Navigating the AI Revolution: Practical Insights for Entrepreneurs

Full transcript

Arun Jagannathan: One example of how people would tend to resist was: We have someone who’s a counselor and a large part of the counselor’s work is dependent on his or her knowledge of the application process. So when we first demo’ed ChatGPT, I threw up all the questions that you could possibly have a student asking, and her first reaction when she saw it was, well, this is my job. This is what I do for a living, right? And it’s gone.

Darius Teter: Welcome to Grit & Growth from Stanford Graduate School of Business, the podcast where Africa and Asia’s intrepid entrepreneurs share their trials and triumphs with insights from Stanford faculty and global experts on how to tackle challenges and grow your business. I’m your host, Darius Teter, the executive director of Stanford Seed.

Consider the first airplane invented by the Wright brothers. It was a groundbreaking achievement that took the dream of flight and made it a reality. Now imagine replacing that tiny propeller engine with the jet. Suddenly we’re not just flying, we’re transporting millions of people around the world at incredible speeds. It’s not a perfect analogy, but this leap in technology parallels the integration of AI and entrepreneurship. Just as jet engines revolutionized aviation, AI has the potential to transform how businesses operate, bringing new efficiencies and opportunities. We recently hosted a master class with Ethan Mollick, a professor of entrepreneurship at Wharton. Mollick is at the forefront of thinking about how AI weaves into business, education, and society. Today’s episode is about applying the tactics and skills that we discussed with Ethan to your business. We’ll explore how the integration of AI tools may create new opportunities as well as unique challenges for your business and your team. Our guest today is an innovative leader who has embraced these obstacles and opportunities head-on. He’s the founder of two Indian firms, and he’s rethinking his business models by giving AI a seat at the table.

Arun Jagannathan: Hi, my name is Arun Jagannathan, and I’m the co-founder of CrackVerbal, which helps people make smarter career decisions by taking test prep and admission consulting. I also have another startup, Yzerly, which is helping companies communicate better through our training programs.

Darius Teter: Unlike many, Arun isn’t discouraged or dissuaded by the acceleration of AI. In fact, he’s been quick to find ways to integrate these tools into his business processes. And he’s doing all of this while thoughtfully navigating AI’s impact on his business and the world that his students will graduate into — a world that is hard to predict.

Arun Jagannathan: So I was in the technology sector, so I spent about 10 years, and that’s because I grew up in the nineties in India. I ended up being a project manager soon enough, and I could understand the nuts and bolts of technology, but my heart really wasn’t training. So I made that shift. And this was before startups were sexy, so this is 2011. But there was a part of me that was a trainer, that was a mentor. I liked coaching people. So I moonlighted while I was still in tech, and then I got into teaching and training, and that’s how I started off with test prep. And when I got into test prep, what I realized is that there is a larger problem with regards to careers, right? People are getting more and more confused about what’s next. And I think that is also a very interesting problem that we solve.

GMAT prep and MBA admissions are two products we have, but the larger problem is more around career services.

Darius Teter: Okay. That’s an interesting challenge, right? Because teaching for the GMAT in a classroom, physical classroom, then going online, then expanding your offering, it sounds like going online has actually enabled you to think more broadly about what services and products you provide and what market segment you reach. Is that a true statement?

Arun Jagannathan: That’s right. So I knew that there had to be a component that went online, and in March 2020 COVID happened, and I had a plan that week, so that’s when I really got executing. So now what we deliver is virtual. So that’s been a very big shift in our business model. I kind of took a step back and I realized that a large part of what I’m doing, whether it’s helping people write their application or whether it’s helping people write an email in a business context, is about words, and generative AI can actually solve that problem.

So we came up with a lot of use cases where we are currently using AI. For example, we help people write their application essays. Stanford GSB has the famous question: What matters most to you and why? It can be a very profound essay that you need to think through. So earlier, a bad essay would be a three on 10, but today with GPT, anybody can write a six or a seven on 10. And I think that’s where we are bringing our values, saying: How do we help people find their voice? How do we help them articulate that thought in a way that’s unique to them? How do they tell their story? So we are also able to identify using GPT where necessary because now the applicant doesn’t need to stare at a blank Word document. That day is gone, right? He has a decent first draft that AI can do and at least it helps them structure, and we give them a lot of frameworks. So they apply the frameworks and they’re able to get decent. But then “how do we elevate it” is where we end up adding our value.

Darius Teter: When I sat down with Ethan Mollick, it became clear why so many are interested in his opinion. That’s because his blogs and his latest book, Co-Intelligence: Living and Working with AI, explore the symbiotic relationship between humans and artificial intelligence. Ethan’s experience, deep research, and unending curiosity make him a fitting guide for understanding the profound changes that AI brings to business — practically, systematically, and creatively.

Ethan Mollick: I have four questions I ask organizations in the face of AI. The first question is: What special thing you did is now no longer important? So we have to be clear-eyed about that. The second question, though, is: What impossible thing can you now do? You said everyone’s a coder inside these organizations. Great, so now what can we do at scale with these things? What change can we make? The third question is: What can you move down market or democratize? There was something where the costs were prohibiting you from going after a set of customers? Now the costs are decreasing. How do you go after a wider set of customers? And the fourth is: What can you upmarket or personalize? Now you can provide personal emails to everybody, additional content. What can I offer that’s special or new? That goes upmarket as well. And I think if you think about those sets of ideas, you end up in pretty good shape.

Darius Teter: So Arun and I talked about that first question. I wondered: What are some of those things that people no longer need to do because AI can do them for you?

Arun Jagannathan: So the other thing that’s come out is a whole cottage industry of tools that are built on what is largely the OpenAI OS. And these are tools. For example, we have Gamma, which is a tool that we use for documents because there is a lot of document management and it does, as I said, the good first version. So instead of going to ChatGPT and getting back, we have it within that tool. Something as simple as we have a daily standup call and we have a bot that summarizes the call and pushes it on to the Slack channel. We revamped our website using aurelium.io and that flow, and we wrote the copy using ChatGPT and all of it within a month. We got the website up and running. And the whole idea is we feel that it’s a lot easier to do things than maybe a year ago when I would get a designer a PSD, give it to the WordPress developer, and it’s a fairly long journey. So I think there is also a speed to execution that AI allows.

Darius Teter: Moving on to the second question that Ethan posed, I asked Arun: What impossible thing can you now do?

Arun Jagannathan: We have now become a lot more productive in what we are able to do. Like I said, we had to transition from getting this ed tech to the whole personalization bit. And in between, GMAT changed the format. So creating a replica of a GMAT question is very tough because GMAT really has, that’s their IP. But now AI does a good job of simulating that question, testing the underlying patterns, and you still need to oversee what is written, but it does a good thing. And at end of the day, you can’t patent English or math. So we are able to create more questions, which is one example of how —

Darius Teter: Wait, wait. So let me make sure I understand. So the GMAT, you’re not having them practice on old GMAT tests or — because like I said, in my era, you buy a book full of GMAT questions that are presumably — they’re old ones. You’re actually creating questions that are like a GMAT question, but they’re completely, uniquely yours. That’s — you’ve designed them. That’s interesting.

Arun Jagannathan: And we do give them the official questions. This is supplementary practice because if we were to create our own questions, it depends on the quality of the question creator, but over here we can model it where we are not infringing on any copyright, but at the same time able to help our students with the right kind of practice.

Darius Teter: Interesting. Okay. So that’s one really powerful use and impact in creating your new product. Next, let’s consider Ethan’s third question. What can you move down market or democratize? AI often reduces costs and that makes it possible to reach a broader audience.

Arun Jagannathan: So, for example, just on the point that you mentioned, the other business, Yzerly over there, for example, email writing. So we had a client, which is a large bank, and they had about 20,000 emails that we had to evaluate. We literally wrote code overnight where it could pick the emails from a spreadsheet, do an evaluation according to a rubric, and give the output into another spreadsheet. So we were able to accomplish this, which otherwise would have taken a lot more time if someone were to do it manually.

Darius Teter: Finally, can AI help you go upmarket by personalizing your services to potential customers at scale? Arun is already exploiting this opportunity.

Arun Jagannathan: I’ll give one more example of a use case.

Darius Teter: Please.

Arun Jagannathan: So we have a profile analysis tool. So what it does is, if you upload your resume, it picks up the relevant points, and the backend picks up from the open AI APIs. We spit out a report and the report is based on a prompt that we have given. So now this is a very simple use case, but what happens is now when someone comes to the website, there is, (a), an engagement tool. Number two, when the report gets created, the lead feels that there is something of value that he has gotten. And what it also does is for my sales team, we already have a very clear profile and we know how to pitch our offerings based on what the report has generated.

Darius Teter: This is really interesting. So it’s a free service. You go to their website, say, “Hey, upload your resume and we’ll analyze it for you and give you some suggestions and recommendations.” And that’s a way to capture a potential lead. Now you have their email address, you know a lot about them. So tell me: What’s the next step for how that touches your sales team?

Arun Jagannathan: So what happens is, at this point, when the sales team reaches out to a prospect, there are a couple of things that come into play. One is there is a custom report that we have already provided, which means that we have understood the problem, the pain point, and we are able to offer something of value. So there is an automatic reciprocity that builds when you are reaching out in a way that is beneficial to them. Because what happens with the services product is there are ways to customize it. So we look at the pain point, we look at what the student’s requirements are, and based on that we give a solution. So it helps us be more specific in the call. If I were to get a new salesperson, I would need to train him to get that kind of experience. But now with the report, and we also have a bot which the salespeople can use to get frequently answered questions about the product. So that way they are equipped to answer any questions during the call that’s in front of them.

Darius Teter: I see. The sales team has their own internal bot that they can use in the middle of a sales call so that they give the right answers on everything. Well, Ethan emphasized that AI can handle a vast and growing array of tasks. For him, the human element remains indispensable. It’s what he called being the human in the loop. It’s about leveraging AI for its strengths while ensuring that human oversight, domain expertise, and creativity still guide the final outcomes.

Ethan Mollick: You might still be much more innovative than the AI. Maybe you’re the top 5 percent of innovative people, 10 percent. AI is offered at the 8th percentile right now, but you should probably be using it to help you generate ideas. And so I think that that’s a fairly profound change. Right now your job can change, but if your system doesn’t change, you’re a doctor and AI does your job really well, does that mean patients would work with your job? Who’s training residents? How will insurance companies … how will you work with the nurses in the room? There’s a lot of open questions. Our jobs are many different things. So I don’t try and bet on job changes too much right now because I don’t know where the world is heading. I think you do what you’re good at and what you love. It’s like I talked to Hollywood writers and I’m like, well, it’s a long way off from replacing you as a .001 percent best writer in the world.

Arun Jagannathan: I think what AI does is, it makes “good” very easy, but “great” is still very hard. So that’s what’s going to start mattering more. In fact, what you say is true. So what is happening is that a lot of tools are getting to a point of advancement where the only reason that you exist is because you bring something to the table. So it could be true for programmers, it’ll be true for content creators, because, as you said, even now I see a lot of my feed on LinkedIn. I can see a ChatGPT post, I know the words, right? Navigate and some giveaways, and I know that this person has not put an effort into making that post, and that reduces my interest to read.

Darius Teter: This is a takeaway I might just pin to my wall. What AI does is, it makes “good” very easy, but “great” is still very hard. There are moments, many moments, where human touch makes all the difference. In education, for example, personalization often comes from human interaction. It’s a balance. And Arun has found that it’s crucial to combine AI’s efficiency with the empathy and support that only humans can provide, at least for now.

Arun Jagannathan: So I think really the idea is, in education, the focus is on personalization and how does that personalization come from a human. For example, if I’m teaching someone to communicate, it’s very different when I’m by myself and when I’m in front of someone else. So having that human does make a difference. For example, in test prep, we call students before their GMAT exam just to wish them luck. And sometimes it’s just having that talk and pep talk and making them feel good, which is where I think humans really add value. So how do you get humans to do what they are best at and leave the rest to AI?

Darius Teter: You’ve described a number of tools that you’ve incorporated, both in terms of your sales process, your own internal efficiency. How have your employees responded to all this? Do they feel threatened? Are they excited? Do they think their jobs are going to disappear in a year? How does this conversation go inside your company?

Arun Jagannathan: Right, great question. I think what really happened was, I had to disrupt myself. So when you look at AI, the question is not to ask, okay, how do I optimize my process? The question to ask is: How can AI just do this end result? How can I achieve that through AI? So when you are doing that, you go through a process of micro experimentations, you’re constantly trying to figure out a new tool, and none of whatever I’m saying could have happened had it not been for the team. So many of the ideas for which I may take credit have come from the team because when you unleash these tools and you give them the power to think, they come up with use cases that you may not have thought of. So I think that’s where the real beauty is, to think of looking at AI as a cost optimization tool. It can assist a human. The way I look at it is, it’s an exoskeleton like Iron Man coming in. You can just lift more weight, you can be 10 X more productive. So that’s how my team has taken up. So today I could say everyone in my team is a technology person. Why? Because they know a programming language called English, and that’s what you really need.

Darius Teter: Personally, I love this approach, but Arun may be ahead of the curve. Not all businesses are as quick to embrace experimentation. If you ask Ethan, this willingness to experiment, to learn to adapt, those are the keys for reaping the full potential of AI. Ethan highlights the importance of company culture and the challenges that employees face when integrating AI into their own workflows. Let’s hear his thoughts on how organizations can navigate these complexities and foster a culture of trust and innovation.

Ethan Mollick: Best practices there. So I mean, that’s a hard question. It’s an interesting one. There’s a lot of dimensions to that question. The first is: What’s the attitude inside your company? People don’t show you they’re using these tools. They secretly use them all the time. They don’t show you they’re using them because they’re worried that — what if they get punished? What if they get fired? What if you realize that their cool work was being done by AI? What if you don’t value the AI work and just take it from them? What if you realize that you don’t need as many people and you fire them or their colleagues? What if you steal their ideas and don’t give them compensation for it? There’s a thousand reasons why people don’t show their work. So it begins with having a good — I’m sure something you talk about is that a good company culture gets you part of the way there to start off with, right? That’s the beginning part. And then it’s the thing about the incentive reward system around exactly these kinds of tools. But the goal is to help human thriving here, not as a … “How do we give the tedious stuff to AI and then expand what we can do” should be the answer. And organizations that outsource work to AI and then fire people are going to suffer in a world of experimentation and change. We’ll see what happens. If you don’t trust your people, you’re in trouble.

Arun Jagannathan: When I made that shift from a classroom to an online, I think I also had to change the DNA of my company, so to be more innovative, and we realized that for us to have the culture, we need to have the rituals and the routines and the stories internally for us to encourage those things. So we started enabling people. So we have a ChatGPT license that we give out to employees. We also have tools we buy. So we have a certain innovation budget where we keep — because it’s very hard to justify ROI at a point where I don’t know what I don’t know, but I do know that, okay, if I don’t put this percentage for experimentation, I have a lot more to lose going ahead. So I think what one example of how people could tend to resist was, we have someone who’s a counselor and a large part of the counselor’s work is dependent on his or her knowledge of the application process.

So when we first demo’ed ChatGPT, I threw up all the questions that you could possibly have a student asking. And her first reaction when she saw it was, well, this is my job. This is what I do for a living, and it’s gone. And what I told her is, that’s exactly what you should think. If this can be done by AI, what is it that you bring? And we realize that what she brings is a certain joy to work when she talks to students, when she motivates them, when she has to remind them before a deadline. So to decouple the human, the person that you bring to the knowledge part of things, is a very empowering thing for people because we live in this fear that, okay, maybe my job is because of something I know. And really to decouple that. Look, knowledge is no longer the differentiation. It’s something else that you bring to the table. And I think one more thing that I said is, if you look at it even as a CEO, a lot of what I do, decision making, I’m pretty sure AI can do a better job, but then I hope that when I come and speak, I’m able to get the team together to get to the goal. So I think the human is there, the value really stands.

Ethan Mollick: I think the first thing is getting people access to a frontier model, and then it’s about setting up education, reasonable policies, and incentive structures. So you have to do it organizationally. But I think the first thing I tell CEOs to do is just play with the system enough. They need to put their 10 hours in. They cannot depend on direct reports. They also can’t expect hiring a consultant will solve all their problems. It used to be that like, okay, McKinsey, Ernst & Young, they knew everything. They don’t know anything anymore. Nobody has a playbook, right? They might be able to help you with transitions and other sets of stuff, but they don’t know how to use AI for your systems. So you have to build the systems to make that all happen.

Darius Teter: As you’ve thought about your journey in adopting generative AI into your business, what advice would you give?

Arun Jagannathan: I heard this very interesting term, which is called the Amara’s Law, which is the impact of technology. In the short term, we tend to overestimate it, and in the long term we tend to underestimate it. I think we are in the journey of getting to AGI at some point. AI is going to disrupt fundamentally what we are doing. But I think at this point, what we do have in our control is just to see what is it that we can adopt as entrepreneurs, and the Grit & Growth podcast talks about internet entrepreneurs. So I think we have to respond to situations, to challenges, to competition, and to look at AI as a very important impact. So it’s coming your way, and the sooner you are able to react, the better. So I think one thing that I’ve noticed is, as I said, many employees today are asking the management and the companies: What is our AI strategy?

Because nobody’s in a bubble. Everybody’s hearing this, right? And they know that if we are on a growth path, on a growth trajectory, then AI has to be a part of the strategy to be open about it, to talk about the fact that, look, we may not necessarily know right now what we are doing with it, but the idea is we want to be an adopter. So having that mindset and telling people and asking them to contribute, especially if you have people in their twenties. So we have younger folks, and the kind of innovation they are able to come up with is incredible. And I think the advantage that Seed companies have is in the size. So we are smaller, we are leaner, we are hungrier, we can move faster. So I think that’s definitely something that works in our advantage.

Darius Teter: Ethan Mollick posts some key questions to help entrepreneurs think about how AI can turbocharge their business. Sometimes that means automating routine tasks, doing the crap that you don’t want to do anymore. Sometimes it means augmenting your own ideas to help you be more creative, to reach new customers, and ultimately to grow your business. Arun is making these concepts real as he transforms CrackVerbal and Yzerly by doing things that would not have been possible even 12 months ago and welcoming his whole team into that experiment. I think Arun’s story exemplifies the possibilities of AI for small and medium enterprises. And remember, the big companies have no advantage here. There is no playbook, there is no manual. Each of you need to figure this out for your own context and your own business through experimentation.

I’d like to thank Arun for sharing his story and Ethan for his deep insights. And if you want to hear more from Ethan, check out my interview with him in episode four of Grit & Growth.

Erika Amoako-Agyei and VeAnne Virgin researched and developed content for this episode. Kendra Gladych is our production coordinator, and our executive producer is Tiffany Steeves, with writing and production from Nathan Tower and sound design and mixing by Ben Crannell at Lower Street Media. I’m Darius Teter. This has been Grit & Growth. Thank you for joining us.