Transforming Customer Service with AI: Landbot's CEO Jiaqi Pan Shares Insights and Strategies

Auto-generated transcript. Speaker names, spelling, and punctuation may be slightly off.

Mark Evans: Welcome to the Marketing Spark podcast. Today, I'm excited to talk with Jackie Pan, the cofounder and CEO of Landbot, an AI powered platform that helps businesses create and deploy conversational chatbots. As AI continues to have a huge impact on a growing number of industries, we're seeing a significant transformation in customer service with AI chatbots leading the charge. Today, Jackie and I will explore how these advanced chatbots reshape customer interactions, making support more efficient and personalized. And Landbot has carved out a distinctive position in this competitive landscape, leveraging innovative technology to deliver exceptional customer experience. Join us as we explore the evolving world of AI driven customer service and discover how is leading this exciting change. Before we get started, I should mention that Landbot sponsors the podcast. Welcome to Marketing Spark, Jackie.

Guest: Thanks for inviting me, Mark.

Mark Evans: I like to start my podcast with the backstory, the origin story, trying to discover what motivates, what inspires entrepreneurs. What is the story behind the founding of Landbot?

Guest: Our story is because we did a huge pivot when we started. In the beginning, we were actually a b to c consumer service business. Right? We essentially help, other businesses, sorry, other users to get things done online. We are like a personal assistant, but with actual human behind the scene talking to people and then chatting with them. The unique value proposition of, our consumer model was that all the interactions of having having happening over a chat. Right? So it was, like, quite conversational, and people can, right away, go to us, come to us, and then chat and and getting, their request done. Right? People are coming to us to say, hey. Walk me this trip, or I need to do some errands. Can you help me to do that? I need to buy stuff. Can you go to search it and buy it for me? And with that, we started just having a lot of customer coming to ask any sort of things. And we started building a lot of technology, usually just to help us to improve our own customer operations. But over time, we start to see that the technology we have built was so powerful that it's not only interesting for us to use it, but can also potentially help other businesses in their way of communicating with customers. And we also realized that the business model for the personal assistant service was not very scalable because we end up having to hire a lot of human agent and having lot of people to just assist all our customers and having those chats. Obviously, we had lot of automation already in place, but human was always needed at some point.

Mark Evans: What was the point in which you realized that this model, this business that you were excited about and had grown wasn't going to work either from a financial perspective or in terms of scaling the business? Was there a was there a straw that broke the camel's back? Was there a moment in time when the senior executive team said, we're excited about what we're doing. We've got a lot of users, but this is not a business. This is not a a business that's gonna be successful financially.

Guest: Quite interesting question. It was basically when I was starting to prepare for a new round, new financing rounds, and start gathering a little bit the data on how the business was performing. Right? And I I saw a a graph, which was, like, mainly the, the thing that made us feel it. Our the graph was basically a combination of how our cost was evolving and how our income or revenue was growing. Right? And there was basically a graph of a divergence where the cost is growing, essentially exponentially. The revenue is growing, like, more linearly. And that was what makes me understanding that in order for that business to scale, we either have to raise, like, 100 millions of revenue dollars to finance the whole operation, or it was impossible to to grow it, and which is why we end up pivoting. That's what become Lambo today. So at that time, I'm talking about 2017, maybe, chatbot was suddenly becoming big, right, because Mark Zuckerberg was announced that that they want to revolutionize the world with chatbots. That was, like, when chatbot become popular for the first time or when the term chatbot was first used. And we had, obviously, a really good fit with that because we built this, amazing customer communication technology and had a lot of automation already in that. So we rapidly positioned us as a chatbot solution. But unlike most of the solution that was available in the market, we went with a very intuitive and ease of use approach, which at that time, nobody voted the term no code, but not even a thing yet. Right? But that was our main value proposition. So we wanted to help essentially business users, nontechnical users to able to have the capability to build their own bots, to automate their customer journey without relying on those technical users.

Mark Evans: As an entrepreneur, failure, whether it's technical failure or financial failure, is often seen as a negative. In this case, you built a business, you grew the business, but recognize that the model wasn't gonna work for a variety of reasons. Looking back at your experience and the journey that you've gone through, kind of advice would you give to entrepreneurs who are looking to start their own B2B SaaS companies? Many entrepreneurs have a problem, discover a problem, recognize an opportunity, and there's a lot of enthusiasm because these days relatively speaking, it's easy to build a product. Tough part is obviously attracting customers and establishing a business. So when you look back at your own experience and some of the ideas and hypothesis that you had for the original business, what lessons did you learn? And what are some of the biggest pieces of advice that you would give to entrepreneurs looking to jump into the B2B SaaS world?

Guest: My advice would be applicable to both B2B SaaS or any other type of businesses. And I think you mentioned one of the key piece, which is identify the problem. Right? That's the first step. I think the problem many people or many entrepreneurs has is that they are starting with the solution in mind. They, they fell in love with an idea. Right? Hey. I want to build this solution because I find it cool or it's interesting. Right? And many times, they suddenly realize that's a solution, but, they don't have a clear market. They don't even know who they can sell that solution to. Then you fell into the trap of having a solution in search for a market, which is usually the best, the poor way to approach it building a business. So the first advice is trying to have a very clear idea who is your target customer. Right? What is the pain they have? And very importantly, what makes your solution be better? Because if we are talking about the b to b SaaS space, truth be told, there is thousands, if not tens of thousands of SaaS solutions available today, and many of them excellent. Right? Very good solutions. How can you convince your potential customers to choose your solution instead of whatever is the status quo and what is already on the market? Right? I'm not saying that you have to be 10x better. That's bullshit. But at least you need to have some sort of angle to go to those potential customers and tell them, hey, use me because of x y z.

Mark Evans: It's interesting that you mentioned brand positioning because obviously and we'll talk about this later in in the conversation because obviously, it's important to differentiate yourself from the dozens, if not hundreds of companies that walk and talk just like you. I did wanna ask you about another pillar when looking to launch a B2B SaaS company. And that is, yes, solve a problem that people want to tackle. But the other thing is that you have to discover whether people want to pay for your product. And when you look at your own journey, how did you get a better understanding of the fact that what you were building people would actually pay for? Was that were people were customers telling you that they would pay for it? Was it an idea that you hoped would work? What were some of the signals that you got?

Guest: I'll tell you what we did, and then I tell I will tell you also what we should have done. What we did is we just look at the market, what are some of the comparable product, and then we try to put a pricing very similar to them. It's, just 30% cheaper. That's what we did. Okay? But I think what we should have done, right, and that works relatively well in the sense of we immediately start getting people paying us, and there was relatively low friction, so to speak. But I think what we should have done is doing a proper customer discovery and then understanding what is really the value of our solution representing for the customer. Like, for all SaaS products, especially in the b to b space, right, you are solving a business problem. And that business problem usually has certain value, either in terms of cost saving, right, what is the amount of cost you are saving for that business, or revenue increase, like how much additional revenue can you help the business to generate. And once you find that value, you can identify it, Right? Then the price comes to what part of that value can you charge. And usually, the recommendation is either that's a tenth of the cost, right, or tenth of the revenue you can create so people can get 10 x the the ROI, so to speak. Some other times, maybe it's not that much. You can get also five x, which is also great, but that's usually the better way to approach pricing conversations because, otherwise, you will fail into the pricing war or who is end up being the cheapest price of the cheapest solution in the market. And you always have people coming to you saying you are expensive. Right? Because you don't have an anchor or you don't have a reference of what is the value of the what solution is providing.

Mark Evans: I suspect that you've been working with AI for a while. And one of the realities of AI is it's not new. It's a technology that has been around arguably for fifty years, and many companies have embraced it long before it became the best thing since sliced bread. In your experience, how has AI technology evolved since you first launched the company that was the predecessor to Landbot Landbot itself? And how has it influenced your product development?

Guest: Yeah. I would say it has evolved drastically, and it's it has going through different stages. Like, when we first launched Landbot and we made that pivot, the the AI technology that was more important for chatbot was natural language processing, NLP. And that was a technology that was trying to identify the pattern of how people are communicating, what are the some of the keyword they are using, and then try to provide some sort of a categorization to say, hey. The intent of the user is they want to know about your price. The intent of the user is that they want to know what is the integration you offer, whatever possible categorization you can give, and then you try to provide certain type of structure response to them back. Right? That's how the previous technology works. Proline of that technology is that first, it's very hard to map out all the intents because people can say the same thing in a millions way different in a million different ways, and, which makes it very easy to fail. Right? And failing means the bot doesn't understand what the user is saying. And one of the typical jokes we always have in the industry is, like, you have a bot coming, and then the people say, how can I help you? Then the users say, I only speak Spanish, please. And the the bot is using that as the what the user is needing, and then they say, hey, Spanish, please. How can I further help you? Whatever. So it's basically misunderstanding whatever the user is saying all the way. Then what the next evolution came is the large language models, which is really the technology behind JGPT. Right? With large language models, the benefit you have is that you don't you no longer need to map out any intent. Right? Because it has such a large dataset behind all the models that it understands every intent in the market, but it can have hallucination, which is the problem with this new technology. Meaning, although it understand all the intent or whatever people is saying, it can sometimes misinterpret or saying things that is just not true. And therefore, you have some sort of a imprecision in the way how the bot respond back to you. And then depending on the use case, this can be more or less, damaging to the business. Right? So what we are seeing right now is that a lot of businesses, they are trying to have a hybrid approach, especially when the businesses have certain scale and has, you know, certain volume of customer interaction where they cannot permit or they cannot allow the bot hallucinating and saying some false, information to the customers. So for very sensitive topics, typically pricing related product information, things like that, they will try to have the bot using, what we call the NLP part or, like, the standard linear flows. And then for information that is not a critical, right, they can have the bot using the large language mouse to answer it, and which provide more flexibility and better user experience because it always understand what the user is saying.

Mark Evans: Couple of questions from that, explanation. How do you address concerns about AI and automation potentially reducing the need for human customer service agents? And the second would be, what strategies do you recommend for companies looking to implement AI chatbots without compromising their customer service quality. Because obviously, technology is amazing. A lot of companies definitely want to leverage the power of AI, especially for customer service and the ability to do it at scale in a very productive and cost efficient way. But at the same time, there's many brands that have products that are complex or sensitive. As you say, maybe a hybrid approach is the optimal way to go. The question to you is, obviously, it would be one, what's that balance between human powered customer service and AI powered customer service? What are the strategies that that companies should be exploring when they're ready to take the leap into AI powered chatbots?

Guest: There is a different way to look at this. First, on the risk of AI replacing humans. It's the same risk of every new technology when they are coming out. Right? When cars came out, those horsemans, they are in risk of losing their jobs, right, by replaced with this new technology. Whenever this new technology is being adopted massively in the in the economy, it will open up new opportunities for people. Right? And that's the way we are looking at it. We don't know yet what are the potential ways this technology can help us to unlock new opportunities, but it will, for sure, having more opportunities and then allowing people to do probably more interesting type of work. Right? Because, frankly, a lot of human agent, the type of task they are doing today is not very satisfying or motivating. They're just answering very repetitive questions people are coming to them or filling out the CRM or their contact center solutions, which is not really motivating to anyone. The AI is right now really focused on replacing some of those low value tasks and then enable those human agents to focus their time and energy on things that will really add value to customers and also be more motivating for them. In our own case, just talking about our customer service teams. Right? We have a team of 10 people in our customer service team. We are encouraging them actively for ways to automate their work so that they can slowly position themselves as a trusted adviser and focusing on more strategic conversation with customers instead of just focusing on low value repetitive questions.

Mark Evans: It does beg the question of how a company measures the success of AI driven customer service solutions and AI powered chatbots. In your view, how do they quantify the ROI? Let me back up here. In your view, what are the metrics, the signals, the outcomes that are the most important? And then how do they measure the success and the ROI of implementing an AI powered chatbot to support or drive their customer service?

Guest: Great question. Here, we have to separate between input and output metrics. For the output metrics, it's usually what is the main business use, the main kind of value the business is trying to drive. Right? It it can be when it's very focused on customer service and customer support, how much ticket the AI has able to handle without having human assistant or without having to getting humans involved. Therefore, how much time has it helped to save humans? Then, obviously, you try to complement that with some sort of leading indicators or input metrics like, hey. What is the satisfaction level on on the different conversations the AI is having? Right? What are the amount of abandonment people are having in the conversation? Meaning, like, people who start having a conversation without having it completed, they start chatting with that. Mhmm. Right? The the higher the abandonment, the poor the experience is, right, in theory. So there's different ways to measure that. Then if the use case is very focused on sales and marketing, which, is a way also you can then see how many leads, right, the bot is able to capture or to convert. Right? What is the ultimate, also, a conversion rate, of those leads to become to opportunities and to deal close? Right? And then input metrics can look at how many people are engaging with the bots and if that is growing or decreasing, right, and how does that influencing your pipeline. You can also look at from the quantitative perspective, which is really important in the beginning when you don't have high volume. Go randomly just choose five conversations and see what are the conversation that is happening between the bot and the user. Are they seeming to ask relevant questions that bot can answer? Or, well, they are asking just a to asking the bot to to tell some jokes or totally unrelevant, questions. Those are the inputs that you need to then slowly iterate your system. Right? You mentioned what is the ideal ratio or balance between AI based interaction versus human based interaction. Mhmm. Really depends on the journey and the type of customers you are facing. Let me give you an example. Right? In our own case, we are actively using AI bots across our entire customer journey. Right? And in the very initial stage where we are a business quite global and we have the customers all around the globe. Right? The problem we have is that sometimes we can have low quality people or low quality leads, so to speak, coming from some low income regions and that we know have very low conversion rates. And probably those are the people we don't never want to talk to your human agent because it will just waste your time and waste your businesses. Right? We ask down the line on the end end side of the customer journey where we have customers already converted and actively using Lambo and paying us, we have these managed account type of segment where they are the best customer category we have and they are paying the most to us. Right? Those type of interactions are probably predominantly human driven because we don't want to let the customer lose the relationship or lose trust in us when when we want to be very close to them. So it really depends on where in the journey you have and who is the type of customer and what is the unit economics you have from those customers that allows you to have more humans or less humans.

Mark Evans: Mhmm. Mhmm. I wanted to circle back on a topic that we touched upon earlier. Something that I spend a lot of time focused on. It's actually the the first place that I start with all my b to b SaaS clients and that is brand positioning and the not the ability, but the necessity to differentiate your product in ultra competitive marketplaces. To be fair, every B2B SaaS vertical is super competitive, including AI powered chatbots. Wanted to get some insight from you into the role that brand positioning has played in Landbot's business and your ability to attract customers. And was it something that was top of mind? Was it something that you tackled proactively? Was it something that simply evolved? Can you give me some insight into the approach that you took and some advice to other companies that may or may not recognize the importance of brand positioning as a asset or tool that underpins not only marketing, but sales, product development, HR, raising capital, and the list goes on.

Guest: First, I would say the first recommendation I would give people is follow someone who I admire for positioning type of work, especially in the b to b space. She is April Dunford. Okay? So if people are not familiar with her work, she's a positioning expert. Right? She has this amazing framework that allows you to understand. We did a workshop with her to help us to define our positioning. And then, how we are approaching it, it's all first coming from a, like, a very customer centric way. Meaning, like, you the first thing you try to do is with your existing customer base. Right? It's, probably a different, way to approach it when you are just starting a note, right, and and you don't have any customers. But assuming you already have some customers, even if it's just 10, whatever the number is, you go to talk to your best customers. Right? The the top customers you have in your, you know, existing base, to understand who they are, what industry they are in, what sub industry or sub segment they are in. Right? The more niche and the more tailored you can go into, the better. Right? What is really the thing that they are struggling with and they are using your solution solve? What are the key differentiation those customers see in your solution that makes them perceiving you being different? It's very important what they say, not what you think you are Mhmm. Using the customer own words to come up with those definitions. And then finally is the category theme, which I believe it's usually for smaller players or people who just start out. Pick a existing category. Don't try to reinvent the wheel. Right? If chatbot is a thing already, don't go suddenly to go something like conversational automation of conversational, whatever random stuff you have invented. Just go with chatbot and then find your niche. Find the space where you can be unique and be relevant where you are a target customers. And I think that's a little bit the the way how we have reached it to our current positioning. We don't have it perfect yet. Right? And we are constantly just keep improving and integrating it because the positioning is also dynamic. Meaning, like, market is constantly evolving. New players coming. Right? The technology is changing. Like, two years ago, probably all the AI part is nonrelevant, but now suddenly everybody is talking about AI. So you have to factor that in and and see what how does that change your positioning. I think that's a little bit the framework we use.

Mark Evans: It's important to remember that positioning is not a set it and forget it kind of proposition. I love the fact that you recognize or highlighted the fact that things change. Customers change, technology changes, a hot new competitor appears on the landscape, and they seem to capture the imagination of customers and prospects. And as a result, you're constantly tweaking the dials. You're constantly looking for new ways to position your product in a new and interesting way. So some great insight there. The $64,000 question that I I did wanna ask you, and this is something top of mind Mhmm. For all CEOs these days is how to drive leads and sales. So while brand positioning and strategic planning are important, and they matter as long term propositions. Mhmm. The here and now for most CEOs, most b to b CEOs is how do I generate leads and sales? There's a balancing act between short term needs and long term Mhmm. Propositions. When you look at your own marketing and sales strategies right now

Guest: Mhmm.

Mark Evans: To drive the business forward? What are some of the successful things that that the company has done to attract, engage, and close customers in a very competitive landscape? The reality is that if any entrepreneur can provide some insight, I think it's you because AI is hot, chatbots are hot, but at the same time, budgets are tight, A lot of companies are prioritizing their spending on especially on marketing. Mhmm. So what's working for Landbot? What's keeping the business moving forward? Any secrets that you can reveal? Or is there anything that you're just fundamentally doing day in, day out that you continue to lean into?

Guest: I will share more like the mistakes and the things we haven't done so right because to be honest, I don't think we have figured it out. I will share what are the things we realized that doesn't work and then maybe how we are changing it. First thing is not all leads are, being equal. And in the beginning, we felt, hey, we need more leads. We need to just having more people. And we have this product like growth motion where most of our customers, like 80% of our customers, are just being self serve. They come to our system. They sign up. They can try it out by themselves, and they can even buy just using credit card. And on a monthly basis, we are generating, like, 10,000 sign ups. Right? 10,000 all sort of people trying out our solution, which, you might think it's a great volume of people than great volume of leads. And we have been thinking that for some time as well, but the problem is, like, our conversion rate is really low. We tend to get, like, maybe a 10 sorry, 1% or, like, sometimes 2% of conversion rate out of those sign ups. And we always struggle, hey. What's happening? Why are we not converting people more? And one thing we just realized very recently is that because of the buying intent of people. Right? One question we are starting to ask people in the onboarding, form and when they sign up is that what is the reason they are using Lambo four? And we soon saw that the only 30% of our customer base are trying to use it for business purposes. The other 60% is either for personal use, like side project, things like that, or in additional purpose. They're student, professors, university, project, whatever, which are people who will never convert. So we are actually only able to convert, like, a third of our overall pipeline, which is great, but that pipeline has not grown if we see, like, the historical performance, which is why we need to be very careful about our acquisition strategies. One thing we are not doing well in the past is optimizing for low cost. Meaning, we were most most of our marketing strategies is all around how can we, you know, doing campaigns that are cheaper and bring more volume, which kind of a could work for some categories, but I think in our case, it definitely is the wrong strategy because in the end, it brought us people who are just looking for free Tradjibbit alternatives or Right. Who are looking for random bots that can make jokes and will not be very serious about using it for their business. That's the problem with a lot of the lead generation strategies business has. Right? They end up optimizing for the low lower cost and lower quality of the leads. It's much better doing the opposite, which is what we are trying to do. Like, we know, like, when we acquire a high quality, high buying trend business, this is really thinking clearly what is the business challenge they have and how they want to solve it with above, the the conversion rate is going through the roof and the retain and all that's the unique economics tend to work really well. Okay? Then what are the different ways how we are trying to acquire customers? Right? One of the things that we have invested long time ago is content. Right? Content marketing. Mhmm. To be honest, lately, it has a lot of noise. A lot of people say with all these AI generated content, SEO will die. It's been two years, and content has still growing nicely for us and keep growing, keep driving volume. So I don't really see the content marketing going away, but you have to be very strategic about it. Right? You have to understand what is the type of keywords because in the end, you you are optimizing for the search engine. Right? Mhmm. How are really the keywords that will why it help that would bring the type of leads that have high volume. We don't want to just optimize for high volume keywords, but in the end, the quality is very low. Right? And the only way to do that, I will relate back to the decision part, is doing customer discovery and customer interviews. Right? And one mistake we did, and I personally regret it quite a lot, is not doing that personally myself as the CEO or after that.

Mark Evans: Right? Interesting.

Guest: Delegating that to the team, like, the the customer facing team, maybe our customer success team or our sales team. And then having them bringing up the insight from the customer discussions and then us as the kind of founders or leadership, we try to extract conclusions to save time. Right? But the reality I saw is that there's so much nuances that is missing in the way when you are not hearing directly from the customer. And you as the founder, you have much wider perspective of what is the problem that you can solve for the customer and what is really the words they are using because those end up making huge difference. Right? One advice, if I can only give one advice to people, is do not skip this. Right? It's hard work. It's boring. It's not giving you immediate results, but it's necessary for your business to work.

Mark Evans: Yeah. Yeah. Patience is a virtue. A lot of companies and CEOs need to realize that marketing is a long term proposition and that it's a slow burn as opposed to a sprint. This has been great. Covered a lot of ground, talked about AI, talked about content marketing, talked about brand positioning, talked about chatbots. Where can people learn more about you and Landbot?

Guest: People can always, follow me on LinkedIn, and then they can also visit our website, landbot.io. Very, very appreciate you doing this session with me.

Mark Evans: Thanks, Jackie, and thanks to everyone for listening to another episode of Marketing Spark. If you enjoyed the conversation, rate it. Subscribe via Apple Podcast, Spotify, or your favorite podcast app, and share via social media. If you're a b to b SaaS company with revenue of 1,000,000 to $10,000,000 and you're looking for strategic and tactical support to scale, we should talk about how I can help you do marketing that drives outcomes, whether you're looking for better brand positioning, leads, sales, brand awareness, or simply getting key projects completed. You can reach out to me via email, Mark@MarkEvans.ca. Connect with me on LinkedIn or visit marketingspark.co. I'll talk to you soon.