Founder to Founder: An Honest Conversation on AI, Capital, and What’s Actually Working in Ecommerce Finance

Webinar Transcript

Introduction 

Stephen Brown: Welcome to today’s webinar. I’m Stephen Brown, COO at LedgerGurus, and also a co-owner of a small DTC brand. With me today is David Segal from Highbeam. David, would you like to introduce yourself? 

David Segal: I’m the President of Highbeam, but before that I was actually a Highbeam customer. I built a consumer brand called David’s Tea from zero to $200 million in sales, took it public on the Nasdaq, and sold it. While building other brands afterward, I got fed up with my bank — zero percent yield on cash, no intelligence on the data, clunky software, bureaucratic service. That frustration led me to try Highbeam as a customer, and I was blown away. 

David Segal: I ended up becoming President. We’re a cash management platform for consumer brands doing anywhere between $1 million and $200 million in sales. We’ll process about $15 billion in flows through our pipes this year, with well over 600 customers and growing about 10% month over month. 

Stephen Brown: Awesome. Today we’re going to dive into a wide range of topics — ecommerce finance, tools, and capital needs — drawing on the perspective of two people who have owned or currently own brands. David, let’s dive in. 

David Segal: Sure. You want to walk through the agenda quickly, Stephen? 

Stephen Brown: Yes. We’re going to cover three topics: what’s going on inside ecommerce brands today, the true cost of capital, and how AI is fitting into finance. We’ll take questions at the end, but feel free to ask as we go. 

 

What Ecommerce Brands Are Up Against 

Cash Flow as the Core Challenge 

David Segal: We pulled data from our CRM across thousands of brands to find their biggest challenges, and a clear cash flow story emerged. It all comes back to visibility — understanding cash flow needs, identifying where gaps will occur, and knowing how to finance them. That starts with inventory visibility, which I know LedgerGurus goes deep on. Understanding inventory buys, cash flow needs, and accessing the right financing are top of mind for most brands. 

Stephen Brown: Exactly. The second point — inventory and COGS tracking — may not seem like a direct money issue, but I would argue it causes cash flow issues. It’s directly related to everything else. The long cash cycles of smaller brands, the time it takes to buy inventory — these are what create cash flow problems. For services businesses like us, cash flow isn’t usually a major issue. But with product-based businesses, it tends to be a huge challenge. And over the last couple of years, market shifts have made it more challenging than ever. 

Stephen Brown: David, when you were at David’s Tea, how significant was cash flow for you? 

David Segal: Massive. Brands can be profitable and broke at the same time. If you don’t turn your inventory, what would happen is we’d get excited about a SKU without really understanding how much money we had available to go shopping. The brands that buy at a world-class level start with what’s called an open-to-buy — essentially, how much can you afford to spend? It takes into account your inventory position, sales, receipts, and months of inventory on hand. It becomes very directional in determining what new opportunities you can buy into and what stale inventory you need to promote and clear. It wasn’t until we were consistently turning inventory at least three times a year that we were not only profitable, but had free cash flow as well. 

 

Poll Results: What Brands Are Struggling With 

Stephen Brown: Based on our audience poll, the top issues closely matched what we see in our data. Cash flow and working capital, and cash flow forecasting and visibility, were the top two concerns — closely followed by inventory and COGS tracking, and access to financing and credit. 

 

Levers for Improving Cash Flow 

Supplier Negotiations 

David Segal: Constraint number one for most brands is cash, not demand. You have product-market fit, but you can’t get the terms you need from your vendor. I’ll share a story: I was talking to a successful mid-eight-figure founder who had been in business for nine years and was very profitable. He mentioned he couldn’t afford a nanny despite having a new baby. When I asked about his buying terms, he told me he had to put 30% down at the time of order, wait 90 days for production, pay 60% when goods left the port, and 10% on arrival. He had a negative 90-day buying cash conversion cycle. I told him that if he could negotiate to 60-day terms with his vendors — showing his track record and lining up alternative suppliers as leverage — he could potentially free up $1-2 million to take out of his business. The number one way to improve cash flow is to negotiate hard on buying terms. 

Stephen Brown: That’s a great point. When you’re addressing cash flow, there are many levers to pull. People often go straight to lending, but you want to exhaust all other options first. Working with your suppliers is one of the most powerful and underutilized. 

 

Other Cash Flow Levers 

David Segal: Stephen, what other levers do you see? Obviously supplier negotiations is a big one. The right working capital line of credit — ideally not financing working capital through equity, because that dilutes you dramatically. Utilizing idle cash effectively? 

Stephen Brown: Using credit cards wisely can buy you cash flow. Another big one — and it ties directly into inventory tracking — is simply ordering the right amount of inventory. That’s a skill that takes multiple reps to develop. A lot of brands got burned during COVID when just-in-time inventory disappeared. But there’s real value in inventory optimization. 

David Segal: This is probably the main point of the entire webinar, and it was your number one identified issue. Jason Panzer of HexClad put it best on a panel: many businesses are heavily Q4-skewed. The hay is in the barn — you’ve already bought the goods and have a plan for what you’ll sell. But sometimes we make the mistake of reducing ad spend to chase profitability, forgetting that it has a major cash flow implication. If you don’t sell as much as planned, even if each sale is more profitable, you can end the season with far more inventory than projected and a lot less cash, right when bills are coming due. Sometimes when the inventory is there, you have to make sure you’re pushing the right SKUs at the right time — turning inventory into cash, even at the expense of slightly lower margins. 

Stephen Brown: And with multi-SKU businesses, it’s very easy to let zombie SKUs just sit on the shelf. Inventory management has many layers — not just ordering the right amounts at the right time, but also not letting slow-moving SKUs build up and consume cash. If inventory isn’t moving, it’s cash that’s locked up, consuming precious resources that could be deployed toward faster-moving SKUs. 

 

Inventory & COGS Tracking 

David Segal: Inventory and COGS tracking — it’s your biggest asset and often your biggest blind spot. Stephen, your firm goes deeper on inventory than most fractional accounting firms. Talk a bit about your approach. 

Stephen Brown: Inventory is genuinely complex for companies in the seven and even eight-figure range, because it requires solid operational processes — it’s as much an operational exercise as a financial one. We work with clients to help them achieve better financial visibility on their inventory, but to do that we have to work on their operational visibility first. There are many places where operational shortcuts or sloppiness create data problems: how you receive goods, how you handle returns, how you track items sent to influencers for marketing. All of that adds up to inaccurate inventory data, which creates inaccurate financial data. What we try to do is help people understand the processes they need and the tools they can use to build that discipline. Once you have that discipline, forecasting becomes much more accurate. But until you build that operational muscle, there are many reasons your inventory and financial data will be off. 

 

Cash Flow Forecasting & Visibility 

David Segal: We’ll talk more about this in the AI section, but Highbeam just launched Luma, our AI engine, anchored on producing a credible, real-time cash flow forecast that you can interact with. Getting a clear, live, accurate cash flow picture with a solid forecast is the foundation for understanding where money is going in your business. 

Stephen Brown: Cash flow is tricky to forecast because so many inputs feed into it. When I work with new customers, I ask about their revenue plan, and that requires discipline. That revenue plan drives a marketing plan, which drives inventory planning, which all feeds into a cash flow plan. A lot of times, when someone asks me to forecast cash, the foundational inputs haven’t been built yet. There’s a lot of muscle that needs to go into it. That’s why I’m excited about what Highbeam is doing — these aren’t easy things to build, and the more you can automate and streamline, the better projections people can make. 

David Segal: It reminds me of the Mike Tyson quote: everybody has a plan until they get punched in the face. You put together a cash flow in Excel, and then the show starts. You always buy too much or too little. You’re constantly adjusting your marketing and inventory plans. What we’ve solved for in Highbeam is capturing all of those micro adjustments happening day by day, updating the model in real time in a conversational way, and keeping it live and current. 

Stephen Brown: There’s a great quote from statistician George Box: “All models are wrong, but some are useful.” You plan, nothing goes according to plan, but if you have good flexible plans, they can guide you as things shift. 

 

Data From LedgerGurus Customers 

Stephen Brown: We did a study a few months ago across our customer base because there’s a lot of analyses out there, but we can only trust data we’ve overseen ourselves. A few things stood out. Leaner COGS clearly led to better outcomes — no surprise. We saw gross margins being compressed by tariffs and inflation. Side note: if you aren’t pursuing tariff refunds, you really should be. Top performers had more cash, which creates opportunity and a virtuous cycle. We also saw a lot of wholesale among top performers, though wholesale alone isn’t a guarantee — you have to execute it well. Overall, the data reinforced best practices we already know: good gross margins lead to better profitability, which creates more cash. 

 

The True Cost of Capital 

Merchant Cash Advances: Understanding the Real Cost 

David Segal: You all identified working capital funding as a key concern, and this is a topic that fires both Stephen and me up. Merchant cash advances are widely misunderstood. The fee they quote you is not an APR — it’s a flat fee. That’s where the confusion begins. 

Stephen Brown: I’ve seen customers get into really bad situations from predatory loans. Part of the challenge is that when you’re small, you often can’t get a decent business line of credit from a bank — even brands doing $10 million or more per year sometimes can’t get bank lending. It takes time to build those relationships. So founders chase what looks like easy funding. These merchant cash advance lenders are like the payday lenders of the ecommerce industry. They don’t communicate real costs and they frame their fees in ways that obscure the true interest rate. When you see “10% fee,” you think it’s not too bad — but when you do the math, it’s dramatically worse. 

David Segal: Let me explain why. The fee is constant — it doesn’t decrease as you pay down principal. So you’re paying the same fee on day one as on day 90, even though your outstanding balance is dropping. A 10% fee on a four-month payback works out to a 42% APR. That’s predatory. Highbeam has a true line of credit — no fixed fees, no origination fees, true APR. You only pay for what you use, calculated daily on outstanding principal. You can pay it off at any time without penalty. It’s transparent and significantly less expensive. We have a calculator at highbeam.com/APR that lets you plug in your flat fee and payback period to compare what you’re paying versus what you’d pay with Highbeam. 

Stephen Brown: That calculator is a really great tool. One important nuance: merchant cash advance lenders look at your sales velocity and project a payback period around that. The faster you’re actually growing and selling, the faster you pay it back — which sounds good, but it makes the effective APR even worse, because you’re paying the same flat fee over a shorter period. Easy money is not always good money. The hoops involved in getting lower-cost capital are well worth it. 

David Segal: To Stephen’s point — at Highbeam, we can give you an answer in 48 hours. You connect your QuickBooks and we can tell you where you stand. And if you want to compare costs, go to highbeam.com/APR. It gives you an apples-to-apples comparison of merchant cash advances, asset-backed lines, and term loans. The sooner you pay back a merchant cash advance, the more expensive it actually is — and the calculator will show you exactly that. 

Stephen Brown: I’d especially encourage you to look at the bottom section of that calculator — it covers penalties and early payback fees. Those finer details can come back to bite you. 

 

AI in Ecommerce Finance 

The AI Landscape for Brands 

David Segal: AI is all the hype right now. We’d love to talk about what we’ve done for ecommerce finance — what’s signal and what’s noise. Stephen, what are you seeing from your customers? 

Stephen Brown: People are all over the spectrum — from still figuring things out to trying to go all in. Ecommerce founders tend to be tech-friendly. The challenge is understanding what’s genuinely good versus what’s just noise. We’ve been running experiments internally — feeding financial statements into various tools and analyzing the output. We’re seeing real improvements, but there’s still a judgment layer required. You have to apply experience and context on top of what the AI tells you. The hallucination concern is real. That said, I’m seeing accelerating improvements that are genuinely expanding what we can do in normal business processes. 

 

How Highbeam Built Luma 

David Segal: Let me talk about what we built with Luma, which is the AI engine powering Highbeam — officially launched today, after over a year of development. What we found through that process: first, when you just dump all your data into the latest model, three things happen. One, the models aren’t great at being deterministic and maintaining context. The same prompts on Monday may not replicate reliably on Friday. Two, you end up spending a fortune on tokens because it pulls everything into the context window and uses the most powerful model for even basic tasks — like taking a Ferrari to the grocery store. Three, it struggles to produce an auditable record of what it actually did. 

David Segal: So we took a different approach with Luma. Our advice for brands: pick one painful workflow, prove it out, then expand. Start with forecasting — the most common need across cash flow reporting. Automate that workflow while keeping humans — including experts like LedgerGurus — in the loop. Make sure you’re in a secure environment where your data is protected. And make sure whatever the AI does is auditable, especially with financial information. 

David Segal: Luma connects with QuickBooks, NetSuite, your banking, cards, and inventory data. You can interact with it in the Highbeam app or in Slack. A key architectural decision: Luma actually writes code when it executes a query. So when you ask it to do something, it writes a repeatable process — not a one-off response. It has the context of $15 billion in consumer brand flows through Highbeam’s pipes, while keeping each customer’s instance unique and secure. 

David Segal: Luma tracks every update — who made it and when. When you look at a forecast, you can see how it built it: did it do back-testing? What was the confidence interval? Did it have enough data? You can interact with it in Slack to answer questions around open-to-buy reporting, SKU sell-through, capital needs, comparing different capital offers, debt vs. equity considerations, money movement, borrowing decisions, idle cash optimization, live P&L, and reconciliations. 

Stephen Brown: There are two schools of thought around AI: general AI tools will handle everything, or focused tools will use AI more effectively within a specific domain. I’m firmly in the second camp. What Highbeam is doing — taking general AI capabilities and applying them in a focused, expert context — is the right approach. A general AI tool may or may not give you what you want for finance. A purpose-built tool trained on your data and use cases is a different proposition entirely. 

David Segal: Let me use an analogy. You can make your own ice cream — buy all the ingredients, read the book, get the machinery. Some of you will produce great ice cream and become power users who could practically work as engineers. But most won’t. And you’ll eventually realize your ice cream costs more, doesn’t taste as good, and took far more time than the artisanal shop on the corner selling $3 cones. Highbeam is that ice cream shop. We have ~25 MIT-caliber engineers spending every waking hour perfecting Luma, training it across our entire customer base, cross-pollinating learnings, staying current on models, and knowing which model to use for which query to reduce cost. As a founder playing with it over the past couple of weekends — the quality is genuinely better than what I get from Claude directly. I’ve had real “holy shit” moments around complex queries: margin breakdown, inventory analysis, cash trough projections, understanding when and how much I’ll need to borrow. 

David Segal: We call it “success by a thousand cuts versus death by a thousand cuts.” Most founders didn’t go into business to optimize small cash flow decisions — they went in because they have a product the world needs. Luma puts those small financial decisions on autopilot. When founders lack clarity on financial data, they make lots of small suboptimal decisions that individually aren’t huge, but over the course of a year add up to tens or hundreds of thousands of dollars. Luma gives founders clarity, helps LedgerGurus get more leverage on your business, and automates money movements to borrow smarter, maximize idle working capital, and pay bills in a way that optimizes your free cash flow. 

Stephen Brown: Awesome. I’m excited to see how this evolves now that you’ve launched it. 

 

Q&A 

AI and Legacy SaaS Platforms 

Stephen Brown: Question from the audience: which legacy SaaS vendors do you see integrating AI in ways that allow small businesses to automate repetitive tasks without the expense and risk of building bespoke LLM solutions? 

Stephen Brown: I’ll take a general stab at this. Before joining LedgerGurus, I spent 20 years in enterprise software. Everyone is doing something in AI — if your SaaS vendor doesn’t have an AI roadmap, that should concern you. There are a lot of new “AI-first” and “AI-native” solutions emerging. But the devil’s in the details. Software isn’t just features — it’s supportability, velocity, long-term viability. If one engineer built your great AI integration and then leaves, can someone else maintain it? My first question when evaluating AI-powered software is always: who can support this, and for how long? That said, I think the vendors that are genuinely leveraging AI effectively will roll out capabilities faster than ever before. We’re in a transitional period and haven’t seen things settle. But the leaders will pull away. 

David Segal: You’re right. You can spin up something flashy really quickly. We learned this early on. But it fundamentally won’t stick if the underlying infrastructure isn’t right. It’s like an iceberg — the surface can look great, but if the contextual layer, the harness that delivers higher-quality answers at lower token cost, the ability to create repeatable workflows — if all of that isn’t solidly built, you can’t build on top of it. 90% of the value gets created in the last 20% of the race. The good news: everyone on this call has access to free trials. For us, you can connect your data in about 10 minutes, try Luma for 14 days, and make your own judgment. We’re putting it into market because we’ve tested enough to know the wow moments are more likely than the disappointments. Try it. If you’re a tinkerer, you’ll go far with this stuff. 

 

Inventory Management: QuickBooks vs. Dedicated Systems 

Stephen Brown: Question from Arthur: do you run inventory tracking through QuickBooks Online for clients? 

Stephen Brown: Short answer: no. QuickBooks Online’s inventory management capabilities have not been adequate for ecommerce businesses. Inventory management for ecommerce is extraordinarily complex — you need a system of record that provides real-time available inventory to all of your sales channels, with integrations to those channels. Intuit has not built what our customers need, though they are investing in it. The things I look for in an inventory management system are: depth of capabilities, how they handle different use cases, integration quality, reputation, and — increasingly — their AI roadmap and how quickly they’re executing on it. That last point is going to be make-or-break for software vendors. 

 

When to Move to NetSuite or an ERP 

David Segal: We often see people move to NetSuite and then migrate back to QBO. They go through these complicated integrations with NetSuite only to find it has more horsepower than they actually need. How do you think about when to make that transition? 

Stephen Brown: NetSuite is very robust — it’s become the ERP of choice for many vendors since Oracle acquired it. Oracle is aggressive and tends to sell down-market, often pitching ERP as the solution to inventory management by putting everything on one platform. But here’s what you need to understand: when you go to an ERP, you’re moving everything — financial software, inventory management, warehouse management. ERPs are very expensive to implement — often tens or hundreds of thousands of dollars — very expensive to operate, and require dedicated expertise to support. They’re also designed for complexity, which means usability suffers. Three clicks in QuickBooks might take 10 or 20 in NetSuite. Our general opinion: put off moving to an ERP as long as possible. The flexibility and speed of best-of-breed systems — a dedicated inventory management system, maybe a separate WMS — outweigh the pain of integration for most brands. As a rule of thumb, I wouldn’t even consider an ERP until you’re at $10 million in revenue, and often $20 million. Don’t let an Oracle salesperson convince you to buy an ERP at $2 million in revenue. 

David Segal: When we moved to ERPs at David’s Tea — Microsoft Dynamics and an inventory system called Just Enough — it was a nightmare. Nobody loves their ERP. I’ve never heard anyone say they love QuickBooks or NetSuite. None of them have truly solved it, because it’s genuinely hard. It always takes twice as long and costs twice as much to implement as you expect. And I actually think AI is going to start converging some of these systems — making it possible to extract the information you need without a formal ERP. The longer you can wait, the better. 

 

WebGility for Perpetual Inventory 

Stephen Brown: Another question: thoughts on WebGility for perpetual inventory tracking? We have worked with WebGility. Unfortunately, we were not able to get it to accomplish what we were trying to do, though it’s been a while since we last evaluated it. So don’t completely discount it based on our experience — things may have changed. But that was our experience. 

 

Trying Luma for Free 

Stephen Brown: One more question: can we try Luma for free? 

David Segal: Yes! There will be a trial period. There’s a waitlist right now — we literally just launched today and have seen incredible demand. Go to highbeam.com and look for the Luma link, or check my LinkedIn. There will be a 14-day trial where you connect your data and use it. I highly encourage you to try it. We could talk about it all day, but there’s nothing like having that wow moment yourself. 

 

Closing 

David Segal: You can reach me directly at david@highbeam.com. At Highbeam, we offer cutting-edge technology combined with old-fashioned banking service. We’ll treat you the way JP Morgan treats Microsoft. Email me directly — I’ll connect you with the right person on the team, and you’ll have our cell numbers and access to the whole team. 

Stephen Brown: Highbeam focuses purely on banking and finance for consumer brands. We focus purely on accounting and finance services for consumer brands. There’s something to be said for that kind of specialization. Thank you everyone for joining us today, and thanks to David. We hope to see many of you in the future. Take care.