: destruction of weeds and crops. How big is that
Jason Lowe: their AI is helping with teaching languages. Their AI is helping doctors correctly interpret X rays in medical imaging. I mean, it’s just. It’s huge. There are strides being made everywhere. AI is generating a lot of value in the economy. It’s estimated 13 trillion of additional value in the economy today.
Jason Lowe: Look at retail. That’s a big area. Why? Well, because the area where a I is most effectively commercialized, packaged, and able to quantify
Jason Lowe: Roi or Tco is Cx.
Jason Lowe: Another area of expertise that I have right in CCaaS or UCaaS. Cx. That’s where AI is being used the most. And that’s where you’re getting a lot of bang for your buck retail is using it very effectively. But so is travel and other logistical areas. Look agriculture with those machines that’s up at the top. But that’s still a heck of a lot of value
Jason Lowe: being created by AI today. So how are business owners using it?
Jason Lowe: Okay. customer service, as we suspected with Cx, that’s the area where it’s being used. The most 56% of companies are using AI and Cx. But what does this mean?
Jason Lowe: 44% of companies are not. That means you as a technology adviser, there’s a good chance that a number of your customers are not using AI today in their Cx environment.
Jason Lowe: Go talk to him about it. That’s huge cybersecurity. Another prominent area where we work here at Telarus. Look at that.
Jason Lowe: 49% of companies aren’t using AI for cybersecurity or fraud management.
Jason Lowe: big big areas of opportunity today.
Jason Lowe: So how big is this market? Oh, kind of a big deal estimated at a hundred 1 billion now estimated to hit another
Jason Lowe: 500, you know, 500 billion by 10 years from now sixfold increase. It’s kind of significant globally. you know, how does close to a 9 times increase in the global, artificial, intelligent market
Jason Lowe: in the next 8 years. Sound lots of opportunity there. Software is the biggest portion of AI, and that’s where you can do the most damage in helping your customers figure out software tools and software solutions to utilize AI to make business processes better, more efficient, and to actually save money.
Jason Lowe: 65% of businesses are using AI in some form.
Jason Lowe: They may not even know it, but they are.
Jason Lowe: And there’s a lot of businesses that are not
Jason Lowe: cell. Microsoft did a really great study. They called it the 2023 Work Trend Index report which had some really fascinating data points about AI and the utilization of AI in business today, lots of companies are already using it.
Jason Lowe: In fact, a I. Deployments aren’t taking very long, most of them. The vast majority of them are taking 12 months or less, and how long does it take to reach a positive. you know. Return 14 months.
Jason Lowe: That’s pretty good.
Jason Lowe: really good.
Jason Lowe: And but this is the big data point down at the bottom that I want to make sure you notice is that there’s a lack of skilled workers out there.
Jason Lowe: A lot of companies are reporting that they just don’t have enough skilled workers to be able to implement and scale AI. What does this mean? They might reach out to their technology advisors for assistance
Jason Lowe: kind of a big deal.
Jason Lowe: There’s a lot of opportunity out there is the point I’m making. But here’s a data point that I think every single one of you should be using when sitting in front of your customers today talking about the utilization of AI and getting them excited about figuring out ways to incorporate AI in their business
Jason Lowe: for every single dollar that a company invests in AI. And again, keep in mind this is Microsoft. This is not some.
Jason Lowe: you know, Random Survey company with an 8% sample rate trying to figure things out. This is a fully vetted, fully capable, index report done on a lot of surveying of businesses today. For every dollar a company invests in AI, they’re getting a return of $3 and 50 cents. That’s 8,
Jason Lowe: 350%
Jason Lowe: return on investment. That’s not insignificant, y’all. And who’s going to help them do that? Well, a lot of them want their internal employees to do it.
Jason Lowe: But there’s a huge portion of them that are also going to be looking to technology partners and technology consultants or advisors. Other words, you
Jason Lowe: to help them venture into this great new world of AI. I’m saying a lot. The bottom line is is that we’ve hit another industrial revolution. I personally feel that AI is the Fifth Industrial Revolution, and you know what happened when steam engines was introduced. Right?
Jason Lowe: You had railroad barons that suddenly made a whole heck of a lot of money because of early adapt adaptation of that technology and utilizing it the right way. Same thing with electricity.
Jason Lowe: General Electric didn’t become a big thing because it sat on its hands. No
Jason Lowe: electricity changed a lot of things. So did computing. Apple and Microsoft are still the dominant people out there today because they were the first ones toeing the line making Pcs. When people were recognizing that Pcs were going to be a big deal in homes today, and then, of course.
Jason Lowe: connectedness being another industrial revolution. AI is, in my opinion, definitely that now, hopefully, you’ve learned enough of AI during this talk that I’ve been giving to make you feel more comfortable. Why? Well, because people are scared of it, and they’re gonna be relying on you
Jason Lowe: to help them feel less scared. Just so you know
Jason Lowe: a number of time over the last few months. I’ve had technology. Advisers call me up and say, Jason, this is a true story.
Jason Lowe: I had a customer reach out to me.
Jason Lowe: It’s not the CEO. It’s some guy in the It department, or in the Cx. Department, or something where the CEO called them into the office and said, Okay, guess what you’re now the Ea. I guru in our company. I want you to go figure out where to install AI in the entire company. Just go do it. You have liberty to do what you want. Go, do it!
Jason Lowe: And the person starts freaking out and says, what do you mean? What should I do? And the CEO goes?
Jason Lowe: I don’t know. Figure it out. Go, do it find places for AI. And so what does this confused person do they turn around? And they call their technology advisor. And since we’re early enough in the game many times, the technology adviser is calling me and saying.
Jason Lowe: How do I help this customer? This is why your engineering resources at Telarus are so important, and why you should be leaning on them. If you’re not already.
Jason Lowe: we can help you with this. We can help you figure out what tools are available to solve specific problems for your customers. If you aren’t able to recognize those yet. use us. That’s why we are here. This is why Telarus has the greatest engineering group in the entire technology solutions, distributorship, space.
Jason Lowe: We are the best. and yes, we know it, but we are humble about it.
Jason Lowe: as you can tell. But we do have great resources available to you. There’s a lot of opportunity, and you can be that trusted advisor to help your customers alright. Let’s talk about specific areas. I know we’re 40 min in. And I want to get to where we’re doing some Q&A. So let’s talk about AI, and where it’s being used in cx
Jason Lowe: chat bots. That’s the biggest area. 73% of people are using AI and chat bots. Look at the bottom. Though, phone calls, only 36% of phone calls are utilizing AI to solve problems. In other words.
Jason Lowe: 64% of companies don’t have any sort of voice. AI, virtual assistant or chat Bot that is interacting with their customers. They’re just not doing it. Lots of opportunities out there, text messaging even only 49%.
Jason Lowe: Lots of opportunities there to have AI automated customer interactions. Now, we’re not talking complete replacement of agents. Humans are still necessary for some of the bigger cases. But wouldn’t it be nice to be able to deflect a lot of those and have AI solve the problem
Jason Lowe: pretty cool, and it costs a lot less money to hire AI than it does to hire more contact center agents to deal with increases of volume.
Jason Lowe: This is significant enough that we have really
Jason Lowe: taking a different paradigm and helping you sell different things right? There’s the whole platforms you cast cpaass ccaas platforms. You can forklift entire systems, and we can help you help your customers
Jason Lowe: get new systems that are huge and that encompass the entire practice area of unified communications and contact center and other functional areas of communication that are often handled with Cpass platforms or contact platform as a service.
Jason Lowe: But we also have a number of providers in our space that allow you to add on or bolt on to existing solutions, all of the different AI things.
Jason Lowe: including Rpa, including analytics and employee engagement lots and lots and lots of opportunity out there, either for wholesale changes.
Jason Lowe: Or for these add-on Bolton products. That’s where we can help you out. Lots of providers that you can sell today. These are not all of them. These are just the ones that I could fit on this slide in a hurry.
Jason Lowe: But this is to demonstrate to you that there are so many
Jason Lowe: possible providers that have AI solutions that you can talk to your customers about. And again, if you want to have a conversation about AI, and how to identify those opportunities reach out to your engineering resources. That’s what we are there for.
Jason Lowe: But CX. Is the place where most of the damage can be done. 4 main areas in Cx artificial intelligence bots or agents to interact with people. AI driven quality assurance or analytics to make sure that agents are doing the right stuff.
Jason Lowe: AI driven knowledge management that can be presented to customers or to agents in a non live environment to give them resources to help them out. That includes training them on new areas.
Jason Lowe: And then you have real time. AI assistance for agents. Real time agent assist is what it’s called. We have a number of providers that are doing that today. And then there are other areas where AI is really being utilized. So
Jason Lowe: just trust us. I mean, there’s even one that is an AI product that will mitigate foreign
Jason Lowe: accents. I’m not kidding.
Jason Lowe: We’re talking a Filipino or an Indian agent. I’m talking East Indian agent, someone from India.
Jason Lowe: you know, it’s not a language problem, because they speak English, really. Well, it’s just an accent thing. There’s an AI provider that we have that will actually make them sound more Americanized for their American customers.
Jason Lowe: It will blow your mind. There are practical applications for AI everywhere in the Cx. Space.
Jason Lowe: All right. Let’s talk about it in cloud. There are lots of ways that AI is being used in cloud. Probably the most
Jason Lowe: known way is through scaling and effective scaling and cost reduction. Having AI help with automatically increasing or reducing resource utilization in your hyper scaling environments. That’s just one example of how AI is being used. Reporting and analysis is another. These are different providers today in our cloud space. And again, these aren’t all of them. But these are
Jason Lowe: many of them that are using AI solutions in the cloud space today. It’s worthwhile to have conversations with them. But then, again, you also have your entire engineering group that can help you identify these different AI solutions, and where to talk to your customers about them
Jason Lowe: in IoT is probably one of the more prominent areas where AI is being used. We’re talking computer vision. Did you guys know that there are computer vision systems? Now that can tell you when someone who’s not an employee has come in through the back door of the company
Jason Lowe: or into the storage unit or the warehouse area and set off alarms. It’s pretty big deal.
Jason Lowe: So computer vision is big for security, recognizing when someone puts something in their pocket, I mean in retail. This is where retail is being so big, not only with Cx, but also AI and computer vision. But it’s not just vision. There are olfactory sensors out there that can tell the difference between someone’s cologne and someone’s vaping.
Jason Lowe: you know. Wouldn’t we be able to tell if someone’s vaping in the bathroom at a school
Jason Lowe: that’d be kind of cool. Oh, hey? But we also have auditory sensors that can be in a school, and can listen to see if there are arguments happening and maybe things escalating to where a violent altercation could happen, and immediately send an emergency text to all of the teachers in that hall, and the principal saying, Get out there and stop the altercation now, because things are getting heated.
Jason Lowe: This is what AI can do right now today, lots of opportunity there
Jason Lowe: and then. Cybersecurity. As I mentioned, it’s not to the point where it can be creative cybersecurity. AI can’t say hmm. You know what? I don’t think anybody’s ever tried this. I think I’m gonna try this and just see if it works and go and try and find its own exploit. A. I can’t find its own exploits quite yet. There’s still the human in the loop
Jason Lowe: area. Now, where a human has to be creative enough to try and figure out a new way to come up with an exploit.
Jason Lowe: but once an exploit is found and determined, AI can be used to very rapidly and effectively test for exploits that can be done for good, or it can be done for bad. That’s why these cybersecurity companies are very important, because they have
Jason Lowe: tools that they can sell to your customers, or that they can work with you and identifying for your customers that use AI to actually stop those possibilities of weakness, or to fill in those exploits or defend against them effectively, or to very quickly shut everything down and mitigate any exploits that might have been done or penetration that has happened. These also are AI driven penetration tests that can be much more effective and identify weaknesses much more quickly.
Jason Lowe: So AI and cybersecurity is a big deal. It makes cybersecurity analysts and personnel
Jason Lowe: tons more effective, just like generative AI makes people that write copy tons more effective.
Jason Lowe: It’s a big deal. Now, Tellers wants to
Jason Lowe: B, your number one. Tsd, Tsd for AI. That’s our goal.
Jason Lowe: I think that’s evidenced by some of the things we’re doing. We’re adding high value providers. But we’ve also extended our provider portfolio by adding ascent business partners to add even more providers to our portfolio that are AI centric, and that are cutting edge
Jason Lowe: assent is more in the Cx space, which is again the more mature area. But you know they have a lot of really great providers that are doing really incredible things, including that accent mitigation, one that I told you about. You can source that through ascent business partners. They are an extension of our supplier portfolio. We also did all of those AI lightning trainings that I told you about.
Jason Lowe: And I am in the throes of a final recording hopefully tonight where I’m gonna record my fourth module. And then through the day today and tomorrow, actually recording modules 5 through 8
Jason Lowe: of an education track that we’re gonna put in the Telarus Lms. That’s gonna teach you more about where you can use AI with your customers. And then we’re also going to have various webinars and summits for next year planned where you can get together with us and learn more about AI Products and the different providers and their AI offerings.
Jason Lowe: Again, I’m sorry to beat this drum over and over. But, goodness gracious! You’ve got a deep bench behind you. Let’s help you.
Jason Lowe: Let us help you.
Jason Lowe: and for those of you that want to learn more about AI. I am very excited to announce that we do have our first upcoming AI summit. We’re targeting late February. We’re targeting it for late February. I can’t say that for sure.
Jason Lowe: but that’s going to be in a prominent easy to reach national location. It’s going to have keynotes. It’s going to have a panel discussion with providers in our space that have good AI products. It’s going to have breakout education sessions. And then at the end of it is going to be a trade show floor
Jason Lowe: big event.
Jason Lowe: We’re anticipating. This is going to be successful. We’re hoping a lot of people come, because if it is successful, we’re going to do it again and again and again
Jason Lowe: wherever it’s wanted. So keep your eye out for that. We’ll announce it as soon as we possibly can. AI Summit coming and based on its success. Who knows? Maybe we’ll add 3, 5, 8, 1015 of them
Jason Lowe: next year. But that’s what we’re doing.
Jason Lowe: Alright. I know that normally a Tuesday call is very conversational between someone like Lila and myself or Doug and myself. I know this has been more of a lecture and a presentation. But now I think it’d be really good to go ahead and answer any questions we have. So, Leela, tell me what people have been chatting about and what questions we’ve got that we can discuss.
Telarus Marketing: Yeah. So outside of just, you know the freakiness of AI, and how scary it can be. We did have some great questions come through. Someone pointed out
Telarus Marketing: how, when AI is used for customer service, it can tend to lose a bit of a personal touch, and some people are afraid about the worries of it taking away human jobs. What do you have to say in response to that.
Jason Lowe: I think you’re right. The personal touch is really important. This is why retrieval augmented generation is a big deal. But this is also why Llms are not being used in conversational AI
Jason Lowe: applications quite yet. There are some companies that are experimenting with that, and have started down that road. But by and large the more reliable technology is conversational. AI. The personal touch can be obtained, though, because the AI engines have been improved enough that they can actually reach out to data sources like crms or erps and different sources of truth to personalize those interactions. I don’t know about you, but I like to fly Delta. And one of the things I really enjoy about Delta is when I call their.
Jason Lowe: you know, phone number for Delta sky miles. It answers the phone. Welcome back, Jason. I see that you have a flight on Friday at this day. Is that what you’re calling about? Does it get more personal than that.
Jason Lowe: That’s pretty personalized. And then I can have a conversation with this AI conversational bot to maybe reschedule my flight or to change my flight time or cancel it, or, you know, add luggage. I don’t know. All of these different things can be done using AI. So advancements are there. It’s happening today, the conversational AI engines that we can sell to customers. You could sell the customers today. All of them have that ability to reach out to data sources and personalize that experience. Now, as far as job removal.
Jason Lowe: there’s this economic principle called creative destruction.
Jason Lowe: You know, it happened with, you know, agriculture machines suddenly taking the place of people out actually harvesting things in the field. It happened with manufacturing, you know, with robotic arms and manufacturing plants. Jobs don’t go away. They just change.
Jason Lowe: The employment is still going to be there.
Jason Lowe: AI is expected to create 60 million jobs by the end of the decade that are AI specific and AI-centric. So it’s going to be an economic stimulus for us.
Jason Lowe: There will be people that will need to transition away from what they’re doing now? Yes, but the ones that are going to be harmed are the ones that are going to be stuck in their ways and aren’t going to look for other ways to learn new skills and maybe do something with AI. So you know, AI,
Jason Lowe: are we going to reach the Utopian level of Star trek anytime soon. Probably not. We’re going to have an economy that’s going to change in fits and spurts based on these advancements and these inflection points that AI are going to add to the economy, but I feel positive about it. I think what we, as people need to be sure to do is
Jason Lowe: be really flexible and continuously be learning and refining and expanding our skill sets so that we can adapt to the positive application of AI in business today.
Telarus Marketing: Thanks for that great answer, and I saw Jim Tennet, who represents one of our wonderful suppliers, providing some some great insight in in the comments as well. Another kind of question that I’m sure a lot of our advisors have on their mind is, do you see a time when AI enables it exos to kind of go through the buying process on their own without the channel.
Jason Lowe: you know. That’s a fair question.
Jason Lowe: and I think
Jason Lowe: eventually. maybe way down the road.
Jason Lowe: That may be possible for those that actually want to do that. But I think you and I and everybody on this call probably know that the strength of the channel is in the relationships.
Jason Lowe: and AI can never take the place of a relationship.
Jason Lowe: And so I think, as far as purchasing an acquisition, and sales are concerned. The Channel is probably going to be the last place where AI is going to have a negative effect on the sales process, I think, in the Channel. AI tools are going to be a great place to help your customers, but I don’t think AI is going to negatively affect the sales process in the Channel for quite a while based on that paradigm. But that’s just my personal opinion.
Jason Lowe: AI is being used in sales processes and traditional sales, processes and outreaches today, a lot of B to C applications.
Jason Lowe: But people that are scared and people that aren’t familiar with different things. I don’t know about you, but when I’m on a customer service call and I have something really sticky, I want to talk to a human being. We’re not past that phase in human development. Yet where humans are completely comfortable, talking about their most intricate and scary problems with an AI, they still will have a huge subset of people that’ll just go.
Jason Lowe: you know. Tell me what your call is about. Representative. Give me a representative. The channel is going to be the same way.
Jason Lowe: Eventually there will be people that will want to do that. I don’t think it’s anytime soon.
Telarus Marketing: Yeah, no, I agree. Another question we had was someone who’s new to AI, and they want to start having conversations with their clients. And maybe they just want to start with something basic like helping them with like helping an Smb with their day to day processes. Where do you recommend that? They kind of get started?
Jason Lowe: I think, because Cx specifically, is an area that is the most mature, and that it’s most easy to demonstrate a positive roi and a lowered Tco.
Jason Lowe: I think Cx. Is probably the best way to start having some of those initial conversations. Another area is IoT. I mean, if people are in the manufacturing industry or the retail industry or areas where security might be a problem or they need sensors. You know, Internet of things is huge with AI today, those are probably the quickest and fastest conversations.
Jason Lowe: We are going to help you there. One of the things that we’re working on and what we want to get up as soon as possible. In fact, Leela and I were talking about this the other day is that we want to build a library of use cases. We are working with our AI Centric providers today, and we’re asking them to please provide us with good solid use cases that demonstrate what the problem statement was how it was solved
Jason Lowe: the product that was made the solution for that problem. What the economic impact was before and after what the return on investment was, etc., so that you have some compelling evidence to bring to your customers and say, have you ever thought about this? Look at this use case? This is amazing.
Jason Lowe: That would be a great way for you to train yourself up. And we’re working on that. We’re also going to increase our Lms modules to make it so that you can more easily identify and ask customers where you have opportunities to make their lives better with AI. So we’re working on that, too. So stay tuned. Our training activities are just gonna go up and up, not only to make you more familiar with AI, but also to make you more familiar with the AI use cases, the providers that we have in the AI solutions. They have to meet those use cases
Jason Lowe: and onwards and upwards.
Telarus Marketing: Yeah, I agree. And epic Ios, one of our early adopters at Telarus University, and they have a lot of great content already. If you’re interested, yeah, go ahead and check those out, and I’ll just ask one final question. You did mentioned your lightning, your AI lightning events? And someone asked, what’s the process? They would have to go through to, you know, schedule one or be a part of one. So what’s your response? There.
Jason Lowe: you know, that’s a great question. And it’s something we’re considering right now. The AI lightning events that we threw were very targeted. We were very, very fortunate to have assent business partners who have a bunch of good friends of mine that work there, and a fantastic group of of providers that work with them. They did it with us. It was a co-sponsored event with Telarus, and with these providers that work through ascent business partners, and so
Jason Lowe: In conjunction with the scent, we very quickly and rapidly threw together a set of cities that we wanted to throw these lightning events in, and we marketed them as quickly and effectively as we possibly could. But Telarus was interested in getting stuff out there as quick as possible. So we just targeted marketed in those particular cities. Are we going to do more lightning training events in other cities.
Jason Lowe: I think we’re working on that right now and trying to make the decision if you want them. And if you think they’re very important, please, for the love of all it is holy. Reach out to your Spd. And make sure you provide that. Input
Jason Lowe: and also, if you want to provide input on how else we can serve you and give you better education on AI, give us that. Input I, personally have been seeking it with just about every partner that I’ve been talking to. I have a number of partners that I’m friends with, that I’ve been having these conversations with, and they’ve given me great insights. But I can tell you that we are very, very, very, very AI focused. We want you to think, Telarus, when you think, AI.
Jason Lowe: And so we’re gonna do everything we can in the next year to make that established and to make it a no brainer. That includes lots of training events. We’re even gonna explore.
Jason Lowe: Leela’s going to kill me for saying this, but we’re even going to explore the possibility of an AI-centered Telarus certification because I know that certifications are really important to people.
Jason Lowe: So we’re we’re kinda explore all those avenues. Now, as far as who has sent business partners are. They are an aggregator. They are someone who, some of these newer, less channel. Familiar providers
Jason Lowe: are working out a relationship with ascent business partners who’s helping them get into the Channel and providing some of those channel specific capabilities and activities for them.
Jason Lowe: And so if you want to source regal I/O or Sanis, or replicant, or, you know, Apt Edge, or I’m I’m sorry I’m leaving a bunch out because they have so many good providers, but if you want to source one of these different providers. You register the opportunity through a sent business partners just like you would with any other provider. By the way, there’s no discount in commissions to you. It’s made whole. You get to deal with them just like you would with any other provider. It’s just the descent business partners is this kind of middle layer that’s helping them.
Jason Lowe: So that’s what ascent is doing for us.
Telarus Marketing: Thanks, Jason. Incredible presentation. Today. Every single time I hear you talk about AI. I learned something new, and I think that everyone agrees here. Really, great job. Really appreciate it. This recording will be into Larry University. Go back and and watch it over and over again if you’d like. I appreciate everyone for joining us. And thanks again. Jason really appreciate it.