Interview with Firmly CEO Kumar Senthil on Solving the Integration Tax and the Journey to Zero-Code Merchant Connectivity

Speaker A: Welcome to the Retail Tech Podcast. My name is Darius Wasseffi and today I am speaking with Kumar Senthil, founder and CEO of a company called Firmly.

Speaker B: Hi Kumar, nice to meet you.

Speaker A: Likewise.

Speaker B: Thanks for having me.

Speaker A: Thank you for uh, joining me on the podcast. Let's start with a uh, brief introduction about yourself, your background and how did you come to start, uh, firmly.

Speaker B: Sure, definitely. So, uh, you know, uh, before starting Firmly, I used to uh, manage the samsung.com. uh, and uh, one of the common challenges I was running into was a lot of partners would come and pitch to me saying hey Kumar, I can help sell your products in different places. Uh, but then the moment they ask, okay, great, what do I have to do for that? They'll say, oh please come integrate with me. And I said, how many integrations can I do? I did my fair share of integrations and realized man, this endless integrations is a non starter and more importantly I can do the integration but then I don't know whether the ROI is there or not. So I put a bunch of people work there for a few months and get this integration up and running and I get one or two orders and then that's it. That means, you know, it's, it's really, it becomes a uh, painful situation internally and people question saying okay Kumar, you asked me to integrate with this partner, but guess what happened, right? So then next time I ask the team, they look at me and saying, are you really sure this is going to work? So I've been going through this motion and I said man, I really wanted to try new things but at the same time this tax of integration and maintaining is becoming untenable as a merchant. But then I also thought, you know, it's not going to be a Samsung problem, it's going to be a global problem. Every merchant is going to go through the same situation. How do we solve it globally for all merchants? Uh, that's how I start thinking. But then I said, okay, being in Samsung, I don't think I can solve this problem, so why don't I just start my own? And I left Samsung to start this. And uh, it's been a fun journey. You know, it's just riddled with just, just challenges uh, to simply, you know, because I was trying to create a two sided marketplace, right? One side I needed merchants, the other side it needed platforms or destinations. And uh, we, we were way ahead of its time because uh, when I pitched this to the merchants they just didn't understand why would you ever need this type of tech, right? Uh, and uh, the, the, it required a lot of evangelizing the market and merchants as well as the destinations to saying okay, you really need some sort of a uh, tech which connects to multiple destinations to multiple merchants. Uh, because even if you are a destination, you want to enable native shopping, you won't be able to go merchant by merchant and integrate. And as a merchant you want to go sell in 10 different channels. You don't want to go out and integrate one by one. But, but I needed to bootstrap somewhere. So fortunately we had a partnership. I, I cannot share the name of the company but it was one of the largest uh, social platforms. Uh, we part, we had an opportunity to partner with them and uh, they recruited merchants for us.

Speaker A: Oh that's great.

Speaker B: About 50 merchants. They recruited uh, large merchants that actually bootstrapped things for us. Right. And then that gave us base at least some sort of a merchant spool to start, uh, you know, talking to partners and say hey, I have this pool of merchants we can enable. And that brought interest from different partners. Uh, so then we started you know, building the company from there. And so now, now we're in a growth stage. A lot more merchants are there, a lot more partners are there.

Speaker A: Uh, okay, yeah, okay, that's great. So um, yeah, definitely starting a company is a brand new, it's like having children. You know, it sounds good until you have one and then you really understand what it is.

Speaker B: That is true.

Speaker A: I know a lot of people that have been in startups and have been even number three or four. It's completely different than being number one where all the, the weight of the world is on your shoulders and uh, it's a different experience. I know exactly. I mean I've, I've lived that life myself for many years and even with the marketplace, uh, currently I do understand, uh, and we have, we're building m probably like a three sided marketplace which is like unbelievable. So I, I, we went from uh, being B2B to changing to B2C and that became a part of our, our business and now we have B2B as well. But like Marketplace is probably one of the hardest things businesses you, you can build. So yeah. So um, now the main uh, point of this integration is of course like Zapier I think probably popularized the idea of easily integrating with different products as far as like even functional functionality and productivity software, all kinds of things. And uh, you know, it just makes sense to have the same for commerce. So I think it's uh, it's a, it's a really interesting idea. Um, how you know of course now everybody's talking about agentic. Right. And AI and how AI and AI has changed the business and will continue to change the business. So I want to dive in a little bit m more about that side of your business and get your thoughts on how AI uh is actually changing your business and your customers lives.

Speaker B: Absolutely. So AI ah, you uh, know uh, when we started the company, um, uh, I mean especially when we were onboarding these merchants, this 50 merchants, it would take anywhere between four to six weeks of integration work. Uh even, even when we say integration work they're the key North Star we had two North Stars we had not gone to the number one North Star is merchant should not be writing a single line of code zero change for the merchant. Right. Um because if we ask the merchant to make any code change means our mission is failed. So that's a very high bar we set up saying we cannot ask the merchant to make any code change. Whatever the website they have, they should continue to operate as is. And the second north star is merchant has to be remain at the merchant of record. Okay so these are the two primary you know uh clear uh goals we had. Whatever solutions we built it should meet

Speaker A: those two things and that if we can just dive in a little bit more. Merchant of record specifically means that transactions are handled on their platform. The catalog, everything stays on their platform.

Speaker B: Exactly. Okay, everything. Uh, and then when we do that we are also not doing browser automation because people immediately agent E commerce means immediately they think oh it's uh browser automation. We don't and we clearly knew from day one this browser automation is not going to work because it just breaks the whole commerce flow from natural experience. So it's just not going to work. And it has so many drawbacks. It's for another day. But uh, that being said it used to take about six weeks, four to six weeks because an engineer has to go understand how a website works, identify the right APIs and whatnot and then integrate with the merchant. Now later because we built a library of merchants integrations, hundreds of merchants integrated. Now we, we could use that infrastructure to train AI to go and do autonomously integrate without a human in the loop. So AI is able to do the integration for us now because we have that infrastructure in place if for somebody brand new starting today you have to have that type of a solution in place to you know solve this problem. And ah, and, and also there sort of lot of know how has to be needed to do it. So that's what we did and we built that last five years that basically research ended up leading us to this point where we're able to automate uh, end to end. So now if you give a merchant it takes few hours, even few minutes to a few hours depending on what the merchant infrastructure is to be able to integrate and uh, bring their catalog. And a lot of times most of the merchants don't have a proper catalog in place. So what we do is we take that catalog, pull data from different sources, synthesize it, normalize and then we push it to a different destinations. That's a very big value add for a merchant because for them to do themselves it's, it's an expensive affair. We've automated significantly that part.

Speaker A: Can you explain a little bit more what you mean by most merchants don't have the proper catalog?

Speaker B: So if you take for example uh, what the data needs are uh, for OpenAI, uh in the product catalog feed that all the data elements that the LLM, uh, because they are asking this data for a reason because LLMs need that data to better serve the user query, right? So because the customers weren't coming up with a specific intent, for example I uh, may go to ChatGPT or Gemini and say I'm looking for a um, pen. Okay, I'm m picking a random example that has 5 raw star uh reviews for, for me. Okay so which means now the reviews information has to be there in the M, in the LLM for LLM to respond saying okay, here are the five star rating pens are here. Then I can choose if I want to buy or not to buy. That's second but the discovery itself will fail if the rating and review information is not pumped into the, into the LLM. Now uh, that is where we come in. This is just one piece of information like that. There's different pieces of information that a merchant has. They don't have it in one system. They all come from disparate systems. We pull all of that together, normalize it and then we pump it into the destinations and each destination may require it in different formats in different ways. All of that we handle it.

Speaker A: Now when you do this automation, integration, automation, are you focusing on specific platforms? Like once you get Shopify for example set up then all the Shopify merchants would easily be added. Or if you do BigCommerce or Oracle or Salesforce or is that how.

Speaker B: Great question. So we're, we're platform agnostic. So wherever the merchant may be we can work with any, any merchant. We're, we're not tied to any specific people platform. Uh, we can work with any merchant. That's, that's our foundational aspect because we're not arc, uh, is not location specific or uh, you know, platform specific. Uh so therefore we can work with any, any, any merchant. That's the uh, beauty of our platform.

Speaker A: So even a custom um, built commerce

Speaker B: would be, that's our sweet spot. But uh, but you know even if you are a standard platform, we can, we can support that as well. Uh, and, and uh, so we, we make it easy for the merchant. The solution is the same. Zero change for merchant. We ask the merchant to do only two things. Okay. Number one, we ask the merchant to sign our merchant agreement. Okay. So uh, that is justly a paperwork to say that okay, I authorize firmly to access my catalog, access my website legally. Without that we don't do willy nilly things. We want merchant to explicitly accept our agreement. Uh, so that you know, we, we have a permission from the merchant. If the permit merchant doesn't want to give permission then we don't do business with them. Right. So we don't. Right. The second thing is we uh, I mean this, this is not a requirement for all merchants but for some merchant, uh, we need them to whitelist us. Uh, so once they do whitelist us then we're good to go. Then they're off to the raises. They can just sit back and relax and then uh, we just recently launched a platform called uh, Firmly Connect in that what we provide is a menu of options like you know, you go to an app store in the phone or a play store in your Android or iOS phone. You can choose the apps that you want to install in your phone. The same way we are providing an option where you can choose which destination you want to sell to. Okay. There's this catalog of app options we provide for the merchant so that they can say yeah, I want this, I want this, I want this. Great. We enable all of that. Literally no code needed, zero coding needed.

Speaker A: Okay. Wow. Uh, that's great. Yes. Like a, a la carte menu.

Speaker B: Yep. And um, and, and that's where we make it easy for merchants. And already we have about 10, 15 destinations ready to go. So, so there it makes it very attractive for merchants.

Speaker A: Um, so by destinations you mean AI agents.

Speaker B: It could be AI agents, it could be publishers. The beauty is we're not just tied to, we're not just tied to uh, only LLMs. Right. So uh, it could be uh, you know, publishers like you know uh, uh, you know, USA Today, Or Conde Nast or First Media or you know, Zip Davis. Could be anybody.

Speaker A: Okay.

Speaker B: Shopping, native shopping for them. Uh, that's one type of use case. Or it could be CTV or it could be, you know, um, uh, what else? Marketplaces, like furniture.com is a marketplace we partner with. It's a vertical marketplace just focused on furnitures. We onboard, you know, four to five merchants a week into that platform. Uh, so that, you know, they, they built a very robust furniture, uh, marketplace. Uh, you know, you can go play with that. And furniture.com is a, it's a very nice furniture shopping marketplace.

Speaker A: Yeah.

Speaker B: Yeah.

Speaker A: Um, so just to understand the architecture and we're firmly sits in the stack for a merchant. Um, they still have their E Commerce platform. That doesn't change. They have their marketing stack. Everything is the same. Nothing changes.

Speaker B: Yep.

Speaker A: They make this connection or you. And they don't make you make the connection. Right. There is, is there like an endpoint, a uh, series of endpoints that they have to give access to or you give them access to somehow making that connection.

Speaker B: They just whitelist my ip.

Speaker A: Okay.

Speaker B: That's it. Right. So, so that I can access their catalog. I can access other endpoints that are publicly visible and that's it. They don't have to do anything.

Speaker A: Okay.

Speaker B: Uh, we handle all the complexities.

Speaker A: Okay.

Speaker B: And then we expose a, ah, UCP endpoint or ASCP endpoint that's needed for the merchant or if they were, uh, if a partner of ours wanted to talk to us in MCP protocol, we have a global MCP endpoint which can, you can talk to any of our merchants or uh, we also have a standard API. One set of APIs we have which has more capabilities than other protocols as well. Uh, so that if a partner wants to use our API, they are welcome to use that as well. Uh, so that it gives a lot more capabilities. For example, uh, furniture business is the most complex vertical you can think of, uh, in retail. So where it requires delivery scheduling, Holloway or White Glow service and so on, there's so much complexity involved. All of that is handled in our simple APIs. So it's a standardized API, uh, so that our partners can, once they do integration, they can connect to pretty much all of our merchants.

Speaker A: So it seems like you spent, you invested, uh, a good amount of time and resources to build out that. The furniture vertical. Do you have to do that?

Speaker B: Vertical is, uh, is a. Basically we provide the plumbing. But our partner, punish.com they are on their business. Right. So they run the business. They would do the marketing whatnot.

Speaker A: Yeah, that plumbing side is, is key and that's where, okay, that's, that's, that's why it's easy for the merchants is you've done all the hard work.

Speaker B: Yes, exactly, exactly.

Speaker A: So um, okay, now what about other verticals, other uh, you know, product categories do you have to do the same for let's say jewelry or apparel or

Speaker B: uh, I mean jewelry and apparel are pretty straightforward. That's a standard, you know, uh, bread and butter for us. Uh, ah, because that's going to spread across different partners. Whether they're publishers or there are LLMs or somebody else, uh, they will all use our uh, uh, APIs for other verticals. Uh, and also uh, one interesting fact that actually we've been evangelizing lot of these destinations to not be a merchant of record. Because early on, before we came into existence, firmly came into existence, if anybody wanted to build a native shopping experience, the only way to do that is to be the merchant of record. And also they have to build some sort of a store sink. Right. Where all the partners, uh, merchants have to come and publish their catalog into that store sync and also keep that store in sync with their website or their inventory price promotion, whatnot. That's how the industry was functioning in the past. But then after we came and we solved that problem for the platforms saying listen, you don't have to be the merchant of rakat, let the merchant be the merchant of record. That's one evangelized we did. And uh, because it was a little bit of an alien topic for them to initially understand, but later they understood the key benefit. The key benefit is it saves you anywhere between 6 to 8% of GMB. Why? Because you don't incur credit card fees. You know, it's anywhere between 2 and a half to 3%. You have fraud fees about a percent and then you have tax uh calculation another half or half a percent and there's so many other customer support and returns, refunds and so on. So many uh, things going on. All of that cost is now become zero. Okay. Because if uh, you use my pipe, it goes directly to the merchant and the order gets dropped into the M merchant and merchant has all that infrastructure in place. They uh, are, that is built into their model. So, so that way that they get the order and they, they, they handle all of that. So as a destination you don't have to incur all of that 6 to 8% cost uh, by using Medpack. Right. So that's a big savings for that. So furniture.com. initially we are planning to go as a merchant of record. After uh, they came to us, they realized oh that's a big savings for us. We don't have to incur that cost. I mean obviously they have different costs but uh, that's that, that anyway they will incur whether they are in merchant of record or not. But they have different costs. But this is a savings they got out of this. So which is a happy customer for us. Uh, they're using it and the orders are flowing every day and it's, it's working well.

Speaker A: So right now when a customer goes to furniture.com places an order.

Speaker B: Mhm.

Speaker A: That payment is not actually taken by furniture.com it's passed through to whoever the merchant is. The merchant is y. But the furniture.com still is from the point of view of the customer is the seller.

Speaker B: They, they saw that transaction happening in furniture.com.

Speaker A: right.

Speaker B: But clearly the messaging in furniture.com will clearly say that you are buying from 1 Kings Lane or you're buying from to Furniture or you're buying from whatever that merchant is. And you will immediately get an order confirmation email from the merchant.

Speaker A: Right.

Speaker B: Thank you Darius. For the uh, order. Here is the order confirmation number and whatever the next steps are. From that point on you become customer of that merchant.

Speaker A: Okay. The um, so customer service warranty, all of that is handled by the merchant merchant immediately. Um, all right. So that, I mean those are some really important things and where you have to spend money to take uh, care

Speaker B: of that, manage all of that. So that is where it is made very smooth. And the merchants are already set up to handle all of that.

Speaker A: Okay.

Speaker B: And they like the customer support to be rendered to their customers rather than somebody else disintermediating and handling it, which means they lose that touch with the customer.

Speaker A: Right.

Speaker B: We're enabling is making sure merchants are in the control end to end.

Speaker A: So in some form of like a classic definition they are. Furniture.com is more like a lead generation system for merchants, not an actual marketplace.

Speaker B: So typically a marketplace, we don't have a term yet. I would say this is like a order generation platform.

Speaker A: Okay.

Speaker B: Right.

Speaker A: Okay.

Speaker B: It's not lead generation because this is generating orders for you.

Speaker A: It's. You're right, it is actually generating. And that's what Chat GPT wanted to do.

Speaker B: Yeah.

Speaker A: Right. And they decided to go back on that. No, let's talk about that a little bit.

Speaker B: Yes.

Speaker A: What this.

Speaker B: I mean based on um, what I heard when they were in Shop Talk when they made the presentation because this, some publishing houses make some uh, you know uh, sensational stories which may not be the truth actually because uh, uh, OpenAI seems to be continuing to pursue the uh, shopping experience. Uh and, but, but they're going through an evolution process of making it as a scalable solution. That's my takeaway from, from that conversation I had as well is that they are pursuing it and the idea there looks like the way they are pursuing is actually have this shopping experience natively. You go search for a product chat GPT would render natively the response results saying okay, you're looking for a uh, a watch. It's going to show you the results for whatever based on your criteria it's going to show the results for the watch. Now if you want to buy that that is where uh, I think they're going through different approaches to say okay, if you're buying it from a specific merchant then that the merchants app would automatically get launched and now you would check out from the merchants app but it will natively be within the ChatGPT experience. Right. So I mean that's again I'm hypothetizing it based my understanding. It is still evolving. Uh, how it will eventually end. The end result would show, we have to see that. But if it goes in that route, what we will be offering for our merchants is that app experience will be branded for our merchant. We will provide that. Okay, so uh, that they don't again they don't have to do any engineering work. Our value prop is very simple. No matter what the destinations adopt, we will be able to absorb that shock and give you a good experience for your customers and then eventually you know no code is needed. That's, that's how our thought process is. That's how we're thinking right now.

Speaker A: Uh, do you know of any LLMs that actually enable end to end transactions?

Speaker B: I mean they're both ChatGPT and Gemini has enabled um, one or two merchants but just to as a PoC initial PoC. But the robust build out I think is going to start rolling out in coming weeks and months. Uh so we're just eagerly waiting. Our ah, merchants are ready to go as well. Um so we're just, I don't want to quote time for their time. Uh so but right. But our understanding is coming weeks and months they're going to start rolling out and we will be able to enable our merchants.

Speaker A: What does the term agentic mean to you functionally?

Speaker B: Okay, agent.

Speaker A: That's a trick question.

Speaker B: That is trick question for sure. I'll I'll explain in two parts. One is, you know, you, you, you ask a simple text query, uh, rather than a specific keyword search. See, in the past it was always a keyword search, right? A Nike shoe or running shoe or I'm looking for a dress. Blue, blue dress or green dress, whatever. We, we were trained to search those type of keywords. Now the LLMs are giving us an option to be more descriptive on the search, saying, hey, I'm planning for a marathon. Um, I'm looking for running shoes. What would you recommend? Or I'm going for a hike to Mount Rainier. I need some equipment for that, whatever that may be. So that's a specific. You're giving more detail your intent with the use case tied to it. Then the system goes and finds a product and now you want to purchase. You're making a purchase right there. Right. There's no redirects, nothing. They're buying it right there. So that's one side of the story. The other one is what I call as an utopian story where agent autonomously going and buying things for you. Okay. So that I still don't feel is in more realistic in in lot of context other than two use primary plausible use cases I could think of. Okay. So one use case is I'm planning to buy a TV for my living room. I want this specific TV and uh, buy it for me if the price drops below X. Okay, I know what I want to buy. I'm very clear on what I want to buy. I just want you to monitor and buy this for me. The price is the only factor is unknown. If it drops below this, buy it. Okay. Uh, and, and, uh, so those type of constraints I can put to an agent. Now the agent can monitor this product on regularly checking the price and whatnot. And if the price drops below that amount, whatever the customer has given us an intent, then an agent can go buy that. That's one use case. The second use case could be, I have a regular grocery list that I want to buy. Okay. I can. This is almost like a subscription, right? So I, so I want a gallon of milk and a loaf of bread and some fruits and vegetables. Just get it delivered to me. So those are the two plausible use cases I could think of. I don't think you would ever tell an agent saying, okay, order a jewelry for my wife and have it delivered.

Speaker A: And expect to be married.

Speaker B: Exactly.

Speaker A: Expect to remain married.

Speaker B: Exactly. Maybe if you don't want to be

Speaker A: married, perhaps not sleep on the couch for a week.

Speaker B: Exactly. Sleep in the dark house. Right. So I don't see that happening. And uh, also keep in mind people take some joy in shopping. So I don't think you could just take it away by saying okay agent, go do the shopping for me. I, I, I don't see that as a plausible use case. However, what I do see plausible and I do see things more realistically happening is the, the, the, the discovery time it takes to discover a product has shrunk. Okay. Earlier it used to you go to search, keyword search and you did you go through 10 different sites to find a product and then you, if lucky you may end up buying it. A lot of times you, you may lose the intent to buy because the time you took to find, try uh, to find a product and you really didn't end up you know, finding the product or the experience was so bad that you just left it and you're gone. Now if you shrunk that discovery time saying okay, I'm looking for something specific and the LLM was able to smart enough to go find that specific product. Give few options, not thousand options, you know, five or ten different product options. Yeah, I can look at that. Uh, okay, I like this. Maybe different color or different design or whatnot. And then once you choose a product you really want to know inventory and the price real time, not you know, cached data. Two days ago, three days ago data. Because when you try to buy then the size 7 you're looking for, it's not available. It's frustrating to corn and you try to place the order, you want to know at that moment of um, finalizing the research. Is this available to this merchant? Right. That's where formly comes in and gives you real time inventory and price.

Speaker A: Okay.

Speaker B: Okay to make that decision. Mhm.

Speaker A: Yeah. So I think I, I totally agree with uh, from my point of view with your um, your thought on it being case by case agentic across the board as a, as a, like an overall umbrella that covers everything is definitely not gonna take wings in any, any, any at least future that we can envision. But in case by case definitely uh, can help. And I mean that whole idea of an agent agent is, could be like what we used to have as a personal shopper. Right. Personal shopper is an agent. Somebody said you know what, I want to change my wardrobe. And he would ask them all the questions to determine what kind of products. So that's all the uh, really the agent is replacing a personal shopper?

Speaker B: Yep. And that is an area really agent E Commerce can do a Lot uh because uh, I was playing with uh this app called Glance uh today and uh, they have built a very nice user experience uh what they call as a try on feature. So it may takes you, it asks you to take a simple selfie and then you know gives a great uh, variety of options. Uh in terms of sunglasses, in terms of uh, dress and whatnot. Uh that's, that's something you would serendipitously identify over. You never went to buy that product but when you see that with you wearing that dress and it, it would look how nice it looks and it says oh I want this now. Right. And then you may end up buying it. So that's, that's where the agent E commerce is going to evolve to more

Speaker A: you know, to give you ideas things you didn't know. Uh, I mean there are like there are so many use cases and it could be even. I mean that's the, the wonder and complexity of shopping is in the morning I may want to just do chat at noon. I may be okay with the video, live video somewhere in the evening it's different on the weekend it's different. It's like it's all over the place. And you can have agents for each one of these specific areas. And you know for example I want to buy, buy a uh, a gas range for my kitchen right now. What's the latest technology? Who has the best product? Take all of that research away from me and show me the products and show me what I can purchase. And if you do a good job, maybe next time I will let you buy it without even me. You know, I mean that's, that's the other thing is like a uh, personal shopper. If they do a good job, you trust them them. So trust could be something that could be built up in an agent. AI agent.

Speaker B: Yeah. That's going to take a while for somebody to give up that to some um, the control.

Speaker A: The final decision making the agency.

Speaker B: Yeah, exactly. It's going to take some time but definitely I'll give you a real example. Uh, my wife was trying to buy a dishwasher and she spent at least four hours researching. Right. And then getting frustrated. That is a real place where an AI will help solve that problem of uh, identifying and bringing the right narrative on saying why this product is better for. Compared to the other one. All the comparison and all that. Uh, and, and that is where if the merchants are not present in, by providing, participating in that you will be left out.

Speaker A: Yeah.

Speaker B: So, so you need to be present in those moments and to be present means you have to do all the plumbing.

Speaker A: Right.

Speaker B: Uh we come and don't worry about the plumbing. We'll take care of it. We'll make it happen for you.

Speaker A: So what's next for firmly?

Speaker B: Oh there's massive roadmap for us so uh we're actually building a lot of partnerships across the spectrum and uh make this a reality Agent E Commerce making a reality. It's still in that theory stage to make it the first summit we will reach uh because we are at the base camp where we are equipped now we have an agent that can go and integrate with the merchant. So that's the base camp.

Speaker A: Right.

Speaker B: Right now we are fully geared up. We're going to get to the next summit of making this all this ecosystem work in a harmonious way. And there are still a lot of unknowns in terms of how this uh fraud systems are going to get the signal. That's an open issue. There is some sort of a model has to be identified meaning how do we pipe this data uh what type it may not be the same signal the fraud fraud systems are receiving today uh may not be the same. There may be different type of signals has to be sent. All of that has to be figured out. Uh there's still lot long way to

Speaker A: go on that now what is the um, the uh. I guess the revenue model how do the customers get uh charged

Speaker B: uh the shop customer who's shopping they pay the regular price. There is no change for them as

Speaker A: a consumer the merchants I mean who how firmly makes Right.

Speaker B: So so we're still uh evolving on that. Uh that's uh a an area that's been evolving. Uh different models have been discussed. Nothing has been set in stone yet uh but we try to make it simpler for the merchants.

Speaker A: So what are the most common ones like what other companies or tools are using is like a percentage is pretty common.

Speaker B: Yeah but uh right now we're not doing a percentage uh uh for LLMs. So we're still I mean it's a very early stage. I don't know if I can yeah

Speaker A: I don't want to get any like confidential information. I'm just trying to understand how the

Speaker B: still it is being worked on.

Speaker A: Okay.

Speaker B: It's not a public information.

Speaker A: Okay.

Speaker B: Okay.

Speaker A: All right.

Speaker B: So if somebody beauty is uh something I can say about it by working with us ah you will save money than what we would otherwise.

Speaker A: Okay.

Speaker B: That is something I can tell.

Speaker A: Okay.

Speaker B: Because it costs you know if you were to go and enable uh do an integration with uh, with each of these LLMs, each one would be a bespoke integration you're doing and that would cost at least half, uh, a million dollars for merchant.

Speaker A: Okay.

Speaker B: We're going to be significantly less compared, uh, to that amount of cost.

Speaker A: Is there a specific merchant size that you're looking for or you're working with right now as far as uh, dollar volume, M or.

Speaker B: That's a great question. I mean we, we don't discriminate merchants. Okay. All of them.

Speaker A: Okay, good, good.

Speaker B: But, but our uh, sweet spot is obviously larger merchants, uh, and uh, midsize. But, but doesn't matter. We, we want to help everyone.

Speaker A: Okay.

Speaker B: We don't discriminate.

Speaker A: So you will, you will actually spend time with a small merchant to see

Speaker B: if you can help them because it's easy for us to do it.

Speaker A: Okay.

Speaker B: You come and you come to formally AI and you go firmly connect. You sign up.

Speaker A: Okay.

Speaker B: It's a self service for most part. Right. And then once you sign up we, you have a menu option where you want to sell your product. You want to sell your entire catalog, partial catalog. All those decisions are self sign up.

Speaker A: Okay.

Speaker B: Once you do that and you sign our merchant agreement, then you're off, uh, to the races.

Speaker A: Okay. That was. Yeah, you answered my next question is the process is if somebody wants to check this out, they go to firmly AI, go to the connect connect, sign up and go through that process.

Speaker B: Yep. It's very simple. It'll take probably 10 minutes of your time.

Speaker A: Okay.

Speaker B: That's it. And then you're ready to rock and roll.

Speaker A: That's very interesting. Okay. Um, now what's the um. As a, as a founder and CEO of a company that is just getting things going and getting into the growth stage, what is the biggest challenge? What keeps you up at night most these days?

Speaker B: That's a great question. I mean, you know, in last five years, every day was, you know, it's a different set of challenges. Right. Uh, early on, uh, the, the market was not ready for our product. Uh, so it was such a deep struggle to get the market ready for it. So we had to do so much evangelizing to, to convince people this is great. This is going to help create, create a good shopping experience. Uh, instead of blind redirects wherever the product is identified by the customer, let them transact. Right. So they should be buy that. So that evangelizing took a while for people to grasp and then know two, three years it took. But while we were doing that, we were building the product as well. And There is so much, you know, we 99.9% of the time we'll get a no. And then there will be 1% of the time we'll get a yes. That yes actually leads to something bigger. And then eventually, you know, we, we were able to turn the tide to saying, okay, yes. This is actually. I don't know why nobody did it until now. That's the uh, point I hear. Now I wonder why nobody did this. Of course.

Speaker A: Well, you know, I told you 100 times before, now this is, makes sense. And you said no, and suddenly now you're telling me that.

Speaker B: Exactly right. So as we went through that, now I don't have to tell that story anymore, right? People get it saying, yeah, now the question is, okay, how do I make it happen? So to do to that extent, when I say, by the way, you don't have to write any code. I'll enable it, and says, no way. I don't believe you. How do you do that? And says, okay, allow me to prove. Okay, you don't have to do anything. Just sit and watch. I'll make, I'll show you. Then we just do the integration and we show them and say, oh, this is great. You already integrated. Perfect. Okay, now what do I have to do? Says, okay, sign the agreement and we move on. It's been a, it's quite a journey.

Speaker A: Okay. All right. So right now, uh, like are you, I'm sure you're building your team out. What are some of the key roles that you're looking for?

Speaker B: That's a great question. So right now I think I have more people than I need.

Speaker A: Okay.

Speaker B: Because, because the, the, I mean, uh, they're all busy, uh, right now, but I don't think I need a lot of people right now at this moment.

Speaker A: Okay.

Speaker B: Because of uh, all the enhanced improvements that has come in. AI, we are able to achieve things with, with uh, with the tools that available for, for our M employees that they're able to produce more than uh, you know, what they would used to produce before.

Speaker A: Okay. Okay. Do you have the typical sales motion that all the other B2B. Um, or is it like product led growth or is it, ah, you actually have SDRs and teams that do that kind of work?

Speaker B: Yeah, great question. So what we are looking at is to, as a partnership led growth. Uh, one is product led growth and the other one is partnership led growth. So product led growth is more organic. You know, word of mouth. Customers see our uh, product and they hear about us and they come sign up and they get the benefit. So that's one side. And, uh, then there's a partnership side. We. We focus primarily on partnerships. On the PSP side. Uh, we've been working on many partnerships. We're going to be announcing some of them in the coming weeks and months.

Speaker A: Uh, okay.

Speaker B: Bunch of large partnerships that we're working on. Uh, so they would recruit the merchants for us. Right. So there's a lot of motion in play. And then obviously, the destinations also. There's a lot of them in motion as well. And as. As both sides power up, uh, that will lead to significant growth.

Speaker A: Okay.

Speaker B: That's all.

Speaker A: Okay. Great. Well, Kumar, it's been, uh. Uh, really fascinating speaking with you and learning about what you're working on with. Firmly.

Speaker B: Oh, thank you, Darius. It was. I was. Enjoy. Enjoy the conversation.

Speaker A: Likewise. Uh, all the best.

Speaker B: Thank you. Thank you so much.