Transcript
0:00 You've described finance as the decision engine. How is that different from what most CFOs think their job is? >> The old school method of CFOs, it was like we'll run the dashboards, we run the scorecard, here's all the metrics. And like that's great, but our job is not to be the historian. You know, my job is to be the navigator. >> A lot of people would say, oh, you know, we're spending too much on R&D. By plowing money into there, you're increasing the resolution rate.
0:25 >> One of our key metrics that we look at is the success rate. Our equity rally will grow the fastest by having the best product that's everyone in the marketplace says like hey this one is the best. >> You've said that order to cash isn't a back office process. >> The order to cash process is really kind of a strategic growth asset. I think if your monetization is messy like your whole growth engine is pretty fragile.
0:47 >> A lot of people listening right now who are just saying I only want the CFO title. How do you think about that? >> One of the most influential conversations that I had was like sometimes you got to slow down to move fast. Is this thing on? >> YESTERDAY'S PRICE IS NOT TODAY'S PRICE. >> Welcome back to Run the Numbers, the show where we talk with the world's top CFOs and finance leaders. I'm CJ at Tech CFO and my goal is to unpack the frameworks and operating principles that make you better at your job. On today's show, I'm speaking with Canal Agarwal.
1:26 Panel is the CFO of Gorgeous, an AI customer support platform that's purpose-built for e-commerce brands. On this episode, we go deep on AI pricing. We talk about how Gorgeous thinks about outcomebased pricing and forecasting the LLM costs that go along with it. We touch on order to cash as a growth lever and why Canal believes what most companies treat as a back office process can actually become a competitive advantage if architected correctly. If you like the show, please remember to like and subscribe. It helps us with the algorithmic overlords. If you're looking to hire the best finance and accounting talent, I'd love to help you. I run a recruiting service that pairs you with thoughtful, qualified candidates from our warm community of finance leaders, people who voluntarily research renewal rate calculations on weekends. If that's of interest, shoot me an email at talent mostly.com and we can talk. On to today's episode with Canal Agraal.
2:14 Canal, thank you so much for joining me on the pod today. >> Yeah, it's great to be here, CJ. >> So, you're a CFO. You've led high performing teams at companies like Survey Monkey, 8 by8, and Non. If you'll bear with me, I actually want to rewind a bit to a different chapter because I think it says a lot about your appetite for risk. So, you left private equity back in the day to start a venture incubator. What was going through your head at that time?
2:39 >> I think a lot of people have that same question uh at the time. I think I remember I had just gotten married. Most people don't make the transition from, you know, a cushy, well-paying VP job at a software private equity firm to kind of start your own company. That's just not a traditional kind of job uh jump that people make. But I've always kind of had an outsized appetite for risk. And I think what's what's really compelling for me was just like what the thing I'm intrinsically motivated by, you know, and I spent time prior to that in investment banking and then in private equity. I got to work with some amazing founders, you know, really got to see them kind of build stuff um and build like durable lasting businesses.
3:18 And I think when when I was looking at that, I said, I would love to kind of feel a little part of what they're trying to do. The deal business is great. There's a lot of really smart people on that side of the world, but just felt like I need to push myself out of my comfort zone. And I wanted to learn what it meant like to go from zero to one. you know, when you grow up in the world of accounting and investment banking and private equity, it's like it's pretty structured. You know, there's a playbook on on how to do things. And I think for me it was like, hey, push yourself out of the comfort zone, learn how to do something you've never really been exposed to to doing.
3:49 And so, you know, when you join kind of a venture incubator and, you know, I was kind of like the GM of like these different products and you're supposed to basically come up with an idea, build a product, find customers, find users, and go from zero to one. I, by the way, had zero knowledge of product management. You know, this was before the good old days of VI coding where if you didn't have technical understanding, you could just create a product. But I learned a ton. I got to build a couple companies when I was there, actually get paying users. Uh, one, I think I realized that product management is not my superpower. It's not the the day-to-day thing, but amazingly has created a lot of differentiation for me in the latter parts of my career. uh just to be able to really understand product at a much more deeper level, be able to really understand how we think about building solutions for customers, what that actually means from a a technical perspective. It's given me more of like an operational empathy that I think has been really critical in leazing with my partners on the product and engineering team.
4:45 >> Do you remember the feeling of that first paying user? >> Yes. Oh, it was the best. I remember when we got a couple of people. There was I think we had three people sign up on one day and actually like giving their online credit card to to pay for stuff and it was just like an amazing feeling. So I I totally know why people would want to build something and when you take an idea and you create it and then people are actually paying for it, it's an incredible feeling.
5:09 >> It's kind of a funny story. I was the director of FPNA and treasury at a startup and on the side on nights and weekends me and my wife had a company in the tourism tourism industry. It was last minute uh activities that you could do in your location when you landed. And I remember being on the phone with a banker from SVB and I'm putting through a spot trade to to to put through $5 million of USD to Sterling and somebody beeps in on the other line. It's a Florida number and I pick up and it's this mom who's trying to find the location for where the jet ski excursion goes off from. I put the banker on hold so I could do the customer support call for the jet ski person and it was like an $80 purchase and we probably made like 10 bucks off of it. But I'm like this is what I'm getting hyped up about today.
5:57 >> It's really cool, you know, to to kind of like have that feeling of building something. And so you kind of also gain a little bit of sense of that euphoria of like what it means to to build something that customers really like. And so I think it also helps enforce a little bit of the responsibility you have to customers. Uh which you know sometimes in the finance world you're a little bit of removed from all of that.
6:17 So that's been that's been really helpful to have had that experience. Non and survey monkey they're very engineering ccentric orgs. Can you say more about how the experience with the venture incubator gave you more empathy for the product side of the equation? >> I think absolutely. That's why I I alluded to the fact of how beneficial it was actually for me in my career was you know ultimately those specific ventures quote unquote failed but I think I learned so much about how to be an executive in a very crossunctional especially at tech companies where to your point you know you have very engineering organization and so there's this concept that you know I was taught around operational empathy and the idea that like do you really understand what the your executive counterparts like what they're really trying to solve for right where are their pain points, where are their pressure points, and what are their viewpoints on like what they're trying to solve in the business. And I think a lot of times as in the world of finance, we we kind of think about things sometimes black and white. And if you don't put yourself into the shoes of what these other folks are trying to do, what their challenges are, you know, it's hard to help help them solve problems. And so I think being able to have that experience, it really just helps build a lot more contextualization, right? When they're talking about, hey, the product trade-offs or product roadmap that they're trying to architect and like what are the trade-offs. I'm not going to pretend like I'm a product master and superstar, but at least I understand what the approach of like how they're thinking about the customers, how this all fits in. you know, I think going to have more intelligent questions on what are the things that are really moving the needle for the customers like how do you help them prioritize on certain things. So the feedback I hear a lot is like that's not what a typical CFO would say, right? I think I always start first by trying to understand what are they trying to do from a customer standpoint like what are they trying to build? What are the technical challenges? What are the product challenges and then solve back maybe there's an outcome that's somewhat financial related but like that's not where I start.
8:08 >> That is such a compliment to hear. >> Yeah, I know. It's kind of like that backhanded compliment, you know. So, I'm like, I'll take it. >> But we talk about this on the podcast, Canal, quite a bit that if you're willing to just do 10% more work to understand what they're going through or ask the question or say like basically take me for a ride along on like how you ask questions to a user, they give you 10 times more back. So, just putting that extra amount of effort in, they're like, okay, you want to come into my world, I'm willing to reciprocate in this big way. and now you're kind of under the tent, but you have to be willing to feel uncomfortable and it seems like a very stupid question to ask them about their day-to-day.
8:46 >> I totally agree. I mean, I have a lot of time in my calendar where, you know, I meet with like I see sales reps, I meet with folks on the product team, I just meet with people like out there doing the work and like understanding from their vantage point. Like the best part of my day is when I'm in the office and I sit for lunch with some of the other folks on my team, people who are CSNing or aming or building the product and just kind of getting their perspective on like hey what are the challenges like what are you seeing the push back is like you don't always get that you know in the executive team room and so this is like to me like the most helpful to get some of that knowledge really understand what's happening you know with the folks out there in the business who are really at the front lines and like that's incredible intelligence for me to just like really understand the temperature of where people are at and then where our customers are at.
9:33 >> Can you say more about just hanging out with IC sales reps because I feel like there's so many truths there. >> Yeah, I love it. I mean, like, you know, we'll get I'm in the San Francisco office and, you know, I'll travel to New York and Toronto and Paris to our other locations, but you know, sitting down at just lunchtime and not even having like a formal meeting. I think it's actually worse to have a formal meeting with the sales team, you know, member who's like trying to busy closing deals. Plus, it feels like, oh, the CFO wants to meet with me. Like, something must be wrong, you know? But I think just like sitting down with someone during lunch, kind of asking about what's going on. They're just asking casually like, hey, what are you seeing out there? Like, where are some of the deals you're working on?
10:08 Like, how are people responding to this product or this product? Like, where are the push backs at? Like, you know, how's your working relationship with the product team? Like, where are you hearing things are really resonating? That is like incredible intelligence for people who are talking to customers every day, right? to hear and kind of get their opinion on what's resonating, what's not. I won't hear that in such an unfiltered way very often and like to be able to hear that from people. It really is valuable for me to kind of help calibrate and and steer the executive team based on like, you know, multiple signals.
10:38 >> I have this friend Dom, hopefully you're listening, Dom, VP of FPNA at the time. Whenever he had like 10 minutes in between meetings that you can't really fill with anything, he would go over to where the BDR sat in their bullpen and they had this putt putt there. So he would just grab the putter and he would just hit some shots and I'm like this guy just loves golf like classic. But what he was doing is he's just listening to what they're saying and inevitably they start talking to him and and then he gets invited to go get beers with them and maybe they just want the finance guy to pay. He would just hang out over there just to hear what signals and you had context on deals and pipeline that I feel like you wouldn't get by the time it was filtered up through all the different layers.
11:13 >> I think those are kind of like the the best learnings. We're in different regions, continents, countries across the world. But the best times are when we all get together and we all get together like you know just having a beer with someone or having a coffee from someone from a different region who's selling and you be able to kind of hear what people are thinking and and especially when you get to sit in a group and hear kind of different bunch of different people who ultimately you know maybe don't know who I am or aren't don't care who I am. they just kind of have their own conversations like being able to observe that water cooler talk around what's working and what's not working in the company. That's where you really kind of get the nuggets of stuff.
11:49 >> Yeah. And if you can do it, like it's amazing because the CFO title, it does carry some gravitas to it and people sometimes will be like, I got to be locked in right now. But if you can get them in a position where they trust you and they also don't think you're like going to try to collect info and burn them, like you can ingratiate yourself into the org. when I ever hire someone people on my team that's why I just tell them like get embedded in the organization like you know this is not an ivory tower profession you know like you got to go sit in the team sit with them like understand what's happening build relationships with them you know and go to their team events like that's where you're going to be the most impactful >> hey thanks for listening we'll be right back after a word from our sponsors Ali and the team at Tabs are building something that directly addresses one of the biggest headaches I see finance teams deal with today pulling data from your ERP, your CRM, your FPNA tools, and your usage systems and actually making it all work together. If you're running usage based pricing or complex contracts, you already know how this usually plays out. Contracts live in one system, usage data lives somewhere else, billing happens in another tool. Rev gets layered on top at month end, and finance ends up just reconciling everything by hand to close the books.
13:00 This is exactly the problem Tabs was built to solve. Tabs is an AI native revenue platform built for controllers and CFOs. It brings together data from your ERP, CRM, and real product usage into a single system of record. From there, it automates billing collections and revenue recognition without finance having to duct tape workflows together every month. If usage based revenue is core to your business, this is the future of how billing and revreck get done. Go check them out at tabs.com/run.
13:30 That's tabs.com/run. We've all been burned. We've all bought that enterprise planning tool that promised the world only to realize 6 months later that we've basically taken on a second full-time job just to keep the software running. I want to talk about a company that was built specifically to kill that cycle. Abom. I actually remember my very first conversation with their founder and CEO, Julio Martinez, over 3 years ago. Back then, they were just starting out in Spain. Fast forward to today, Julio moved the home base to NYC and they're the engine behind finance teams at Strava, Replet, and JG Wentworth. Abacom doesn't turn you into a software admin.
14:07 The integrations are actually self-service. You don't need a $300 an hour consultant to plug in your ERP or HIS, and they're doing AI in a way that actually impacts the things FPNA teams are doing every day. Things like creating variance summaries, building formulas, and modeling scenarios. If you're scaling fast and you're trying to avoid that legacy platform trap, Abacum is the move. They're building the future of our tech stack and they're doing it with a CFO's perspective. Go to abacom.ai to see it for yourself. That is abacom.ai.
14:39 If you're paying for A-level finance talent, they shouldn't be doing B-level tasks. CFO time is expensive. Senior finance hires are wicked expensive. And yet, in many companies, highly paid operators still spend hours reviewing expenses, chasing receipts, and reconciling systems that should already be automated. That's where BRE comes in. Brex is an intelligent finance platform that combines corporate cards with built-in expense management and AI agents that automate the repetitive work finance teams usually handle manually.
15:10 Transactions are categorized automatically. Receipts are matched. Policies are enforced in real time. Reconciliation just runs in the background. So instead of adding admin as you grow, you increase output per finance hire. Rex is already automating hundreds of thousands of hours of manual finance work every month across 35,000 companies including Anthropic, Coinbase, and Door Dash. If you want your finance team focused on performance instead of paperwork, check out brex.com/metrics.
15:41 That is brex.com/metrics. You've described finance as the decision engine of the organization, which I love. I'm totally going to steal that. How is that different from what most CFOs think their job is. >> Yeah, I think this has certainly gone through an evolution. You know, I've been doing this in the field for 20 plus years, but I think, you know, when I saw the old school method of CFOs, it was like, hey, like I'm going to run the scoreboard. We'll run the dashboards.
16:08 We'll run the scorecard. Here's all the metrics. And like that's great, but our job is not to be the historian. you know, my job is to be the navigator of where we're going. And so, like I think about it as being kind of like in I don't want to run the scoreboard. I want to run the playbook of like what's the plays we should be doing on the field. And the role of the CFO is like I think even just way beyond just what's happening at finance like I said like we're running the playbook for the entire team. And that means go to market engineering product marketing just that whole suite of like people who are doing stuff across the company. Can we get an integrated system of like the data of people and processes so that we can pipe in what we know are the right signals, the right context and help turn those into decisions, right? And then have a feedback loop on like are those the right decisions? Do we need to revisit those or move on? And so I really think that is a critical part of my organization is to really help architect those things like the data, the people and the processes so that we can continuously make better and better decisions, make them quicker uh and be more agile with the decisions that we are making. But it's it it's kind of like it's got to be all part of one system.
17:17 >> Something that I know you're really passionate about architecting and having a hand in is pricing. and it's it's come to the forefront with AI now. So, I wanted to dig into how you've applied AI pricing to your own product, which I went down like such a rabbit hole on your website just preparing for this conversation. It's really cool. Maybe to start, can you just give us a 30 second overview of Gorgeous and and maybe what your company does?
17:41 >> Gorgeous is the conversational AI platform for e-commerce. So, we are the leading provider of customer experience software on Shopify. So, what does that mean? It means like, hey, if you're a merchant who sells on Shopify today, you want a way of managing the conversations you have with your customer. So, when a customer comes to your website, you want a way of helping that customer understand what to buy, suggest things based on what's targeted to them, individualized for them. We help you purchase. The legacy of our company has been on the post-ale side of things. So, kind of support for people who have purchased from an e-commerce merchant.
18:13 And then kind of the last piece is like re-engagement. And so like we kind of help maintain that entire conversational life cycle for merchants uh and predominantly selling on Shopify. So we have 17,000 customers. Uh we're the by far the leading provider on the Shopify ecosystem. And yeah, so like I think you think about AI agent like that's where you know we think the future of the support is like I think that customer support and the ways that like customers are actually engaging with e-commerce like you know it's it's not us saying it. It's like you know Tobies and the Open AI folks and the Google folks like the future of commerce is changing dramatically. I think customers the way they browse like know people aren't necessarily going to a website browsing through a bunch of nested menus and like selecting a product and off they go.
19:00 Like customer behaviors are changing. So we want to be at the forefront of that with our Gentic AI product. And our AI product serves a couple of use cases. One is to help you discover your product and then to help you with automating support. >> Yeah. The discovery piece is so key. It goes beyond just the basic support that you see coming out with a lot of companies. >> Yeah, absolutely. Like I I think about this like the the moat and the vision we have is around know one platform to handle all the conversations you have with your customer. What people don't want is they don't want to come to a website, have this thing that feels kind of separate that, you know, is not really customized to you and just start suggesting a bunch of things and now you know I bought something and now I need support for that and now I have to go with a different system that doesn't really plug into the rest of the things and it just feels very disjointed. So I think the angle that we're taking is like hey one platform for all your conversations that you have with the customer. The idea is like by owning the customer data, the CRM data meaning like the support data that we have for each of the customers because we have 17,000 customers. It becomes like the system of record that now enables the shopping assistant to be much more powerful and enables the AI support agent to be a lot more powerful because we already own the system of record underneath. And so it just provides a lot more customizable experience for for the for the end user.
20:18 >> All right, man. Now we're cooking. So how did you land on the current pricing structure for your product? It was not definitely like you hit the single right out of the ballpark, you know, and then it's like that's where you land. Like it's just has been constantly iterating. So the company's his history has been in kind of like a help desk support platform. So traditional kind of SAS platform where you know uh help desk support software and we've kind of gone through a bunch of iterations but that's largely been on like on tiers of usage.
20:46 So to date we haven't really gated things based on features, right? If you're at certain pricing plans of ours, you have the same features as the other plans. It's really just that the limit here is like the number of tickets on this on the help desk side. So again, it's kind of like a usage based tiered approach. And then as our AI agent product has kind of grown tremendously over the last few years, you know, there's a lot of kind of schools of thought here on AI agent. And and I think what's interesting about agentic AI products is it introduces kind of new paradigms which weren't previously around in the world of SAS, right? And the world of SAS is great like you can give away a product and your incremental cost for that product is almost zero. It enables you to have a lot more flexibility and like driving for adoption uh by kind of discounting or doing things uh a little bit more aggressively. With an agentic product, you have real unit costs, right? our infrastructure costs are single biggest expense driver uh of our company year-over-year. Like that's the single biggest increase. And so you got to be mindful of that. Uh when you're a startup and you know you're minding cash burn while you're also trying to grow, it's a it's a difficult line to balance.
21:56 So for us, I think where we landed on our AI product was again I always think about like start back with the customers. They feel like they're going to understand the ROI of what you're selling. And I think as we kind of talked to more and more customers, it came to be like where there's an outcome based like a resolutionbased pricing model. So for our AI agent product, like we will only charge you when we actually resolve an interaction. So when we do what we say we're going to do, which is like our AI resolves an interaction that otherwise your human would have to do, right? We charge you a fee that's far less than what that human costs per interaction. But if it doesn't resolve it, if it has to hand over, then it hands over. but we don't charge you uh for that AI Asian price. Right? Now, the the trade-off is like that's still incurring a cost for us, right? So, we have to bear some of that cost. Now, what's what's really interesting if I take a zoom back, the way to win here ultimately, right? If I think about how to grow equity value is you got to have the best customer experience, which means you need the best product, right?
22:57 And the way to grow the best product is, you know, you need the best like agentic architecture in the model. you need the fastest experience uh and the best experience right where so one of the ways of looking at this like ultimately is like hey are our inference costs kind of like sales and marketing costs do we invest a little bit more do I give the R&D folks a little bit more juice to invest in inference costs and say like hey just build a best product because ultimately that's the thing that's going to have the flywheel for us right that's going to have the hockey stick growth I'm sure folks will listen to that and be like okay great I'm just going to just run off the course with LM costs which is obviously you know we monitor that pretty closely day-to-day we have visibility into what those costs are that as a CFO is where you have to really kind of be mindful of the tension right it's you want to have some guardrails and where we go with cost but ultimately can't come as a sacrifice of quality >> so it's usage based in tiers by support ticket do you only retire a support ticket if it if the outcome is positive if you resolve it >> the legacy kind of our help desk software I'll call it help software that is just based on the number of tickets >> here you talk about the product it resonates because a lot of people would say oh you know we're spending too much on R&D etc but what you're really doing is by plowing money into there you're increasing the resolution rate which actually increases the revenue because what you're doing in an outcomebased pricing model is you're saying like I'm putting my money literally where my mouth is the best product is going to win there >> we're not raising you know tens of billions in dollars or rounds of financing like some of the folks in the in the ecosystem are but our equity value will grow the fastest by having the best product that's everyone in the marketplace says like hey this one is the best so to your point like one of our key metrics that we look at is the success rate and our AI agent success rate and success rate all that means is like hey how many interactions came in how many of those interactions were the agent able to resolve we can drive that number up and up and up that just means now like we're going to be monetizing more effectively because we're able to now automate more and more of these tickets that are coming in and resolve them, right? And so that's a net win for everybody. That's when our customers are saying like, "Hey, I'm happy to pay gorgeous more money because you guys are resolving way more tickets. I need less uh humans to do this job or I can put my humans on like the most complicated things that are out there that they don't have time to do normally and so I can repurpose their time and energy."
25:27 And so this ROI is like a really easy sell for our sales teams to go out there and and make that argument. >> Hey, thanks for listening. We'll be right back after a word from our sponsors. >> You just launched your new AI product. The new pricing page looks great. I'm talking crisp. But behind it, last minute glued code, messy spreadsheets, and running ad hoc queries to figure out what to bill. Customers get invoices they can't understand. Engineers are chasing billing bugs. Finance can't close the books. Well, with Metronome, you hand it all off to the realtime billing infrastructure that just works.
26:03 Reliable, flexible, and built to grow with you. They turn raw usage events into accurate invoices, give customers bills they actually understand, and keep every team in sync in real time. Whether you're launching usage based pricing, managing enterprise contracts, or rolling out new AI services, Metronome does the heavy lifting so you can focus on your product, not your billing. That's why some of the fastest growing companies in the world like OpenAI and Anthropic run their billing on Metronome. Visit metronome.com to learn more. That's metronome.com.
26:38 Here's a growth tax nobody talks about. Every new monetization model you ship creates a nightmare for your finance team. Add usage based pricing. Now you're tracking consumption against commitments. Launch product bundles. That's multiple performance obligations per contract. Offer midcycle upgrades. Woo! Good luck reallocating revenue manually. But that's exactly where RightRev shines. Write is the revenue recognition engine built for companies that can't afford to let accounting slow down growth. When your rev is automated, your product team can ship new pricing without asking finance for permission.
27:13 And your sales team can close creative deals without worrying about downstream chaos. To get up on my CFO soap box for a sec, I love talking about creative pricing models. Hybrid pricing, credits, tiered usage. But I've seen too many companies where the sales team is celebrating a huge quarter while finance is still trying to figure out how to recognize half of it. In a world where your pricing model might change three times next year, that flexibility is everything. If you want to scale your monetization without breaking your books, visit writereev.com. That's right.com, where modern monetization meets bulletproof accounting. Being a CFO, you know how much I love tools that actually make the lives of accounting and finance folks easier. One of my favorite tools right now is ReLit, the AI native ERP going head-to-head with Netswuite. Yes, someone is finally doing it. I met Relet two years ago when they were still in stealth. Since then, they've absolutely taken the finance world by storm. Their mission is to make the zero day close a reality, and they're actually doing it. Customers are literally closing their books at 1:35 p.m. on the first day of the month.
28:19 They've got everything you need to scale your business. Complex revenue recognition, native integrations, custom reporting, multi-entity, close management, and much more. They're only a few years in and are already supporting NASDAQ publicly listed companies. Yes, seriously. If you want to scale your business on an ERP that wasn't built in the '90s, you need to check out Relet. Book a demo at reallet.com/cj. Oh, cool. That's me. That's r i l.com/cj.
28:50 r i l.com/cj. Tell them I sent you there. I am curious how you figure out the how to split the value between the value delivered versus what what you charge for it. And I bring it back to a conversation that we had with Dan Griggs who's the CFO of Intercom. And he's like, we know that 99 cents per resolution is like wicked cheap compared to the value we're providing. I bet they could charge5 or $7 because I've had support inquiries that were 30 to $40 just as a CFO looking at how much it it costs.
29:22 >> Yeah. Yeah, I mean I think Dan's right like we have a similar pricing which is about a dollar per uh your resolved interaction and that can vary slightly but like at its core it's it's roughly around there. You know our advantage is we just do e-commerce support and postsale support and now pre-sale support. So like we're very focused on the industry we're doing. We work very closely with our merchants when we on board and when we set them up and to understand their current cost structure. So I think you'd be hardressed to understand find another company that understands e-commerce merchants as well as we do and the way that they run support is going to be a little bit different than like a B2B SAS company. So like we have pretty good data on their what their support structure looks like, what their ecom team looks like, you know, how much they're spending on resolving human interactions. And you know in in the world of e-commerce, it's interesting because like you get a lot of concentrated use cases, right? Like the number one use case for people is like where's my order? For most of the time that takes a human like going and looking that up and like figuring that out and like okay where's their order?
30:25 Let me go figure out where their thing is, find it where it's in at FedEx and then reply to them. Now like we found that to be a very automatable use case, but it's not easy to do, right? It's something that like you now have to hook into like downstream inventory systems, downstream order management systems, be able to understand how that all does. So it's not like people can just walk in and be like, "Oh, I can build that functionality." Like it's not that easy.
30:48 And so like that's where our legacy of being in this business for 10 years doing help desk support software is now like really helping the AI agent just be much more performative and be able to understand those use cases. And every merchant has different guidances on how they want to handle returns or how they want to handle refunds. And so like because we already have that net knowledge starting point, it's just easy for our merchants to layer on our agents on top of the stuff that they've already built. This is why I think players that are vertical specific and customer use case specific are going to be winners in the age of AI because people say, "Oh, where's my order?" That's a very nuance depending on what you ordered. So I ordered a treadmill the other week and I was waiting for it to get here and I was saying, "Where is my order?" It weighs like 600 lb. It's not going to be like they're they're shipping a hat to me, but I had to email them. Then they bounce me out to a third party uh logistics provider and then that person had to call me on the phone. It had all these intermediary steps and it it varies depending on the object that you're shipping and where you live. So, it's it just goes to show that the context matters a lot. People are underrating the amount of kind of specific customer knowledge that's required and the specific flows are required for specifically vertical industries. e-commerce being a very one of them like there's a ton of complexity in the back end with logistics providers software providers warehouse providers like people just don't have full visibility of that it's not the same as a BB SAS business right so there is some stuff that's quite tricky quite nuanced and I think ultimately the experience I'd love to have for our customer you go on to you know your treadmill company and they were like oh yeah hey CJ we know what you have we had a history of all your prior support tickets for certain things. Let us go do all the backend work and tell you very quickly here's where your order is, here's when you can expect it, right? Are the other things you need as part of that? Like are the things you need? Do you need like shoes or you need other things as part of that journey? But ultimately, let's make sure the question you have gets resolved very quickly, right?
32:50 Always on. And you have a great customer experience with that merchant. And like we want to embody what that merchant's tone is, what that voice is, but we can deliver that experience as like you're like, "Oh, this was awesome. This was easy. Then I think Gorgeous is one. >> Yeah. One of the front door questions was, "Did you order a bike or a treadmill?" I'm like, "Why do I even have to answer that question? I ordered a treadmill."
33:10 >> Yeah. And so like being able to tap into like your order history and be like, "Okay, we we know exactly what you ordered. Here's where it is. Let's go do the work and figure out where it is in the back end and where you when you can expect it." >> Canel, you'd mentioned earlier the cost component to this because AI does introduce real variable costs under the hood. How do you model and manage that as a CFO? I'd say the last 24 months it's been an area that you know I just haven't had that much exposure to in my prior 20 years because it's just like you know gross margin was always like some type of AWS hosting cost now like there like I was mentioning earlier they're very real costs for our costs you know we use a variety of different providers but we tool those we monitor those constantly we have real-time dashboards on you know the the cost that we use and so one of the key KPIs that we use internally is like what's our AI agent cost per interaction one of The learnings I've had in sitting with our engineering team and our infrastructure team is there are a lot of related costs that are not just Lup cost. You know, there's ways to like manage the security. You need different software tools to manage all the things that need to go in and around building a Gentic AI architecture. End of the day is like I have a pretty good view now on like what each interaction costs us. And so it's helpful for me to a baseline on like what those real costs are because again as as someone who also owns pricing decisions. It's really helpful to know what your baseline cost is like as we think about different pricing offers and we're trying to experiment with pricing which we do pretty frequently. you know what's our like real margin on certain things that we're doing because now there is real margin on it right it's not kind of like a marginal cost zero when we when we do additional AI agent interactions like there's real costa >> you make an excellent point that it's not just like the tokens it there's stuff around the tokens just architect the system that you have to put in the cost >> you know the way I look at it the question I was asked my engine is like hey if we didn't have an AI agent product how much of this would you need and so that's like to me like when it would be like oh yeah don't know if we need this or certainly not as much of this. And that's the way to kind of be like, okay, then those things need to be factored into like this fully loaded cost. It was been eye opening to me over the last 12 months to kind of understand all the different things that need to go along with supporting you not just your openi or anthropic or your gemini cost like there's a lot of other costs associated with running magentic AI business on the engineering side in particular. How much do you benefit from the falling costs of what it takes to run these models? The price per token depending on like if you're using this version versus that version is going down. But at the same time, nobody wants yesterday's newspaper.
35:54 >> I think the art of optimizing your agentic cost is going to be like a growing discipline. I think our team is better than most in terms like we have pretty good visibility into what models we're using, how many times are caught, what the cost is. And so one of the things that we are also going through is like really just rebuilding the entire architecture of how our agentic AI product works, right? So now like so much of it is like how efficient are you in the actual the number of calls that you need to make. And so like you know when you're starting off we were just you ask one question and then in the background it's kind of probably doing a bunch of stuff taking tokens or taking cost and now it's like if you rebuild the infrastructure in a more sustainable way. It's rare that this happens but you can drive quality and cost down. So quality up and cost down. It's like one of those things that you have to really invest it. You got to say like this is really important for us. we're going to do this and this is going to prevent us or then we have to not do these other things because we need to build the right infrastructure. Certain use cases have more complexity than others and not everything needs the same model. So, you know, there's maybe certain wrote kind of repeatable customer inquiries that are relatively simplistic. We could use a cheaper model for that. But some of these more detailed ones, you know, detailed questions that fit certain parameters, like we probably need, you know, a much better model to handle those and that's more expensive. So now we have to kind of think about uh use casebased models and think about the mix that we're using for different questions and different sentiments that are arising. So there's a lot of room there I think for this field to continue to grow and like how to really think about it. It's so interesting how you say that it'll be somebody's specialty going forward because I can think of people who were excellent at the cloud cost portion of things, but I don't know anybody today who's like probably some of the CFOs have come on the show, but I I've never had a peer who was like, "Oh, that's the person you go to when you need to forecast the per resolution cost using different models." Like that's a new world.
37:54 >> Yeah. I mean, this is exactly your point. And I think the comp is like, you know, back in the day you used to have like the people who are like the AWS experts, right? And like kind of help you really understand think about spot purchasing versus cost and like how you maintain those models. Now it's going to be like, hey, like how do I think about the right composition mix and the right mix and like how do I think about the cost worthy interaction and what are all the surrounding costs and how that go up and down and to have someone who really knows how to think through those and can help partner with the edge team to develop the architecture alongside that.
38:25 That's a valuable skill. It's got to be hard though thinking about we talk about a growth versus efficiency like as the P&L as a whole but even going down to the LLM level you have to decide on each transaction on each customer cohort how much you want to protect your margin versus give them the best of the best at all times to make it so it's delightful to use. >> Yeah, that is the no easy answer question. Right. Again, I think the northstar is customer quality, customer experience. If I'm shortch changing that, then we're just not going to win as a company. So now the question is, how do we put the right guard rails in place to still have that as the northstar? How do we continue to maximize the customer experience, the wow moments that the customer can have, but knowing like some of the things that we can do on our side to build the right agentic architecture, the right mix of different models. And you're never going to get 100% right. And sometimes you have to go tune in of like hey maybe this response was took a little too long or this was like hey not the right model for this side of question.
39:24 >> I want to transition a bit to talk about order to cache something both of us are passionate about and you've said that order to cash isn't a back office process which I totally agree you you see it as a growth engine. Can you break that down for me? >> Historically you know when I've been in different organizations order cashache has really been kind of like a fragmented series of systems. There was no real ownership. that was kind of discreet. You know, you have like some billing tool or some order intake tool the sales team is using and then there's some like homegrown billing team that's like, you know, it's an engineering thing. Everyone hates their billing tool and then you got something where people are collecting cash, but it's like this is just like different pockets of stuff, but it just feels to your point as like a series of like administrative points of entry and they're they're not connected. And I think if your monetization is messy, like your whole growth engine is pretty fragile. I've really kind of developed conviction that the order to cache process is really kind of a strategic growth asset. It's really like a core part of the growth engine that makes it work because a unified order to cache makes you go faster, reduces friction and also like leads to a better customer experience like you know I always come back to like what's better for the customer. It definitely leads to a much better customer experience and so you know I think billing in particular people don't view this as like a real product thing.
40:43 It's like always been like an engineering thing or a backend type of thing. But to me, like billing is one of the key touch points you have with your customer. They're paying you through that product. They're inputting stuff in the product. You know, if you don't have an agile, flexible tool, a well-connected tool that's interconnected, like there's nothing more than would piss a customer off of like, hey, like you're overcharging me like five times or I sent a payment in, it's not getting recorded. Like that's a terrible customer experience. And so like that's because like maybe you have a bunch of inter disconnected systems that no one knows about and are not talking to each other or you signed me up for a plan that I didn't sign up for.
41:20 Like those things happen all the time. So I think for us like one of the first things I did when I came on to gorgeous you know being the first CFO that was there like I took ownership of that order to cash and we kind of just ripped out the disparate systems and try to make sure they were all interconnected. And so we took a new contract intake tool for the sales reps, which like by the way at the end of the day like it makes them go faster, right? Like I want my sales team closing deals and not trying to input the minutia into some contract life cycle tool and like figure out what inputs to put in. Like that's not time they should be spending. The more we can standardize that, you know, we put in ChargeB as our billing tool.
42:00 Um and so now like we feel like we're in a much better place where things are much more interconnected. Oh, I've dealt with that on a micro level, even just for sponsorships for my podcast. My sales guy does not want to be following up for payment with people. He wants to be out there selling. Like, and me and him joke about them like, I don't want to be on you to have to do this. Like, I just want you to be crushing deals so I can pay commission and high-five you.
42:21 It's probably the the point fraught with the most friction potentially if you're not doing it right. So, you can screw up how the customer feels about your product, not because of their experience with the job to be done, but just how they have to pay you. You know what's really critical especially because we have a lot of SMBs as part of our customer base and when you kind of go through like even churn reasons we look at churn reasons it's like sometimes people just had some like really bad experience with like the way they were built or they got overcharged and sometimes you'd be like okay you got charged like maybe five or 10 bucks too much but like you know for a small business owner that could be meaningful and it's just like that experience of like hey I shouldn't have been charged for something you charge me for it or you didn't recognize some other thing that some saleserson told me X, it didn't get into the system and you're charging me Y. That's a terrible first experience to have as a gorgeous customer. So like I think the more we can think about billing and order to cash as a growth product, right?
43:17 Something that actually impacts our customer experience, you think about things differently. You make different architecture decisions uh as you plan that out. If I walked you into a $25 million AR SAS company tomorrow, what would you look at first or audit and order to cash process to see if hey, I think they can make it to 250 million. >> I would look at it with the angle of like where are the things that are going to explode as you 10x this scale. The first question I was asking are like where are the manual inputs into the system and like how many different points of inputs do you have? Like when I came to Gorgeous, there was literally sales reps manually inputting like contracts. There was no set forms. We would input a contract into our contract tool. Someone had to manually take that contract that got inputed and then input it into a billing tool and then that would get sent out to the customer and then someone had to go manually go chase those. A lot of manual inputs and so like that's the thing where you know I would really investigate like first off like where are all the the manual choke points. I think the key is I would spend more uptime work upfront architecting what you want your system to look like so that you can figure out what systems can all talk to each other cuz I think where the mistakes I see are people trying to find point solutions for every single part. Okay, I have a contract life cycle tool here and then it goes into my CRM and then I have some billing is completely separate or I build some homegrown integration between the two.
44:38 Like those things are not going to scale. For those listening at home and they may say, canel, I work at a software company. You can't touch anything. I don't have any inventory. How important can this be? What would you say to them? >> I'd say like, hey, listen, the ordered cash is really it's a manifestation of your business model. The critical stuff that no one sees, but it's the growth engine that's there. So, I get you may not have inventory, but it doesn't really matter, right? If you don't get the foundations right, the world is changing so quickly around this where you have to be agile. have to experiment with different pricing needs like customer needs could be different in 6 months from now like maybe we don't have an outcome based resolution model maybe it's some hybrid that we have I need to have the right infrastructure to change to do that which doesn't require me like six months of engineering work to go off and do that and so if you build the right infrastructure you're enabling yourself to respond to the changing macro environment and the changing customer needs so I'm just thinking through some of the overlook leverage points here inside of order to cache so you said manual input That's a massive one. You said looking for point solutions rather than architecting it as a as an entire system. Anything else come to mind before we transition?
45:48 >> Yeah, I would say kind of two things. One, and this is kind of maybe nerdy specifics, but like just product catalog, like really knowing what your product catalog is. I didn't really appreciate that, but then like every time we do our billing system implementation, it's like when you have real issues, it's cuz like you have this everflowing kind of product catalog. And then I think it's again about connecting the systems, right? When sales reps quote deals that goes right into a CR tool that or goes into a billing tool, you know, all these things are connected. That's going to make things just so much more agile. I want to transition to talk about learning density and career compounding, which are two amazing terms that you had thrown at me when we were preparing for this. And you've talked about optimizing for learning density rather than title early in your career. There are a lot of people listening right now who are just saying I only want the CFO title. I don't care what I have to do to get there. How do you think about that and trying to get learning density?
46:43 >> One of the most influential conversations that I had kind of like mid-career was like sometimes you got to slow down to move fast, you know, and there's probably a lot of type A people you have listening to your podcast and early on in your career and be like, I want the fastest, straightest line to get a CFO title that I can get, right? And I think that I'd say is like, hey, optimize for the learning density. And when I say like learning density, it's just like what is the maximum amount of learning you could have per unit of time? If you did this job for 12 months, would you be 3x better? The experiences, the things you've been exposed to, the people that you worked with, because those things like they fundamentally like just compound. And so like when you become a CFO, you want to be able to like be prepared, right? No one's going to have all the answers. No one I know any CFO has ever seen everything. You know, being able to have pattern recognition on certain things. Being able to have been exposed to different areas of finance. Like, you know, early on in my career, I went out of my way to jobs in treasury on spend times on go to market, on product finance, on accounting and doing all these different disciplines where, you know, if I was optimizing for the highest title as quick as I can, you know, I would have taken a different route, right? I would have done other things and like you know my my background is a little bit atypical where I've done a lot of different jobs in and around finance but all those skills have like really helped me now right because I have a varied experience and so that's why I always kind of tell the folks on my team when I'm doing mentoring or you know people are thinking about leaving I should really think about like what is the thing you really want to learn not the title you want or the comp obviously those things are important for people at certain time in life but really think about what's the set of experiences says that will help you the most 5 years from now.
48:29 >> If you take a look back at your career, is there a 12 to 18month window that jumps out at you like, "Oh that had a that had a lot there." >> I think the the time that was the most formative for me was the time at Survey Monkey that I spent when I joined Survey Monkey was actually kind of I guess quote unquote like a lateral move. I was at another public company 8 by8. I took the same title and the same role at Survey Monkey. And so a lot of people like why are you doing that? like why don't you just go find a higher title?
48:57 And the reason I did that was because I had the ability to learn from two incredible mentors. So Debbie Clifford was the CFO at the time. Justin Glumby was, you know, going to be my direct boss when I worked there. And when I interviewed with both of them, I was like, the these people can help my learning density by like five or 10x. Just to be able to understand how they work, how they think about this. I can see what great looks like. That's an incredibly important thing for people in this world to to see in in in any kind of job you want to do. But being able to work with people who are great, getting to kind of see greatness in day in and day out, that's a learning opportunity that like you should never pass up. The company went through some wild times from, you know, a very successful IPO from getting kind of almost bought by Zenesk for 4 billion, then going through, you know, a little bit of decline and going through a private equity sale. But the learning density I got from there, the types of experiences, the people, the exposures that I got, those things I could never have replicated. And yeah, I'm really grateful.
49:59 >> Shout out to Justin, now the CFO of Miro. >> Yeah, I heard him on your show before. So that was great. >> Hell yeah. Uh, all right, Canal, I'm going to take you into what we call our long ass lightning round. You're a successful CFO, but you got to give me one thing you've messed up on the job before. Could be this role or any other. I think because of the way I grew up in some pretty combative places like in banking and private equity, I used to listen with the intent of responding but not to the intent of listening. What I mean by that is like someone would kind of fire back for me in a meeting and I'd be ready to go with the response right away just to kind of combat and try to win the argument. And I think the maturity that I've had now being in this role is like listen to what people are saying, digest, don't have to agree. If you have a different intent to really understand what they're saying, it's okay to change your point of view. I've asked that question over 200 times in the show and that may be the best answer I've ever heard to that question. I've personally done the same thing where I'm downloading for a rebuttal, not to learn or to add to the conversation in an incremental way.
50:57 >> Yeah. I mean, I I still struggle with that a little bit, but that's been a little bit of the unwiring that I've had to beat over time. >> If you could tell your younger self something knowing what you know today, what would you tell them? >> The most valuable thing you can do is like make other people's life easier. So like figure out ways to do that. Be creative, be curious, like you know, there's tools available now that like we just didn't have for most of my career, you know, and so that if you can come back with some analysis or some insight, something proactive that just like that's, you know, I don't have to go and look or I don't have to write a headline or I don't have to like sit there and reformat stuff like little pain points.
51:34 If you just make someone's life easier, you're going to go a long way in this role, right? And so like just think about how you can always make the people's lives around you easier and that can be really important things like hey driving a lot of like really critical strategic analysis but it can also be just small things where you're doing things the right way that something else doesn't have to spend time on and you build that trust and once you can build that trust that's just so valuable to have.
52:00 >> Next one I got for you more of a technical one. Can you walk me through your finance software stack? What tools does your team use to get the job done? >> Like I mentioned Chargeby is our billing tool. We just moved over to them. Been been thrilled with what we've got from them. Next suite is our ERP. Pigment is our FPNA tool. Tropic is our procurement and uh vendor provider. And then you know recently, you know, our whole company has access to Chat GPT. We've been using Claude a lot now as most people do. And yeah, we've been building a lot of in-house kind of decision engine tools with my awesome uh data team and data engineering analytics team, which has been probably the single biggest productivity increase I've seen in 20 plus years.
52:41 >> Anything neat that you've built built recently that you can speak to? >> Yeah, so we built this internal tool called Cortex. I could probably talk to you for an hour about this. So the data team, data engineering and data analytics team, we we rebranded them into decision intelligence. They've built this internal tool called Cortex which basically the the very quick 30 secondond view is it's taken a bunch of structured data. So all the stuff from CRM tools, your customer success management tools, your financial systems, all these different structured datas, unstructured data like board decks, slacks and it's basically put all of this into a context layer. And now that context layer has different decision products that you can use.
53:22 decision products meaning like I can go in Slack and say hey at Cortex ask them something simple like hey how much AR did I have for the last 12 months by these different segments and it can give you kind of very quickly stuff that maybe take five minutes to look up in different months but now like just this morning I was like hey Cortex I've seen a bunch of churn happen over the last 6 months in this particular segment can you tell me the biggest movers go analyze the customer transcripts and costs and tell me like what the root causes of this were and suggest like some operational fixes and it just goes and like does all this analysis because it has all this unstructured data has all this context and now it's coming back and telling me like why these customers are analyzing the customer transcripts like analyzing all the different reasons that we have in our system the utilization kind of correlating the product usage to potential churn and contraption risk so it's been like just incredibly useful so now I think like 70% of our company is using this day-to-day >> that's amazing we're going to do a part two on that last one I got for Oh yeah.
54:21 What's the craziest thing you've ever had someone try to expense? >> You know, one time I saw someone try to expense like this like three star Michelin dinner with them, their wife, and then like someone else on top of that. They labeled it as like, you know, self-help. >> The wife wasn't a customer is what you're saying. >> And then they added a third person on there. And it was like, I mean, it was a lot. It was like, you know, 2500 bucks or something like that.
54:46 >> Three out for dinner at Nou. Canel. This has been an absolute blast. Thanks for joining me on the pod, man. Yeah, this is great. I had a lot of fun. >> On the Numbers is a mostly media production. Yelling, an intro by Fat Joe. Artwork by Meg Delisandro. Show is executive produced by Ben Hillman. Nothing said on this podcast is intended to be business or investment advice. It's the sole opinion of me, a guy who feeds his dog way too much ice cream and has a history of net operating losses, lol. If you like this podcast, hit subscribe and give us five stars. It will take like 2 seconds and our algorithm overlords love it. Drink water. Call your mom. And have a great day. Peace.
Summary
- Finance should act as the decision engine of the organization, focusing on navigating growth rather than just reporting metrics.
- The order-to-cash process is a strategic growth asset, crucial for enhancing customer experience and reducing friction.
- Emphasizing operational empathy, Agarwal advocates for CFOs to engage with cross-functional teams to understand their challenges and needs.
- Gorgeous employs an outcome-based pricing model for its AI product, charging only for resolved interactions, which aligns costs with customer success.
- The integration of AI in customer support is reshaping e-commerce, requiring a deep understanding of product and customer data.
- Agarwal stresses the importance of optimizing for learning density in one's career, valuing diverse experiences over rapid title advancement.
- Building a cohesive tech stack is vital for scalability, with a focus on interconnected systems to streamline operations and enhance customer interactions.
- Continuous monitoring of AI costs and infrastructure is essential for maintaining quality while managing expenses.