Meet the Founder Using AI Agents to Automate Billing

in this episode of emerging founders with manav I interview David Pang he's building the AI agent for B2B Billing at diim Ley and he also created another company called sheet we which has over 1.8 million shortcuts and Counting David is solving his own problem because when he was creating invoices in his past company he realized he was wasting a lot of time creating these invoices and with his company dimd he's automating that process with AI agents and in this podcast we talk about everything what he's focusing on how he raised money from investors and how you got accepted into [Music] YC David how are you doing today hey I'm doing great thanks for having me on the show so before we discuss timely I'm really curious about you as a person so I want to understand what was your prior Venture sheet whz can you explain what problem you were solving in that company yeah so SheetWiz is a Chrome extension that brings Excel shortcuts and functionality to Google Sheets so we take shortcuts as simple as you know inserting a row or in a column in a Google sheet or sizing a column or row and we bring those shortcuts to Google Sheets or the Excel specific ones we also support more complex functionality like you know Trace precedence or Trace dependence or autocolor or format cycling that exist in Excel plugins that finance and Consulting professionals use and we bring that over to Sheets because we know there's a ton of X Consultants X Bankers X big four accountants that grew up using or started their careers using Excel but now they're forced to use Google Sheets at the company that they're at and so that was my experience I started my career in Tech private Equity um before making the jump into a startup and when I made that jump and was forced to do Financial modeling in Google Sheets I lost maybe 80% of my productivity and so this tool that we created was a way to solve that problem got it so let's segue into the founding story of dim Ley what was the problem you saw in the market and can you explain in more simple terms what dim Ley does dim Ley builds AI agents for B2B SAS Finance teams and so there's a ton of kind of compli ated manual finance workflows that exists today in any Finance team and rebuild agents that are configurable by the finance team in order to solve their specific problems and so why did we build something like this well this was a problem that I faced personally at the startup that I was at called one signal I was there for over 2 years during that time both my co-founder Peter and I were building and running Billing at one signal for their Enterprise customers yeah we experienced the problem firsthand when we were at on Signal where we built and round billing for their Enterprise customers when I meant running billing it was basically me and a spreadshe doing a bunch of calculations quite manually and having to do that work I frankly thought I shouldn't need to be doing this this is not work that you know a human should be doing an AI could do this just as well and so when we you know became aware of these AI agents that you could you know this new technology was coming up back in early early 2024 late 2023 we really saw it as an opportunity for us to kind of use this new tool that we have to solve a problem that plagues almost every single company in the world yeah one problem I experienced personally was using QuickBooks and and creating invoices and QuickBooks was such a pain like creating the same thing again and again every day so with dim Le are you going to be integrated into tools like QuickBooks or what kind of tools would you be integrated into and how exactly are you helping people save time yeah so we launched our first AI agent back in August and that AI agent focuses on connecting where your contracts are stored and where you're sending out your invoices so exactly kind of the flow that you mentioned with QuickBook sending out invoices but your contracts are probably not being generated in QuickBooks it's probably generated in Salesforce HubSpot maybe it's getting signed in docy sign pandadoc iron flad various systems like that and our agent takes the PDF copy or maybe there's metadata available and we read it and understand it and then we decide how we're going to create an invoice for you in your end system whether that be stripe or netw Suite or QuickBooks you know charge B maxio whomever and so all you need to do instead of manually clicking around in a UI you can just review the changes that we're going to make hit approve and will make those changes for you in your in kind of your Billing System or system of choice we've observed that especially for B2B billing that is actually a very complicated and time-consuming process that takes 20 30 minutes for a single contract because these are heavily negotiated very nuanced contracts that somebody with Advanced knowledge of the system and tools needs to go in and read and that becomes a very overwhelming task especially at the end of the quarter or at the end of the month when you have you know a bunch of deals closing sales trying to their quota and suddenly you're the accountant and you get all of these pings and you're processing all of these and it's getting late at night and there's a bunch of other things that you have to do in order to close the books so that's kind of our first problem that we're solving is building an agent that helps with invoice configuration for contracts do you mind explaining what AI agents mean yeah so the way that I think about an AI agent is something that can take information from somewhere and understand it create a plan of action and then can execute on the on that plan and so whether that's you know pull contract read the contract extract the amounts and the products and then identify the products and the amounts in another system and then make an API call to create invoices like that would be an example of an agent inter do you mind talking about like the pricing structure and the business model of D it depends on the use case but I would say that in today it is maybe more of a direct sold motion and so we can determine that based on the clients that we work with and so we always like to think about the amount we charge as a Roi for the customer we don't want to be overcharging you for something that the value that you're extracting is not even equivalent to what we're charging we always aim to be between kind of like that 20 to 30% of the value that we're generating for you and so we always tend to do like an RI calculation right how much time are we saving you how much greater visibility to the tools are we creating for you are you collecting faster because you're able to send invoices faster are we you know enabling sales to sell more complex building structures because now our agents can handle the complexity versus a finance person spending a lot of time to handle that and so I know these things are not necessarily calculable but we always like to think about it from an Roi perspective but if that makes sense and I know maybe it's not as straightforward but the reality is you know when it comes to these pricing discussions we really want to make sure that the ROI is there and so just saying like we charge x amount of dollars per contract that we process doesn't necessarily do justice to pricing that we yeah the biggest thing I've realized is that essentially if you can replace like a whole employee which is anywhere from like $70,000 a year minimum or even more if you're in New York California how do you justify like replacing a one person with just like a subscription right even if you charge like 500 a month still like a great deal for the company that's why I feel like these AI companies they're just making more and more startups like more linear like I noticed even in my companies we used to have like 20 or 18 employees and now we're like two or three we've become so lean because of that and the biggest thing for me is like I was listening to the podcast of the repet founder and he was talking about they actually got rejected by YC four times before they got in we were also rejected by YC and then when they finally got in it was like the most insane like three months where they kind of like really went all in and like put insane amount of hours during those three months well how was your journey like getting into YC and how much work did you put in we were rejected for summer 23 we and we pitched some sheet whz at the time we weren't rejected from dimy but for Sheet whz effectively the feedback was this the chome extension how is this going to grow to be the next Airbnb stride whatever company at the time you know we had just launched we hadn't even monetized the product yet and so that was a little bit harder of sell necessarily and so we were rejected in summer 23 we then applied again in sum 24 with diely after realizing the problem and the the technology had kind of reached a point where we felt like hey this is this is something that we can go after and then with she Wiz at that time we had also grown it quite significantly we had started you know charging for the tool you know it's only $4.99 a month but at least that's what we were charging and we could go back to YC and say hey me and my co-founder we've worked together we're also co-workers at another startup we're also roommates we know each other very well we know our working Styles we built a business already we've monetized it and now we want to go after a larger opportunity and this is the opportunity that we're going after we want to use AI to do it and I think that was a much more compelling value proposition or much more impelling pitch to YC than necessarily sheet whz which I think at that time had maybe like 10 users cool so I want to ask you what advice would you give to other entrepreneurs starting a company I would say there's a lot of advice out there and sometimes it's not advice that should necessarily be taking for example I think you hear a lot about acvs for SAS and where they should be and there's like a dead zone because can you explain what ACB is average contract value it's like the size of the contract got it okay and you know if you're above 2K for example and you can't expense it with a credit card you know you're going to have to go through more but if you're below 20K it's harder to hire sales people they're taking a commission they're not really incentivized to sell cuz the amounts are a little bit lower you're kind of in that dead zone where it's too big to be expensed on a credit card and too small to be to incentivize anybody to sell it and I think you hear that feedback over and over again you need to be above like 20K be above 25k push those values up and the reality is that's larger company I think problem and that that's maybe a hotter take but start with the customers that are willing to pay for you that's like the first step right and if they're willing to pay for you there's more value that you can add on top but if you're just being like okay this idea isn't going to work cuz people aren't willing to pay like 20 25k or you're rejecting customers that that might be even worse if you're rejecting customers because they're not you know you're below a certain ACB when you're just getting started that's not something that you should necessarily be doing because customers is better than no customers and you can leverage those learnings by working with them to build something that is maybe more deserving of that contract value and so that's maybe one example of an advice that maybe you shouldn't necessarily take for face value especially if you're early stage startup that's great advice in my opinion I think a lot of people especially in the beginning they just want to see some traction or you want to see some win right away and I feel like getting that dope mean head from getting a customer might be the juice to keep the company going and I feel like that's that's such a great advice I love that awesome David thank everybody we had such a great talk about AI agents we will do another recap video in a year or two and kind of recap this podcast and see where our Journeys have led us to thank you so much David for coming on the show I really appreciate you yeah thanks for having me please make sure to click that subscribe button if you get any value from this share it with a friend that's all I ask for thank you so much for watching I'll see you in the next episode