The Future of AI, Revenue & Work: What GTM Leaders Need to Know NowOne discussion.

Artificial intelligence is transforming how businesses operate, but what comes next? This panel explores how AI is reshaping go-to-market strategies This was originally recorded at Go-To-Market Live last August, and it is so relevant today, even more now than before. This panel features leading voices in AI and operations, including Aditya Gautam, Tech Lead in Machine Learning at Meta, Jason Lowe, Professor at the University of Utah and host of the AI Diatribe, Kelly Chamberlain, CEO at nClouds, an AWS partner, John Bradshaw, President of AI Utah and Codebase, and Mason McMullen, VP of RevOps at Alicio. Together they are sharing their expertise on how AI is shaping business strategy, operations, and go-to-market execution. It is an incredible panel. You don't want to miss it. I was on the edge of my seat and so is everyone else. During the panel itself, I think you'll find it just as compelling.

All right, I'm really, really excited about this one. I'm excited to introduce many people who, 5 people who are absolute experts in AI. The first one is Aditya Gautam at Meta. Round of applause.

Woo!

Hello, thank you, Arisha. Next is Jason Lowe, professor at the University of Utah and founder of Aptivon. Jason!

Woo!

Thank you so much. The next one is Kelly Chamberlain, CEO at nClouds. Kelly!

Woo!

Next is John Bradshaw, President of AI Utah and CEO of Codebase. John! Woo! And last but not least is Mason McMullen, VP of RevOps and Strategy at Alicio. Mason! Awesome. Okay, I'm gonna pull this one chair over. And what I'm gonna do is, I have questions, like I have lots of questions that I've written. Hannah's written a bunch of them too. But what I feel like from what I've been hearing from you guys is that you guys really want to ask the questions to the experts. Am I right? Yes. Like, I'll ask— I'm just going to get one to get them started and then I'll just come around with the microphone. Okay. All right. Here's my question. Here's my one question. Would you please introduce yourselves and a little bit about why you're so passionate about AI? Go ahead, Aditya. You first.

Okay. Hi everyone. I'm Aditya. I'm currently working at Meta in machine learning. AI is one field which is like actually changing and like in a way right now, especially after that LLMs and the agent AI that has never seen before and the dramatic changes it is bringing into our life and also in the technology that is amazing, that is super I mean, amazing to see. So that's why I really love and enjoy being a part of this field.

Hi everybody, I'm Jason Lowe. I am a podcaster about AI, in addition to being a professor at the University of Utah. I am passionate about AI ever since I saw the robot Maximilian in the movie The Black Hole. Yes, there we go. That was scary enough for me to really get into that. So I've been You know, when ChatGPT decided to go public with its fun stuff, I jumped on the bandwagon and have been riding it ever since. So that's me.

My name is Kelly Chamberlain, the CEO of nCLOUDS. I don't know if it's passion so much as absolute necessity to understand what it is and how it impacts the business. So my passion is about survival and capitalism is why I love AI.

My name is John Bradshaw and I started AI Utah. And the purpose of AI Utah, why I'm passionate about it, is I feel like Utah has a lot of really strong talent here, but a lot of the individuals didn't know each other. And so a large purpose of the mission is to connect other people in that space, help people identify the nucleus of talent, and then allow those people to share and promote their learning as well.

Mason McMullen. I'm the head of rev ops and GTM strategy at Elysio. And I'm passionate about AI because we're an AI-first startup. So it's in the blood. And I think that, you know, I'm not biased or anything being a RevOps guy, but I think there's a lot of interesting applications for AI in revenue. And I'm excited and happy to be on the forefront of that and feel very lucky sitting up here with these smart gentlemen to talk about it.

Absolutely. I'm so excited. And also, as it happens, this movie trailer that just came up is so pertinent to our discussion, isn't it?

I'm like, what? They know us.

Crazy. All right, so who's got— Dave, if you can hear us, there's a movie trailer going on behind us, which is great. You got it? Okay, cool. All right, who wants to start off with the most burning question of AI that they have? Who wants to start? Oh, you guys don't have any questions? I have a million questions. Nobody's got one yet? Yeah. Tell us your name and your company.

Garrett Blood, KJ.

Where do you see AI going in the next 24 months?

I know it's doubling every 6 months. It's supposed to do that for the next 8 years.

What are we looking at in 24 months from today?

Sure. So there are multiple dimensions to look at it. Like, as if you look at the raw capability of LLMs, like the text is almost at a benchmarking. We are going to see a lot more improvement in the audio, video, and images capabilities. We are going to see like a phase where like you just give a prompt and it will generate a 15-minute like high-definition 4K movie which is like as good as the best director. That would happen in the future. So multimodality is one aspect where it'll kind of like go. And the second thing is that the integration layer, it will be like you will— we will be seeing AI integrated like very well in all components of our life, not just chatbot. We are going to see like an OS where you don't have to click 7 buttons to 12 buttons to book an Uber. To do a meeting, you don't have to go like to 10 buttons and click and look into like your email document and everything. It will be just through voice. So we would see a lot more proliferation of AI in like in our daily life and underlying capability of LLM and agentic AI that is going to improve a lot in multimodality stuff. So, yeah.

I think you need to pay attention to robotics. You want 24 months? I mean, with the agility of the robotic body right now, at least in humanoid form, having the advances that it is, and looking at China— sure, we're winning the AI war. Are we winning the robotics war? I'm not sure we are. And what differentiates a fully capable robotic body that has full tactile capability from an electrician, okay? The knowledge in the electrician's head. Well, with AI, you can take that knowledge and put it in this AI body of a robotics. So I think robotics is going to be an area that is completely overlooked, at least as far as the economic and the work impact in the United States in the next little bit.

Well, this goes down the line or not, I think that AI is technology, so we look at cloud adoption, just whatever that looks like. In 24 months, in my opinion, it's just the pace and speed of business. So what we're looking at AI right now is accelerating certain tasks, but as we look at it, the access to the alignment with the customers is going to be so fast because they'll have knowledge that's replacing us if we're not more and more relevant. So we're going to have to be faster at what we do, better at what we do in that differentiated space. So it's just more consumable. Right now, I think it's— I mean, ChatGPT is very consumable, but it's going to be very much more productized and like apps are today. It's just consumption model I see is what it's going to look like. And so now it's going to be the pace of business for us. That's how we view it.

I see us working with fewer apps and more screens and more data from companies aggregating and working together. So a much more simple life with the complexity behind it that you don't have to worry about.

Yeah, I mean, I'm going to talk about this from a purely revtech perspective because that's kind of the space that we operate in, the space that I'm familiar with. I think the biggest thing we're going to see is you're not going to need data transformation in between your data and AI. And I think we're already starting to see that play out in the market and with technology and its acceleration as fast as it's going. You know, there's a lot of companies that are kind of sitting in that space where it's like you need good data and you need to transform your data from source to actually get your AI to work. And I think we're going to be far past that in 2 years for sure.

That's awesome. Have you guys seen Whitney Merrill? Where are you? Whitney, remember what you showed me about Scotty with the interfaces and Do you want to just explain that metaphor for one second? Because it's so pertinent to what John just said about screens. And then Dylan's got the next question.

So the metaphor is Star Trek. And so we're living now in the world of, you know, building the ship while flying. Well, this is going at light speed, right, on a starship. Right. And, you know, we have engineers who they only run the ship. Because they've got AI, because they have AI with them. And this is the whole concept of, I don't even talk, I just have a light view on my dashboard of the enterprise that's there. But we, we look at that as a dashboard, but it's really AI. It's the computer itself. And that's the metaphor.

I love that metaphor. Thank you so much. Okay, Dylan has the next question about AI, and then we'll go back here. Thank you.

Thank you. My name is Dylan Ferguson with Fullcast. I have 7 children, and I'm seeing a lot of entry-level jobs leaving the marketplace that would usually be used to train you to become a ladder mover or a future owner. So especially for Jason at the University of Utah, how do you see the education system changing to accommodate for AI? Because most of my kids, are even allowed to use AI in school. So how are we training them and how is that model going to perpetuate bringing us the right kind of candidates into the workforce? And then for the rest of the group, I had a question. A big challenge we're seeing is that with the next layer of agents, the browser-based agents, they're going to be able to surrender the control we're going to give them access to systems, passwords, usernames. What are your thoughts in terms of how that affects both the softwares that we're putting out in the world and the security we have to deal with? And what are the opportunities that exist there?

Well, to answer your first question, which is a really, really good one, it's really interesting. When the calculator came out, there was a full-blown revolt by parents because they didn't want the calculator to be used in school. And obviously we know where that went. I still know how to do long division. I know how to run a calculator. You can still do the thing. But the calculator certainly aids and abets my ability to do my job or to do my things. AI, at least as it stands right now today, before we hit the AGI level in the next— well, according to OpenAI's teasers today, maybe tomorrow, who knows? But until we do hit that AGI level, AI is just simply an augmentative tool. It's not a replacement tool. And so what I am focusing on with my students and what I know other people at the University of Utah are doing is we are not discouraging people from using AI. In fact, we are encouraging it. And I think the teachers that are discouraging their students from using AI are doing themselves a big time— are doing their students a real bad disservice. Because, I mean, if you can use a calculator, why not use a calculator? If you can use electricity, why not use electricity? You just have to do it. So I know that it's going to affect higher education. How? We don't know yet. There's a lot of things that are going to happen in education as it relates to AI simply because people can go to ChatGPT and they can learn a whole heck of a lot of things right now. You don't need a teacher if an LLM can tell you a lot of the things that it's already found via research or in what was in its training set, provided it's not hallucinating. So to be determined. Let's just leave it at that.

I kind of want to comment on this, Dylan, because I'm passionate about this topic. I was, I was helped early in my career by, by mentors that bought into, bought into my skill set and helped me develop the skill skills to become a good employee and to improve and to grow into roles that I'm in now. And I think it's important, and there's a lot of operators in this room, there's a lot of business owners, a lot of CEOs. I think it's really important to keep those roles open for new grads, you know, and help them develop the skills they need to succeed in this interesting labor market that we're in. I mean, AI is not going away, but use it as a tool to equip your humans to be better. And to help them develop the skills they need because we can't sacrifice humans for the sake of artificial intelligence. We just can't. It's unethical. And I'm pretty passionate about that. So I would say keep that in mind as we go forward in the next 5 years. Let people come in and develop the skills that we developed 15 years ago.

I have a thing on that as well. And I won't speak to the technical password element. We'll let someone else do that. What does entry-level mean now with access to AI? Entry-level is now mid-level or better just immediately. So, I don't think it's losing the job at all. It's just recognizing that it's actually much more advanced and much more capable than it was before. We're an engineering company, so we have global engineers, and we deal with this all the time. Career progression. What does it mean? What does it look like? Am I replaceable? Am I a commodity? What does it look like? But my junior engineers today act like senior engineers 5 years ago at a much, much better price point, much more competitive capability, but it's what the market demands now, and that's what AI is facilitating, is junior doesn't mean junior anymore at all. So I don't think we're losing a career progression at all. So we're redefining what the subject matter expert senior people have to become because junior is now the new baseline mid-level. So yeah, I think we revel in that. Bring that on. We're much more productive, much more capable, able now and celebrate it. But there's one more level above that. That's the one that has to be even smarter and even better. That's the one that's at risk in my mind, not the, not the junior. It's the, it's the senior. You got to be better now than ever.

So yesterday I did a podcast with the founders of SchoolAI, and the ideal teaching scenario in their mind is one teacher and one student. And AI, I think, is being viewed the wrong way in the, in the, in the learning environment. Because teachers— one teacher is looking at AI helping 42 students cheat, and the teacher needs to realize this is where every student can have a personalized training person, whether, whether they're visual, the audio, however they solve problems. I look at it as every student will have a PhD-level teacher in their pocket to help them learn the best things. And I think when it comes into the business world, in the sales world, I run a dev shop and like, like the same type of things. How do you overcome objections? I take out, you can take out ChatGPT. Like one thing I, one of the things I did today for this was I was prepping and I put on ChatGPT on audio and I said, pretend you're Amy Cook. She has a lot of stuff published online. I said, train me, ask, have, ask me the types of questions Amy will ask me. Bring in questions about Fullcast. And your, your sales reps and everything can do that exact same thing, and it's $20 a month.

Uh, I think very good points like is made. I want to just, uh, bring one research which happened recently which shows that in students, the processing power of the brain and the activation of the neural network from the MNI, that is reduced. That is reduced because we are becoming too dependent on the ChatGPTs and like different tools, AI tools, and this will grow. That dependency will grow. We are getting more convenient and we are getting more like it's easy for us to use. So there is this question where like how, what is the intelligence of human like down the line and how much creativity and the innovation that we have in our mind that is like questionable right now.

So that is fascinating and so interesting. John, I want to know what Amy Cook GPT said to you.

She had really good questions. She asked me about the recent acquisitions and how it tied into AI and the future of AI and, you know, the direction.

Did she use a lot of emojis?

No emojis.

She was very nice. Okay, not quite me then.

There was not the Amy Cook sandwich.

That is awesome. This is really cool.

May I say one more thing? There was a colleague of mine at Stanford that had a really good quote that I think we could all learn from. We're not to the stage yet when we hit AGI where AI is replacing people, but people that know how to use AI are going to replace people that don't know how to use AI. So just keep that in mind. And you're absolutely right. They have to have the tools to be able to do the job.

Love it.

And we got to figure out how those neural networks can be activated better because my kids are now doing their homework with ChatGPT. That's a whole other story. All right.

All right, this is Josh Daniels. So one question, this is a nerdy question, meta, meta guy. So are you guys, are you, how are you guys playing around with like hierarchical reasoning models versus like a standard LLM for like, I don't know, it's how the brain thinks, right? We compartmentalize and bring it down. So I just, so that's the first one for you. And then a broader question for the team here, what are like the top 2, 3 skills you guys say people should be learning with AI? And specifically, let's focus on revenue. So what are those skills that teams in revenue should be learning of how to use, orchestrate, and work with AI?

Regarding the reasoning part, the model capability with the reasoning is going to get better and increase over the period of time. So right now, we are at a phase where It's thinking during the inference time. This is a test time coming in, that's why we are seeing a lot of milliseconds, like a lot of seconds of delay and latencies because it's thinking, refining, and improving its answer once over and over again. What we would be seeing, there are two directions which are happening. One is to reduce that time to think. Second is to actually have more like rather than one centralized brain, have small brains which which are called small language models, and each of them are responsible for a small part of thinking. Just like a memory, we have what the cerebral— I mean, we have different modular omegata for different purpose. We have a different for thinking, reasoning, for creativity, there are different things. Similar to that, there are going to be architectures which are already in research and prototypes are out, where you are thinking in that specific manner and the whole brain is not one LLM, but like the small mini LLMs which are individually very powerful and a very specific part of the job that the brain does. So over the period of time, we are going to see the intelligence which is going to be incredibly good, and we are going to have less hallucination. And, uh, yeah, it's, it's going to get more powerful. The second, uh, domain or dimension which is which is make the AI more affordable. And where the unit economics started working is that the models are going to get like smaller, like through the knowledge distillation, and there are different techniques to make a bigger model into a smaller one. That means the cost of the inference, or when we are doing a chatbot, or like putting a query or something, from a dollar it will go to like 10 cents. That's where we are going to see that like these companies are getting profitable. Right now, it's not in a phase where the cost of query is too high as compared to what you are paying in $20 or $100 or $200, like ChatGPT kind of thing. So these are the two trends which is happening with reasoning. One on the capability side, and second on the compute or the efficiency side. So, yeah.

Thanks. Interesting.

Okay, and now the question of the top one or two things Okay, so this is something I've done a heck of a lot of thinking about, so thank you for asking that question. I was on a— I was doing a podcast episode with the regional CTO of IBM, and before we did the podcast, we were actually debating this very question. And the answer we came up with, I think, very much applies to revenue organizations. There's really going to be 3 main things that AI is not going to be able to replace anytime soon. And I call them the 3 C's. That's what I call them on my show is creativity, communication, and critical thinking. Those 3 things right there are what you should be focusing on when you're hiring for people in revenue organizations or in some way, shape, or form training people, because if they can critically think themselves out of a paper bag, they're going to be doing okay using AI as a tool. If they're good communicators, well, hell, they're supposed to be doing good communication anyway, right? They're in a revenue organization. Okay, and if they're creative, they're going to come up with other ways to be effective. So that's my thoughts.

I think that when we look at what makes people amazing is the ability to connect and empathy. If you want to be successful in business, in sales, in revenue, in anything, can I connect with you as a human being and do I empathize with your business? With you as a human, that can't be trained— sorry, by AI. That can be trained, and I think that those are the skill sets that personally we look for, is the empathy element of human connection.

What we're seeing in the LLM space is that there's not one LLM that's just dominating everything. So like a really good example is if you look at OpenAI and v03, like they're very different in outputs. And I think as we're looking at tools and how to become better in the workspace, like one of the things I like about Fulcast is the last 3 companies they've acquired are all AI companies focused on RevOps. And to me, what I see in the marketplace is that they're going to be the leaders in this type of specific AI. They recently, they sold the company for $250 million. They took all their experience Now they're buying AI companies and all that knowledge is being packaged into the full cast AI. So you're not looking for one that's going to replace everything, it's going to be compartmentalized and specific companies are going to get really, really good at training and creating great outputs.

Can I add one C to your 3 C's?

Please.

I was going to say curiosity. I think that's one of the the biggest applications of AI, especially the generative AI that kind of just lives next to you as your AI companions, is like there's no better learning tool than deep diving with a ChatGPT-type tool, you know? And I think that's something that when we're talking about revenue jobs and the roles that exist in the revenue lifecycle, you know, that curiosity serves you well in any role, whether it's a CRO, whether it's a marketer, whether it's a RevOps person, getting to know people especially, you know, we've talked a lot about human connection, but I think coupling up that human connection with curiosity and utilizing AI to just be better at your job, learn more about the companies you're serving, or learn more about the roles that you're interacting with and your internal, external customers, it's like that will serve you better than I think anything that you can do in revenue.

I love that so much. One thing that a lot of you guys said, just to synthesize, is that you are prioritizing what we have traditionally called soft skills. Traditionally, deprioritized in business. And now what you're— what I am hearing here is that you are prioritizing these skills in business because AI can do so much of the computational process. So I think that's really fascinating and super, super relevant. Um, okay, so for— we have so many people who are interested in, in your questions. What I'm going to ask the, the, uh, people now is for you guys to tell us who you want to get an answer from so we can answer more people's questions. Okay. All right. Thanks, guys.

Hey there.

Jacob Strachina.

This question is for Aditya specifically.

Do you think that— we talked about AGI a little bit. Do you think that it's possible to reach AGI under the current day transformer tech that underlies all of the major LLMs, or is it not possible?

Okay. So this is a question which is very controversial. Not controversial. I would say more like Uh, first, there is no clear definition of AGI. Like, everyone has its own definition, right? I mean, for me, the AGI could be like, okay, you are getting the best scientific paper in biology or like some kind of like domain which is very well vetted by people, and this is something which is never discovered. That could be an AGI. So the definition of AGI is very vague, um, but I mean, over the period, like, the model capabilities are going to get better and better. There is no doubt about it. Like, if you look at the expenditure like the companies are doing on the data centers and like there is, of course, there are like news and other things where like the model capability are going to get to multi-trillion. If we go to that level, that means we are exploring something which is like incredibly like out of our reach. So there would be experimentation and those things on that part and right now we are dealing with like at a multi-hundred billion dollars, like, in a day to use, right? So, uh, the model capabilities are going to get higher. Like, when it comes to AGI, it's like, it may or may not— like, because there is no standard definition of that one, there is no benchmark. And that's what scientific committee is also doing. You would probably have better idea that, like, how do we define AGI? How do we create a benchmark to say that, okay, if we reach more than this accuracy on this percentage, we have achieved AGI? right? So those things are still like in a framework and like a different organization bodies are doing. I'm pretty sure you would have like a lot more on this one. Yeah.

Isn't it a sliding scale too? Just it's continually changing.

Yeah.

Yeah. It's a changing. Exactly.

Awesome. I love that. Okay. I know I said just one person, but I actually really wanna hear what all of you think. This is actually so interesting. So let's have it. Let's hear it from everyone.

Stock up your food supply. No, I'm of the mind— we think a lot alike in this— that LLMs are great at language, but they're great at language, you know. Are they great at everything else? Well, can transistor-based technology really do that? I don't know. It can do a lot more than we thought it could 2 years ago, but Is it ultimately going to end up being the main orchestration engine for an AGI brain that's made up of certain composed components, including an LLM, an image generation engine, uh, who knows what else? I think there's another technology out there that we have yet to really refine that's going to end up being the master of all of these subprocesses. And when we find that,