How can your organization leverage generative AI today to reach key goals? Clarence Lee, co-founder of Eisengard AI and a former professor at Cornell’s SC Johnson College of Business, joins host Chris Wofford to examine opportunities for the cutting-edge technology in sales and marketing.
As co-founder of Eisengard AI, Clarence Lee spends his workdays examining how businesses can leverage cutting-edge artificial intelligence (AI) technology to improve their workflows. The use cases for marketing and sales are abundant — from copywriting, A/B testing and customer relationship management to pipeline operations, pitching and cold call strategy. Lee, a former professor at Cornell’s SC Johnson College of Business, shares how companies can apply academic theory to create AI business frameworks for those routine lead- and revenue-generating practices in this episode of the Cornell Keynotes podcast from eCornell.
In conversation with host Chris Wofford, Lee explores:
Discover the latest best practices for AI in eCornell certificate programs:
Additionally, Clarence Lee is an author of five marketing certificate programs:
Learn more about Lee on his website and get the latest updates from his company at eisengard.ai.
Books and authors mentioned in this episode:
Did you enjoy this episode of the Cornell Keynotes podcast? Watch the Keynote.
Chris Wofford: Today on Cornell Keynotes, we are looking at ways to deploy AI technology across your sales and marketing functions. Where this AI podcast is different from the others is that my guest Clarence Lee knows a lot about AI and a ton about proven academic theory and actionable business frameworks. He studies this stuff like crazy. Clarence also reviews some great use cases for generative AI and sales and marketing and how to get the most return on your investment. Great guest and lots to learn in this episode. So be sure to check out the episode notes for details on eCornell's multiple AI courses and marketing certificate programs.
Chris Wofford: And Clarence even provides an excellent reading list to get you learning about what's important here. Listeners, here's my conversation with Clarence Lee. Let's begin by learning how you're helping companies that are using AI for various functions, right? Efficiency, high performance, optimization of sales, marketing functions. Where are the opportunities from where you sit right now?
Chris Wofford: Where the quick wins and the long term areas of focus that companies need to be focusing on?
Clarence Lee: You know, one thing that we need to think about, no matter what industry, you know, the, uh, the folks in the room, you might be doing sales and marketing in CPG.
Clarence Lee: You might be doing sales and marketing in like a FinTech startup. At the end of the day. There is usually some sort of customer or consumer outcome you're trying to drive. Right. And so based on that you can ask, Hey, these are the outcomes I'm trying to basically affect and change, and what are the levers that I want to dial up and down, whether it's an A/B test that you want to run with your marketing copies, whether it's an email message that you may want to tweak so that you can send out to your highest prospects.
Clarence Lee: It could be actually a pipeline management thing where you might be actually having the data and your customer relationship management tool, your CRM, and be able to get a sense of, you know, gosh, my sales team, their most valuable resource is their time and also their relationships. How do we help them think about that?
Clarence Lee: So that everybody has 24 hours in a day. How do we make sure that one hour that you spend can yield returns? I found that that's at a very high level, the overall patterns that if you want to apply before we get to the AI, just data, to decision making. That's the general pattern that I've noticed that is helpful to think about what use cases are and um, reason why you want to do that is because I also unfortunately see a lot of common patterns where folks are just trying to do the latest things, the buzzy stuff, and they have no idea specifically of strategically what are we doing and what are we not doing.
Clarence Lee: And so having this levers in the outcome framework it kind of helps you think about What are the two or three outcomes I'm trying to enact? Yeah. And so again, this is one of those things that didn't necessarily come up all by myself. I actually stole it from a book. It's from, I don't know if folks in the audience have heard of Principles by Ray Dalio.
Clarence Lee: It's a book that was shared to me, by my co founder. And this book is written by you know, the man that started Bridgewater Associates. Right. And so they apply these idea of data driven decision making to investing. What I like the most from the book, I mean, there's stuff I agree with that stuff I don't agree with, but, what I like the most from the book is this idea of being focused on your outcomes and being thoughtful and strategic about what levers you can pull up and down to do that, is something that you know, we try to tell our clients from the beginning and so everything kind of falls from there.
Clarence Lee: So, happy to talk about the use cases a little bit more. Um, you know, we think about applying to marketing and sales uh, whether it's prospecting, lead gen, pipeline management, scheduling. We could talk about that, but I wanted to start with a high level view, if that makes sense.
Chris Wofford: The things that initially appealed to me were the things that are, you know, kind of, kind of mundane tasks.
Chris Wofford: Things that we get in the weeds with. Scheduling. Who do I need to contact? What do I need to do? Right? Uh, Managing lists. you know, getting, getting your stuff dialed in for the day is the stuff that I found really appealing. Again. For salespeople, time is number one, right?
Chris Wofford: That's the thing we need, need the most. So in our invites, I should mention that, you know, we told our viewers in the overview copy for the event that we'd be examining some of the educational concepts, theories, and frameworks that can be applied from your work in the educational space to business, right?
Chris Wofford: And these are the kind of stuff that you teach in your online courses. Can you talk about some of those things where you're taking theory and applying it in your work?
Clarence Lee: Yeah, absolutely. So, just to start out, right? Some definitions. You know, if we think about what frameworks are these business frameworks at the end of the day we can think about there's the theory that I used to teach and, you know, my colleagues are way smarter than I am.
Clarence Lee: Like, they don't come up with their theories of themselves, right? It could be in the form of papers. It could be in the form of books. It could be in the form of you know, podcasts itself that if it's not in published, but that theory, usually there's some sort of science behind it. Okay. And so the rigorous theory um, you know, when I, when I was being mentored as a PhD student my mentors were telling me because we're business academics you got to have the rigorous theory, but without relevance and applying to practice, it's not really helpful.
Clarence Lee: Right? And so if you think about this equation of rigorous theory, plus applying to practice. As the sum of useful uh, business frameworks you know, let's start there. Okay. So that's what a good business framework should do, but what are they, right? You can think about some examples of frameworks, frameworks are really just, you know, in a business school we have frameworks galore.
Clarence Lee: You know, if you're going through the two year program, right, you're learning about frameworks, that's a marketing and finance, you know, like what DCF is and a whole bunch of acronyms, let's say, okay, I'm going to give you a little bit of that applied to marketing and sales, but then we're going to quickly turn to, well, how can these frameworks could be used strategically by our listeners so that it's not just another slide that you hear in exec ed, but rather, gosh, this might be something that, you know, your direct reports could pull from a large language model and then basically apply it, you know, moments before you're about to, you know, like actually use it on real data.
Clarence Lee: Right? And so there's a lot of stuff there. Um, so don't, for the lsiteners that are here, I don't want this to just be another talk about business school frameworks touting the the virtues of it and stuff like that. I wanted to have teeth for you guys, right? And so I think one of the exciting things about generative AI is actually the application of these frameworks to your hyper contextualized businesses in your context so that it actually has teeth.
Clarence Lee: Okay, so I'll give you some examples from the academic side, right? So, on the academic side, when I was a marketing faculty, so, um, my colleagues and colleagues in the field, you know, there's, there's a couple of framers that I found. It's a very helpful way to think about your customers in the sales and marketing domain.
Clarence Lee: One of them is this idea called the customer lifetime value. Okay. We all heard it. A lot of the VCs, any VCs that are in the room, right? Anybody that's from Silicon Valley in the room, you've heard a lifetime value, CAC. You know, I think it's even mentioning the, an episode of Silicon Valley, like the HBO show, right?
Clarence Lee: So fine. Right. Lifetime. What is it? Okay. And so, this idea and the active academic side, you know, you can, I mentioned books that came out of largely Columbia, Wharton and HBS, where, you know, you got folks like Don layman, Sunil Gupta um, our own at Cornell, you know, Young Hoon Park and Sachin Gupta, for instance, they're all masters at customer analytics.
Clarence Lee: And it's this idea of today, if you're some type of business that have customers that are not only making one time purchase. But you're making recurrent purchases. How do you, how might you use those recurring cash flows you're getting from the customers so that you can model out how their lifetime value is to your business?
Clarence Lee: And from there, you can think about, well, how do you, might you prioritize? Okay. And so, you know, that is one idea where when linked with cohort analysis, super powerful to help you think strategically. My team, you know, and I got together a demo to show you exactly what that would look like from a sales director's perspective in just a minute, right?
Clarence Lee: So we'll come back to that, but that's where the academic roots of that looks like. Now on the academic side, there are some folks where, you know, they took that even further, right? So, Dan McCarthy at Emory he's a Wharton grad, his work in Pete Fader, you know, they did a lot of something called customer based corporate valuation.
Clarence Lee: Don't want to steal their thunder, but the whole idea is a really very powerful one where. You know, if you know what the lifetime value of your customers are, and you can figure out and apply other things like the 80 20 rule, the Pareto rule, and all that to it ideally, you can potentially value at least some direct proportion to your entire business.
Clarence Lee: And if this is a public company, then how do you think about actually using that information to actually trade on the public markets? Right? And so that could apply, that has implications on the private equity markets as well. So all of that is kind of tied together. But why did I start there? The reason why I start there is because lifetime value this is one of the most powerful ways to actually think about the outcomes that you want to drive.
Clarence Lee: And if it's directly linked to your, you know, public traded value, right? That all even better. And so, In the C suite, there's a lot of talk of this, but what I've seen actually in my own firsthand experience is that we are just at that turning point where, you know, the last couple of years I find that a lot of CMOs and CEOs are talking about lifetime value.
Clarence Lee: They want to calculate it, but there's always like a little bit of holes in their organizations where that's really preventing them from applying this concept uh, in its full glory.
Chris Wofford: What are those holes?
Clarence Lee: So for instance the lifetime value. Idea, you know, it basically it composes of a couple piece of key pieces, right?
Clarence Lee: And so there's an equation, there's various forms of it, but if you could put a dollar value on your margin, if you could put an idea of what your retention rate looks like, if you have a sense of how many cohorts or how many segments of customers you have and then what their respective margin lifetime uh, margin retention rate is, as well as, let's say, the cost of acquiring a customer, the average cost of that.
Clarence Lee: You can almost mix that up in a way, throw it in an equation, and then you can put it on there, right? But, you can think about um, a couple of the typical holes that we see is, well, how do we get that data? Oftentimes, if you've been calculating the margin, that sits with finance in an organization. Whereas, if you are the head of sales or the VP of sales, you might not have that data readily to disseminate out to your directors.
Clarence Lee: So data all in separate places, they're kind of sitting all around, you know, how do you actually merge that together? So it's consistent and also makes sense. The other thing also is like differences in definition, this idea of a disconnect from the very top, the C suite all the way down to the actual salespeople on the ground.
Clarence Lee: I mean, you know, sales folks, right? They're rewarded for having, you know, very targeted attention spans, right? And being able to focus on the next thing. And they should operate that way. But if you want to sit them down and go and actually sit down and do an Excel sheet to calculate lifetime value, you have to go, they have to go coordinate with black people to figure out what the retention rate looks like.
Clarence Lee: Wait, what is a cohort? Remind me that again, right? You can see it adds up. Now what's exciting about everything I just said, if you go back and actually play the video uh, the audio, each one of those things I You can actually now, at some point, turn it in uh, a query into a language model so that some type of generative AI can actually help you figure the data out and also play it back to you.
Clarence Lee: So we'll show you an example of that, but that's on the academic side. And now on the practitioner side, you know, we've heard of books that are not necessarily the new, newest ones out there, but um, it's been vetted by the C suite, right? C suite, CROs, CMOs. On the marketing side, there's the classic how brands grow.
Clarence Lee: Yeah, there's, these are colleagues coming out of Australia, but you know, this is something where my own students, when they're interning in Proctor Gamble they really need to just know that stuff cold. So reading a book, but then PnG internships aren't exactly a walk in the park. So, when you're late at night trying to figure out, wait, what are the concepts from the book?
Clarence Lee: How do I think about double jeopardy? Didn't you wish there could be some type of assistant that's almost co piloting with you throughout to remind you of that? So that's one book from the marketing side. And there's way more I could give you guys. On the sale side, I'll give you one old and one new one, right.
Clarence Lee: The older one is something called the challenger sale. We've heard CMOs, CEOs, CROs, and kind of like playing that back to us, right? Say, okay, cool. Can you help me? Apply the principles in the challenger cell to my org. A new one, especially for the listeners that are in the tech audience. Um, there is um, there's a Cornell alumnus I believe forgetting his name is right now, his name right now, but um, this, this uh, it's a book called Game of Sales.
Clarence Lee: Um, you know, it's from his um, uh, the, the author's experience as a big tech sales leader at the Googles of the world. Right. And so, that's something where literally in that, and there's like a chapter in there talking about, well, how do I achieve perfect focus through proven business frameworks?
Clarence Lee: And so that's, something where, you know, just want to give the audience, a sense of what, what are frameworks, how are they helpful? And to me, the biggest opportunity is that before the generative AI stage it's really hard to apply these frameworks and there's a certain bar that you need to pass in order to be able to actually apply it.
Clarence Lee: You know, and that's where a lot of these consultants come in. But with the latest developments in generative AI over the last gosh, 24 months, there's a lot we can do now that we don't need to have a call to McKinsey to apply this.
Chris Wofford: Yeah, that was David Perry on Game of Sales. So let's talk about the recent developments.
Chris Wofford: I have you here. I want to get your take, you know, just the last few weeks, we've seen huge news, kind of blockbuster blitzes from Open AI, a really, really big one, right? Chat 4. 0. Google and Microsoft. What do you think about what's going on right now as it relates to what you've just set the table with in the previous 18 minutes here?
Clarence Lee: I'll give you guys one idea that I think all the big tech companies are talking about. That's going to be important for the next 6-12 months. And I'll maybe mention one idea that they're not talking about. Okay. And so, let's start with the idea that they're talking about.
Clarence Lee: So I think for the next 6-12 months the business world will be talking about agents. Agents. Agents. Yeah. And so, Karan is going to give a whole session on this, right? So I think he gave a talk last week. I encourage the listeners to go check it out. I think that's a fantastic overview of what's happening with generative AI.
Clarence Lee: But I don't want to steal his thunder, like, but the basic idea about agents is it's like, that's the next evolution of how we as individuals, as well as businesses can really leverage the power of generative AI. So you think about what the last 24 months you know, you have Chat GPT, you have.
Clarence Lee: Amazon their version, you have Microsoft's co pilot and all that. Essentially the way you access it, is you open up a browser or you can download their app on a phone and you would ask it prompts, you would ask it things and it would give you recommendations, right? It's almost like a search engine.
Clarence Lee: But much smarter, right? And you can ask it, okay, to write code for you. It can, you can ask it to, you know, maybe figure out, verify certain things. Does it make sense? Does it not? And, you know, even now there's like tools that it could call where you can ask it to go and actually invoke a search tool where it can actually go on the internet and search query for you.
Clarence Lee: You can actually have a tool that it connects to a database. Fine. Okay. So that's chatbots up to this point. Now, the difference between chatbots and what agents are is that agents do not just recommend and answer your questions for you, but rather it can actually execute actions for you. So for instance, instead of you drafting, asking it to just write and draft a letter of an email to certain wording, it can send an email for you instead of going to Amazon and actually, you know, do reviews and, you know, summarize some of the reviews on a, on a, on a certain product that you're looking at, you can actually task it based on certain judgment of things.
Clarence Lee: To go and actually purchase the product for you. Right. And so, you know, Chris, if we have time later, I'm happy to kind of go back into that and it showed exact examples of you know, what agents can do. Yeah. But that's what everybody's talking about right now. I'm sure you guys are going to find, you know, dozens of that, like tons and tons of videos about what agents are.
Clarence Lee: Microsoft talked about it, I think more than a month ago, Google talked about a month and a half ago, Amazon just did this, I think it was like last week or the week before. Right. And so, you know, you can expect everybody's going to have their own say about it. So. Encourage the listeners to go out there and kind of Google and just watch every snippets of what Amazon's take on it is, what Google's take on it is, what OpenAI's take on it is, and what Microsoft's take on it is.
Clarence Lee: They all have their own little flavor, but agents, agents, agents, that's the talk of the town this year.
Chris Wofford: You can almost feel the demand for this kind of thing that you've just described. And one thing that's notable uh, you know, we've been looking at this over the last couple of weeks, in the recent advances among OpenAI, Google, Microsoft, and in generative AI, is that there seems to be kind of a reconciliation of the tools, or kind of, a functional integration, right?
Chris Wofford: It's certainly in the case of ChatGPT 4.0, that's exactly what we saw. It was, um it was doing all the stuff. Can you kind of explain a little bit about what is happening? You know, because I, what I'm getting at is that the technical advances, not super groundbreaking, but the reconciliation, the integration, the confluence of those things, functionally working together was really striking.
Chris Wofford: That was the thing that really kind of stuck with me.
Clarence Lee: Yeah, absolutely. So let's talk about a confluence, right? And so I want to give a big caveat. You know, this is one of those things where a lot of this confluence right now that we're talking about it's based on the transformer type of sequential large language model based architecture that came out back in, gosh, 2018, 2017 by Google.
Clarence Lee: So what I'm about to say kind of applies to that train of thought, right? But what's interesting for the audience too, if the folks I want to see also like what might be coming down, you know, further down that's coming out. Yann LeCun from um, I, I might have mispronounced his name, but he's an expert on AI and he's with Meta and also NYU.
Clarence Lee: Thank you. But I thought a lot of the work that he's actually talking about related to uh, the jetpack architecture, right? So join embedding you know, these like that are that idea as well as how to actually get AI to help you plan. That's just beyond the whole large language model. I'm going to pare back.
Clarence Lee: Um, you know, the tokens I've seen before, right. In my training. I just want to start, like, I want to start there. I just got to put that out there. That's a, that's the caveat. But based on that, the language models, I think one of the ideas that I think Bill Gates talked about is with Microsoft.
Clarence Lee: There's this idea of having a smart AI that can really be personalized and knows you. That is something that is like the million or the billion dollar question, right? Everybody's trying to figure out how to build that. And so the question is, how do we get to that and what are the missing pieces to that?
Clarence Lee: Right. And so holding the LLM structure fixed and there's other ways to get to this, but holding the LLM structure fixed um, I see a lot of people talking about how to build smart bots, but I don't hear a lot of people talking about how to build wise bots, having this idea of not just having AI to be able to play back information to you and it knows you, but rather how can you have an A.I. that guides you because it knows best practices. It knows it's seen, you know, given certain industries or cross industries. You know, what are the not just what to do, but actually what not to do and how best to do it right. That's the kind of slight difference that um, it's a little more nuanced, but traditionally in the pre AI stage, these are the roles of the consultants that are out there. The McKinsey's, the Bains, the uh, the BCGs, they're brought into a company to guide a company's leaders, to think about strategically, what do we not do? What are the best practices? How do we actually do the same exact thing, but do it in slightly different way so that it yields maximum output.
Clarence Lee: So Chris, I think, you know, on this, on the slide right now, I put together a visual for y'all. Right. And so we think about where current AI is. Um, and this might be actually in the past now already, but want to borrow this phrase from Karan in the last session, he called it.
Clarence Lee: I think something like current AI is like cheap dumbness.
Chris Wofford: Yes.
Clarence Lee: Is that right? Yeah. It's so, so like, not to disparage all the amazing things that, you know, it'd be an AI and the research that they have done with alums. But the focus on the current AI is this idea that it has reasoning. It has natural language.
Clarence Lee: And so you could think about these LLMs as really good reasoning engines, right? I'm going to steal a page from Andrew Ang. You know, that's what he was talking about, how to think about the current LLMs. And if you look at all the keynotes that are coming from all the big tech companies we're basically getting to here.
Clarence Lee: The smart AI that knows you. Now, how does it know you? In this space, there is a hot word of, you know, the past 12 months called RAG systems, and it stands for Retrieval Augmented Generation. Also came out of academia, but then, you know, it's being applied by not just big tech companies, but some of the biggest vendors out there in Silicon Valley, right?
Clarence Lee: And so everybody's kind of doing the RAG things. But what is a RAG? At the end of the day, I don't know. It's a smart bot that can actually not only reason and have natural language, but it can also connect to your own data and documents. So think about your SharePoint, right? Be able to connect it to your co pilot and it's able to peruse all that, but that's fine, right?
Clarence Lee: It can remind you of things, give you all that stuff. But what we really want is to get to the stage where the AI itself in the moments when, you know, I forget things or I've confused things, right? My brain fails. It could almost guide me in the way that I want it to.
Chris Wofford: What would be an example?
Clarence Lee: Yeah, let me give you an example for that, right?
Clarence Lee: So this is where the business frameworks come in. And so, I actually have a demo I can show you guys. But a very simple case is as a entrepreneur now, I'm juggling 50 different things. I'm sure the sales leaders and also the marketing leaders that are out there are not so different. Every Sunday, I sit down, you know, I think about my weekly planning.
Clarence Lee: What are the most important priorities this week? What am I going to do? What am I not going to do? And so this idea of, there's so many times when I'm like, on Sundays, I'm like, you know, I just finished putting my kids down for, you know, for the night.
Clarence Lee: I'm tired. My brain's not really working. Trying to choose what I'm going to do this week and what I'm going to encourage my team to focus on and what not to focus on. It has to be tied back to our quarterly goals. Our monthly, you know, milestones, right? What are we going to say to our investors and so forth?
Clarence Lee: Just in the same way that if you're a sales and marketing leader, I'm sure you have your annual targets. You have your quarterly business, you know, review. You have your monthly business review, right? So there's almost a thread. And an arc that in the natural cadence of your business you need to basically operate.
Clarence Lee: So consistency is really hard for human beings to do, right? But imagine having an AI that knows all that context, but then are able to pull in frameworks from books, you know, Seven Habits of Highly Effective People, Getting to Yes, Atomic Habits so that I can almost hack my brain to, you know. Do you know what the aspiring optimal clearance could do where most of the times I'm not optimal.
Clarence Lee: Right. And I'm just like, Oh my God, I'm just trying to put it together. Right. And so I think that's one of them in a very high level. That's a one particular use case. I love to show you guys that, right. Like when I'm saying that, you know, this uh, this is very, You know, okay, that sounds great, Clarence, but does it actually work?
Clarence Lee: How does it actually work? And so anyways, I'm going to stop there. That's one use case. There's plenty of ways to actually play into that.
Chris Wofford: You know, the thing that I found kind of intriguing and I go to the you know, the place of unintended consequences, right? What happens when we do liberate ourselves from all these other tasks and functions that we need to do?
Chris Wofford: And we are actually, just empowered to do our work nonstop, right? It'll just be interesting to see how, what happens with our brains once we become that much more focused and have, and are, and are liberated from so many of these things and empowered by, you know, by best practice and texts and all those kind of things.
Clarence Lee: you know, you know, Chris, I think I'd love to, maybe we should have another conversation about that whole session talking about that.
Clarence Lee: I think there is an idea about freeing up time with AI. But then the separate idea is what are you going to do with all that additional time? Yeah. And so is there should it should there be a framework for that? Should it should it be a mandate? My way of thinking about is, you know, yeah. Like, no, it should be you, right?
Clarence Lee: Like deciding what do you enjoy doing? What's going to get you to thrive? What are the, all of us we're, we're a little bit different, right? From each other, right? There's some things that we just love thinking about, you know, some things that we just absolutely hate doing. Is there a version in the near future that we can at least get to where if you can almost spot out and identify, you know, in your own routine, what are the things that you hate, right? These blocks that are just preventing from getting the next level and and be able to potentially have these micro agents to help you. That's one way to do it from a micro level. From the macro level, imagine, you know, as a leader, as a manager, if you're able to identify all those spots, you know, in your organization.
Clarence Lee: Right. And so, and maybe like, I can show you an example of like what that looks like. I've seen how people can remove um, these blocks that people might have in a sales marketing organization.
Chris Wofford: Would you please? Yeah, let's do it. Let's check this out. So what are we going to be looking at here?
Clarence Lee: Yeah. So if we actually bring up the visual right here um, what you're looking at here is I was trying to think about, you know, if today you're a leader in an organization, how might AI help you now? I'm going to give you guys like a one, two punch kind of framework to think about it. I'm not going to tell you what to do, right?
Clarence Lee: But this is very logical, right? It's not rocket science, but what we found, this is helpful in our initial engagements with our clients. In this setting up piece, right, this white glove piece where we're just basically doing integrations. You know, we, we sit down with them we, we talk to the C suite, we talk to the different heads of the orgs that are in the organization, we ask them one very simple thing.
Clarence Lee: We ask them, in your organization right now what are the highest points of leverage that can give you either that increase in revenues or cost cutting or some type of outcome that you need to report out to your investors report. So we start from the top, very top, so very clearly define the level, the magnitude and the outcomes that we need to hit.
Clarence Lee: Now, after that, you know, typically we find that if there's a sales team in the organization, marketing and sales are like the first front lines for that, because that's where the revenue comes from, right? And so, what I'm about to tell you can actually apply to marketing as well, but let's focus on the sales org.
Clarence Lee: So, if we focus on the sales org you know, one way to actually dive into that would be to think about, we can start by looking at the natural cadence of your organization, right? And so, as we mentioned before, for a sales team, the time is most valuable, right? And that applies all the way up to the head of sales.
Clarence Lee: And so if we actually think about today, what their time looks like in a given year, zoom that back down to their quarterly, their monthly, and their weekly. That's how one might actually look at their and diagnose their organization. And so if we look at this we can actually drill all the way down to say, okay, today, as we're getting our team to really, you know, blow their quotas out of the water.
Clarence Lee: Do we start from the top? Do we start from the bottom? Do we start in the middle? Your call. For all the audience listening today, it could be very, you guys could be in various stages of digital transformation, right? In different various stages of data quality. So it might be at the bottom, it might be at the top, it doesn't matter.
Clarence Lee: But one of the things that we hear consistently, that is a, is a block to data quality in a sales organization is the fact that their sales reps don't track, or at least don't track reliably. And why is that? Well, if you think about it, what we hear from sales folks is they're busy. They're on the road, right?
Clarence Lee: They're trying to figure out whom to actually go and talk to. And you know, the best sales reps, right. Are planning this each week, who are they going to visit? Right? It could be 20 people all the way up to let's say 60 people in a given week, if they're crazy about visiting, but that implies, you know, in a given day, right, you might actually have no more than seven to 10 people.
Clarence Lee: And that's a lot of ready to visit. Okay, so they actually might plan out their, their visits, and they actually have to go and drive the uh, actually visit the the customers, and then they got to close the sales to meet the quota. So, not rocket science, this is kind of how their, their thing is.
Clarence Lee: But what you'll notice um, here, what's in red. Is this idea that, well, these are the parts which we found that, you know, planning and also like getting to that closing the sale, that's usually kind of the most painful parts as a salesperson, right? Like, you know, sales folks are really outgoing. They're gregarious as they're visiting the folks that are really good at doing that, but having those tough conversations, how do we actually frame it?
Clarence Lee: As well as whom do they prioritize and extra planning. Those are some of the issues that we saw. Okay, so just you'll find other red spots as we go up. But, you know, this is like very tactically. Okay, in the, in the week of a given customer, you might actually find this. So let's bring back the business framers.
Clarence Lee: Now I'm going to just steal from Sunil Gupta's book, Driving Digital Strategy, who's my, you know, advisor from back at HBS. So, he won't get mad at me, I hope, for stealing his stuff. But one of the things I really like his uh, his, from his book is this idea that, you know, if you're a sales market leader at the top, how do you think about acquiring you're retaining your customers and expanding it and linking it back to a company level lifetime value.
Clarence Lee: That's a, that's the big strategic idea of how to prioritize, right? So if you identify, I think even Tim Ferriss talked about in his book, like his original South by Southwest talk way back before he was famous. He talked about applying the Pareto principle to his actual list of customers, right? So you'll find that if you, you know, for the folks who know who Ferriss is, I highly encourage you guys to check out his book, but long story short this is essentially what Suneo talks about.
Clarence Lee: So if your directors have already did that, that strategy and deciding, okay, who should you plan, right? What are the high value customers? That solves a lot of the pain for you. You can just go and say, all right, if you have something I could tell you, these are the five people I need to visit this week.
Clarence Lee: They're high priority. Let's go double down on that. Let's go and visit them, right? Like that's, that's one example. Another example is closing. Now the challenger sale not to give away the punchline, but it's this idea that all sales reps out there and selling styles can be distilled into I think five or six different types of styles.
Clarence Lee: And there's this idea that um, great sales folks are made, not born. So what that means is if you knew which one of these styles can yield you the maximum closing, you probably want to learn how to actually adopt that style, right? And that's the idea of the challenger style. There's very specific things, like three things that you do.
Clarence Lee: I think it was like differentiating, teaching for differentiation, tailoring for residents, and then there was also like take control of the sale, right? Anyways, that's a framework, right? So, instead of having a talking head like me, talking to you about how to apply that, what if you have a generative AI that can remind you right before you're about to go drive out to that person this week, I'm going to go talk to Chris and close the sale.
Clarence Lee: What if it could actually show me, okay, here is resume state. Here's like where your last talking script that you're planning to talk to Chris about. Here's some new information that's come in about Chris's priorities. Let's adjust the script so that I can go talk to it. And so there's ideas about how agents and actually you could automate some of these things for you also.
Clarence Lee: Right. So I can talk about that, but um, you know, so that's the idea of how we bring it all the way from the very top, all the way down to the ground level in the day of a given salesperson could apply a framework.
Chris Wofford: We've got some viewer questions here. I want to hit Clarence, if you don't mind. And thank you for those shout out to Peter for submitting three or four, which is great. One thing that I was thinking as you were offering the demo here, well, Peter asks this, he asks how can a company be sure that AI acquired data is true, real, and worth using for strategic business decisions.
Chris Wofford: My head went into disaster mode, right? AI is only going to give it what allow it access to. You're providing, you're providing the reading material, so to speak. And if it doesn't have integrity, it could even screw things up even worse. Then if it was never, I think you know where I'm going with this, right?
Clarence Lee: Absolutely.
Chris Wofford: If everybody's sharing bad data, it was always, it was already dangerous. But when we've got AI, trying to guide behavior in the form of, you know, like you've just described, right? The wise assistant. Disastrous, right? There must be increasing priority as it relates to data integrity and all of this.
Chris Wofford: How do you think about this?
Clarence Lee: Yeah, that's a great question. Here's how I think about it. I want to challenge everybody in the room to think about the following alternative. Let's forget about AI for a minute.
Chris Wofford: All right.
Clarence Lee: Imagine you have a new hire, a junior, fresh college grad or MBA grad that you just added into your org.
Clarence Lee: They're smart. They're well meaning. Presumably. Yeah. And then also they could reason and logic. That's essentially, you know, think about applying to your question about making sure that that person that is like your new hire, how would you make sure that they don't screw up your your company?
Clarence Lee: Yeah. And so, you know, there are 10 typical ways of actually making sure that you check up on them, making sure that you're actually giving them certain bounds of delegation, right? What do you have the authority to do versus not? I think there's another book. I think there's a cant remmeber, who the author is, but.
Clarence Lee: Um, in my, In my learning as a manager I, you know, I, this book is called, I think it's called like, if you're doing it yourself, you're doing it wrong. And this idea of, you know, how do you actually think about, and all of you guys in the room have done this, you know, for a long time. It's like, how do you think about being very clear about what can your direct reports do and not do to the authority?
Clarence Lee: Giving them examples of what looks good and what looks bad, as well as if there are certain things of uncertainty, come back and check with a person, either me or somebody else, right? So, let's take this idea for a direct report, a human, you can easily just swap out that statement I just had there for an AI.
Clarence Lee: When you bring an AI into your organization, you don't want to do the big bang approach. You want to be very strategic about it. You want to think about, I think Karan was talking about this too, right? And challenging. The leaders to think about, well, I want to be educated about it. I want to think about, well, where are the points that I can actually go and actually start using some of these things to yield returns, you want to be very focused about how you apply AI first.
Clarence Lee: So you can do what can build trust with it. You can understand exactly what can it do, what can it not do, and what kind of situations will it hallucinate, will it not hallucinate, and so forth. And then lastly, you notice all the examples I gave you? A lot of it's like, hey, remind me of this, remind me of that.
Clarence Lee: It's connected to a certain database, right? That's very intentional because we found that, yes, the language model is going to get better and better, right? The big tech companies are doing that. But their job is to make sure that AI can be used for general purpose domain. Us in business, specifically with business in a, let's say sales and marketing.
Clarence Lee: A lot of times we have certain lingo, certain jargon, certain ways of doing things, even within our own company, that kind of dictates what's inbounds and outbounds. Well, use that information to your advantage. And build in checkpoints, just like you would do with an inexperienced grad. Now there's an actual systematic way of doing that.
Clarence Lee: It's called reinforcement learning with human feedback. And so there's ways to actually design the AI so that such that, you know, you could be having chatbots that are interacting with your employees. There may be agents that you're deploying for very routine tasks around. And one of the craziest ideas about agents I'm not going to talk too much about it on the call today is this idea that you typically have an org chart in your organization.
Clarence Lee: But what's going to happen going forward in the future as you'd start deploying agents, you're going to start having networks of agents as well. So you might have a shadow, not shadow, but rather you might have an AI chart, org chart where you're showing exactly within your org chart, what are the places where you're deploying AI?
Clarence Lee: And be able to monitor that, be able to see that. Right. So you can imagine even having an agent itself that's designed as a supervisor agent, that's going to help you to see, Hey, given the last month, what are the reports of hallucination? Which part of the work is happening? How are you using it? Right.
Clarence Lee: That's all data you can actually create and that stuff I'm not dreaming up. You could do that right now. And people are starting to do that. If you guys want, we can do that for you, but it's, um, it's I want you guys to start thinking about it that way so that you're almost designing in the old world as a leader, you're thinking about what your org chart looks like so that your company can outperform other companies in this new world, you're going to have that, but also an AI or layer that meshes with that.
Clarence Lee: What does that look like for your company? Do you even know? Going through that exercise will be absolutely crucial before you deploy this in large scale.
Chris Wofford: I think that's critical. And it's it's really instructive as well. And it got me thinking about where we want this conversation to go right now.
Chris Wofford: So perfect. You know, you had mentioned that the key and something that Karan mentioned a little earlier is what to build and where to put it, right? So if you envision the org in the org chart, the opportunities start to present themselves. How do you think about that? How do you advise for that? For those that are, you know, just getting into this, right?
Chris Wofford: Sales marketing friends specific or otherwise.
Clarence Lee: Yeah, so Karan, a couple of good points. I want to reiterate and I found that truth in my own experience as well. When I was doing research on AI it was hard to keep up latest research, right? Because new stuff are coming out every day.
Clarence Lee: And if you were just kind of chasing, you know, what's the latest fad? You know, it's really hard to distill the signal from the noise, so to say. So what he said that I thought was very helpful is before you go in and try to keep up, it's important for you to have mental model. A mental model of what's important and what's not, how it's going to help with your your context, right?
Clarence Lee: What do you, What do you as an organization care about? So I think. I'll mention his framework and I'll give you guys you know, one or two to think about one is this idea of I think he talked about having an idea, internal mental model to understand what can I do in terms of capabilities, right? I think he talked about classification.
Clarence Lee: He talked about generation. He talked about inference and there's a couple more, right? So he can go into that, right? And so if we're going to think about, you know, what your agents could do. You can think about, okay, well, is this a task that I, as a human being doing right now? Is it classification? Is it generating a report?
Clarence Lee: Is it inferring something from data, right? And so, being able to at least put those tasks out and lay it out like that, that's super helpful. Simple example, for instance, imagine today if you were to ask me, Chris, let's say I'm your team on eCornell, you ask me to write an email to respond to your customer inquiries.
Clarence Lee: It might be a block diagram where I, you're expecting me to go figure out, well, this email, what is the nature of it? Is it somebody complaining about my course? Is it somebody actually praising? Oh yeah, this is super helpful. Or is this just somebody that's like, Oh, I ate a sandwich. Thank you. It's like some random email, let's say, right?
Clarence Lee: So that's an example of classification. If you just take that and actually give it to a prompt right now in the form of an agent, it can classify for you by giving it bounce. That's super helpful. I encourage everybody out to do out there, if you have access to OpenAI, go into the custom GPTS, GPTs, and just put the prompts in like that, and you can see how well that does.
Clarence Lee: If you want me to actually write that email, that's generation, that's generating a text and stuff like that. So, that's, you guys get the idea. That's basically, you know, the framework of the mental model, what AI can do and cannot. Now, on top of that I want you guys to think about also for any use case that you're thinking about, maybe think about this use case as either a steam engine, a it's either a steam engine, it's a slide rule, or it's a crane.
Clarence Lee: What does that mean? Right? So steam engines, it replaced the horse, right? It could replace a human, right? Much like the loom did, right? It replaced a human effort um, where. Is this use case, you're just going to replace a human and there's no human in the loop. That's the first one. The second one with the slide rule. Chris, I don't know if you ever used the slide rule, but it's basically a calculator, right? It augments our ability to calculate things and so forth. And then so augmentation, that's the use cases there. And then so, and then the crane is basically using an AI to do things that was not possible before.
Clarence Lee: So, you know, today there's actually drug new drugs being candidate drugs that are being suggested by AI, where the research itself is able to progress at a level that is never possible before without AI, right? And so think about strategically, any use cases you encounter, are you going to basically do to do that?
Clarence Lee: And then lastly, you know, I mean, I'll flash up a slide for y'all uh, so you can take away with y'all is um, you know, like in, in just a little bit, but you know, to go using that arm with those two frameworks that went from me, if I'm from how do you then go into your organization and diagnose those red spots so that you can say, Hey, that's a, that's an opportunity for augmentation.
Clarence Lee: That's an opportunity for replacement. And, you know, that's a real opportunity for crane right there. Usually it's the easiest to start out with augmentation. Right. So I wouldn't start by replacing. And in fact, It doesn't make sense to replace initially. So, I'll stop there. I know that we're coming up on time.
Chris Wofford: You know, would you kick up slide 10, please? Let's take a look at that real quick, and maybe give us a little context. Because I do want our audience to take that away. And everything you've set us up with prior to this I think leads us to it. So, let's see if we can check out slide 10 before we go.
Clarence Lee: So, slide 10, that's the uh, the agents, right? Yeah. Okay.
Chris Wofford: Let me mention something while we have our audience here and making sure that gets displayed. Viewers, I want to drop a URL for you. We're doing several AI events.
Chris Wofford: Uh, we've been mentioning one that took place on May 22nd, I believe, and we're doing another one with Karan Girotra on July 1st. We're going to share those URLs, get registered for those, and naturally, the recording for this one will be available right after we're done, Clarence. So if anybody wants to share with friends or colleagues, absolutely do so.
Chris Wofford: So what are we looking at slide 10 here?
Clarence Lee: Yeah, so just a cheat sheet for the folks that are in the room wondering what agents are, right? A simple way to think about agents. I'm stealing from Amazon, right? This is what they talked about in their keynote. Agents are AI that can help you plan, assess, and execute actions.
Clarence Lee: What does that look like? It's basically a confluence of three concepts together. There's these prompts that we have out there. Respond to my customer's email for me. Okay. And so when you're doing something like that, you might want to actually pull the relevant notes from your SharePoint. So you have the information about the customers, right?
Clarence Lee: That's essentially an idea of tool calling. And then there's this idea of what's called a chain where you might have multiple prompts that you want to pull together in sequence, so you can almost like direct and tell the agent whenever you're doing this, don't just go randomly like do stuff, but instead, follow the sequence.
Clarence Lee: Go and help me categorize this email, what it's all about. Based on that, here's how you pull the relevant notes from my SharePoint. If somebody's complaining, go and pull that document from legal, and we want to make sure we incorporate that. Somebody's really happy about this. Maybe pull up their, you know, what's like the notes about them and actually echo back and personalize that note so that, you know, they feel heard, things like that, right?
Clarence Lee: And then respond to it, you know, so putting it all together, that's how you have an agent. You know, I'm really glossing over a lot of details here, happy to talk about that in a separate session, but I think Karan's got a lot of exciting stuff for y'all on this.
Chris Wofford: Clarence Lee I really appreciate your clear headed, measured delivery, the way you kind of go about this whole thing.
Chris Wofford: It's been very instructive for me, as always. For me, my, most desired outcome is to learn a lot and have fun, and you've helped out with both of those. So, I appreciate it very much. Thank you for joining us from New York City, and I wish you best of luck. We'll connect soon, I hope.
Clarence Lee: Absolutely, Chris.
Chris Wofford: Thank you for listening to Cornell Keynotes. Check out the episode notes for information on eCornell's AI courses and marketing certificate programs.
Chris Wofford: And thank you again, friends. Subscribe to stay in touch.