Joe Fox: Hey everyone, thank you so much for tuning into another episode of Retain Grow Thrive, the Growave podcast. I'm your host, Joe Fox, president of Growave. Today I'm super, super excited. We have an absolute e-commerce industry veteran, someone who is very, very well-versed in high-end e-commerce, has been a part of this industry for a very, very long time, worked with some amazing brands like the LVMH Group, knows a lot about AI, machine learning, et cetera. So I'm really excited and I think there's going to be a lot of value that the audience can derive from this episode. So Mike Le, welcome. You've got some really cool products that you work with. You've got a really cool agency. So tell us a little bit more about you. It's going to be hard for me to keep this under 30 minutes, because there's so much I want to dive into. But Mike, can you tell the audience a little bit about you and what you're up to?
Mike Le: Sure, thank you, Joe. I'm very happy to be here. I'm Mike Le, co-founder of the performance marketing agency CB/I Digital, and the inventory AI startup Conative AI. I'm an immigrant entrepreneur — I grew up in Vietnam and came to New York in 2005 to get my master's degree. I randomly met my co-founder on the F train in New York, we talked, became good friends, and decided to start a company together. That's CB/I Digital, which we've run together for 18 years now — since the early days of SEO, the early days of Facebook ads, all the way to the AI era. We've scaled to over 100 people in our offices in New York and Vietnam, working with both enterprise clients and a lot of e-commerce brands. In the last three years we've gotten about 15 industry award recognitions, including the U.S. Agency Award 2025. And three years ago I started my second company, Conative AI — we aim to reduce overstock and understock inventory for retail brands and optimize cash flow using the latest deep learning AI technology to replace the traditional forecasting approach.
Joe Fox: That's amazing. I love diving into this because, as you know Mike, I owned an agency in Australia before moving to the tech side. My first role on the tech side was for a Scandinavian AI/machine learning company using big data sets to optimize Google ads and business intelligence decisions for companies like Toyota. So I feel like there's a lot of crossover here. And also like yourself, I'm an immigrant to the U.S. more recently, chasing the American dream too. Congratulations on the plethora of awards, by the way — we're lucky enough to work with Ari on a regular basis, and I want to extend that congratulations to you and the entire CB/I Digital team, because the agency space is super competitive and to be recognized for that is really cool. I want to unpack two main things today — some stuff around CB/I, then move into Conative and why that's important, and maybe at the end get your predictions on agency and AI. What are the main things CB/I is focusing on for merchants at the moment? Is there a particular use case you'd like to talk about?
Mike Le: For 18 years we've been a performance marketing agency, and we've solved two things: creating high growth for merchants, and turning around decline situations. Most brands come to us with major challenges — I rarely get easy brands in the first days. In the good days, anybody can run your digital campaigns, but in the tough times you need to really know what you're doing and what creates results. Brands and merchants usually have budget, have ideas, but aren't always sure which idea will work. To allocate budget to the right place, you need to dig deeper into the data to find the root cause of what's slowing you down or what the levers for growth are. At CB/I we ensure the brand only spends money, resources, and time on the ideas that give high returns. As a marketer, you might have 20 ways to spend money, but there will be 3 ways that give more returns than the other 17 combined — the question is knowing which 3. To give an example: we worked with a brand called Melinda Maria for about four years now. When we first signed on, the brand was in decline — it had lost more than 50% of revenue in less than a year after a very high growth period, because the structure and operations weren't ready for that scale. We came in as an agency, stabilized the ship, optimized the digital marketing funnel, and over the next three years took the brand from under $10 million to over $30 million in online sales. We've achieved similar growth with various other brands.That's one use case on the agency side. There's another story I want to share about the transition from SEO to AIO and AI search.
Joe Fox: Yes, I love this — it's super topical and blends both.
Mike Le: We worked with a client, First Citizen Bank — not a merchant, but a bank — on SEO for about five years. We got top ranking for keywords like "open online checking account," a very competitive keyword in the financial space — ranking higher than Chase, Wells Fargo, and other leading banks. But in the last two years, our effort shifted to optimizing cost of retrieval and entity schemas so AI engines can pull our information into AI answers. We did that for two years, to the point that the bank now appears in various AI engine answers when you ask about banking services. So this year, when bank leadership started paying attention to winning on AI, we were already there. It's a must for us to stay a step ahead and pay attention to the imminent trends that drive success.
Joe Fox: That's so incredible. I have a lot of agencies come on who focus on LLM optimization, but to be doing it with a merchant of that scale and be so far ahead of the curve is amazing. What is one thing you'd recommend to smaller Shopify merchants right now to focus on?
Mike Le: AI is so broad, it touches every part of merchant operations — but it's critical to know AI today is table stakes. Getting on AI is no longer about gaining competitive advantage, it's about getting in the race just to stay in business, because everybody is using LLMs like ChatGPT, Claude, Gemini, as well as tools like Nano Banana and Veo to create copy and creative. Everybody's using AI for productivity, ideation, and speed to result. As an agency, every week we release two AI agents specializing in a part of the expertise we deliver for clients — our SEO team, our PPC team, use vibe coding to build AI apps to execute specific digital marketing tasks, without needing support from developers at all. It's all non-technical people using ChatGPT to write code for Google AI Studio to build an app, then refining it. That's mandatory now — we have to change not just the implementation, but the mindset of letting go of what drove success in the past, and embrace having AI agents execute the job and AI automation do the work, so people can focus on strategy and thinking. AI can get us data, but analyzing the data to know the root cause and figure out where to focus to drive growth is still on us as strategists — I don't think we can trust an AI agent to do that yet. So I tell our team: if your value is hard work, you're not ready for the future. AI is more hard-working than you, less drama than you, doesn't ask for a raise. We all need to become thinkers and force ourselves to deploy AI in the majority of the execution work. Once you overcome the initial hurdle, it becomes a flow and deploying AI gets easier and easier. Getting on AI is the race to survive, not the race to get ahead.Now, most people use AI for marketing efficiency, productivity, and ideation. At CB/I, we also want to use AI for operational efficiency — inventory efficiency — which a lot fewer people are focused on. The key to a brand winning in the future isn't just marketing superiority, it's also better cash flow and more efficient operations. That's why we built Conative AI — to use AI to optimize inventory efficiency the same way we use it to drive marketing growth.
Joe Fox: Absolutely, there's so much to unpack there. I remember from my agency days doing SEO — it was consistent output, trying to get rankings, create content, build links — the sheer amount of human input was massive because there's only so much writing a person can do. AI is incredibly hard-working because it takes care of a lot of that. I want to stay on the inventory planning topic — you can't just have the best marketing, the highest traffic, the highest conversion; you need to think operationally about how things are managed on the back end. Tell me more about the product — how was it conceived? What are the core components that make it so valuable for the customer?
Mike Le: I first want to share why I created this product. As an agency owner for so many years, I saw a problem that becomes a very big problem but very few people perceive as a problem: the disconnect between marketing and inventory. I've seen marketing campaigns flop because they promote a product that's low on stock and stocks out very quickly — so all the advertising dollars go to waste, or get redirected to sell the rest of the portfolio that isn't performing as well. On the flip side, inventory teams often buy a lot of stock on certain SKUs that can't sell, move very slowly, and ask marketing to spend ad dollars to move that stock — but it's overstocked because people don't want it that much, so you can't spend your way out of it. That disconnect between marketing and inventory creates a lot of inefficiency and hurts brand profit.I also realized that a lot of growing brands — in the $10M-$50M range — still rely on spreadsheets for inventory buying decisions. You can have an ERP, you can have a lot of tools, but then you export everything to a spreadsheet and manually pull it together to make decisions. That's a cumbersome problem creating pressure on the inventory team and inefficiency for the marketing team. So I created Conative AI — first, we built the first inventory AI agents to optimize the inventory planning and buying process. We built a data engine that pulls data from Shopify, marketing channels like Google and Meta Analytics, and ERP sources on inventory and purchasing. We use deep learning technology to do demand forecasting — more accurate and lower risk than the traditional rules-based approach or early machine learning approaches. Deep learning does better forecasting with time-series data, which is great for e-commerce and Amazon sellers because those channels have real-time data to train the models. Second, deep learning can factor in a lot more variables than the traditional approach, which only accounts for a few key factors — so we can solve outlier cases that made forecasts inaccurate in the past. For example: if last year you sold a thousand units of a SKU, but you were actually out of stock for 15 days that month, and this year you have stock the whole month, you can't just rely on last year's numbers for planning. Traditionally there's no way to feed that information in — you rely on memory. With deep learning, we have a historical stock-update feed as part of the model training, so we can get much closer to the truth in forecasting. We also connect marketing signals into the forecasting process — if you triple your ad spend, demand should be higher. Traditional inventory tools don't have marketing signals or a way to account for that. With Conative, we aim to provide better forecasting and inventory optimization by solving the problem at a deeper level.
Joe Fox: I love that. I have to say, Mike, spreadsheets do not like you, but buying planners definitely do — that's an incredible gap in the market you've filled. I can only imagine the tens or hundreds of millions this has helped brands save. And it's not just saving money, it's reducing wastage too, which is great from an environmental perspective — when a SKU nobody wants doesn't move, it often ends up as wastage. What's on the roadmap for product development? Will you lean further into marketing insight, or further into logistics and shipping?
Mike Le: For Conative AI, it took me over two years to build out the model and test its reliability and accuracy to the point where brands can trust it. We also built a risk control layer so that even if a brand blindly trusts the AI, it won't put them into any major risk — the algorithm is very strong, I'm confident about that. What's ahead, and what I'm most excited about, is that we're now launching the full AI agents model for the tool. We're moving from the data UI, to the visual UI, to the agent experience — so you can have an agent handling your inventory almost autonomously. It can flag any risk you need to pay attention to, create a PO for you, show you what changed in sales for a specific portfolio, and do the planning for you. The AI forecasting model we built is proprietary, but we leverage LLMs for the agent layer itself — so we can feed the forecasting data and actual data into the AI agent so it can learn and execute the task for brands, for the inventory team, and for the marketing team. That's what I'm most focused on and most excited about right now.
Joe Fox: That's amazing, my goodness. I'm so impressed. I talk to a lot of founders launching AI-focused products, but this conversation is incredible. I have a lead in mind, actually — that's a good sign. On that note — it's difficult for me to stay under 30 minutes because I could talk forever with you, this is fascinating. I definitely want to do a part two. This feeds into my next question: with your experience across product, agency, and e-commerce in general, working with amazing brands and big banks — what are your predictions for e-commerce in 2026?
Mike Le: It all comes down to how effective we are at acquiring new customers. The landscape isn't as easy as it used to be a few years ago — it's more competitive, and consumer demand is somewhat lower than before. On average, brands often plateau and have to work really hard to acquire new customers, but great brands keep growing — it comes down to how effective they are at acquisition. There are a few directions brands take: being more efficient with paid channels, affiliate, advertising — or expanding into new marketplaces like Amazon or TikTok Shop to find more users, hopefully more cost-effectively. Across all those choices, Meta and Google are still the best-performing channels overall — I've tried AppLovin, TikTok, all the other channels, and as of now Meta and Google still consume the biggest budget most efficiently. But one thing to remember: we shouldn't let the algorithm run our entire campaign, or trust it blindly — algorithms aren't always fair and can be biased at times. We need to take back control of how we structure campaigns and where we allocate budget, to focus spend on the products and channels that actually work. At CB/I, we're big believers in the connection between product inventory and marketing — we want to allocate the majority of spend on the SKUs that can move fast, because that's how you acquire new customers effectively. We need to leverage the platform algorithm to find users, but take control to make sure budget is spent on the right SKUs, the right audience, the right creatives — because algorithms don't always do that for us automatically.By moving into other marketplaces like TikTok Shop or Amazon, one thing to pay attention to is that those platforms require a completely different mindset than DTC. The consumers on Amazon are not DTC consumers — what makes you successful on DTC will not make you successful on other platforms. You have to learn that platform from scratch and apply a new set of best practices. That's the trap — if you're not careful, it's easy to apply the DTC mindset to another platform and underperform there. So the trend all comes down to better new-customer acquisition — the prospecting game — and we're really good at that.
Joe Fox: I love that. As marketers and e-commerce owners, we need to evolve our thinking to match where the potential new audiences are — that's a very good point. What's the best way for the audience to connect with you, hear more about what you're up to, your insights? LinkedIn? The CB/I website?
Mike Le: I'd love to connect with everyone via my email, mikele@cbidigital.com — happy to reach out and share ideas that could hopefully be helpful. Also reach out to me via LinkedIn — my handle is mikelecbi, so it's linkedin.com/mikelecbi. I love staying in touch with people.
Joe Fox: Amazing, Mike. Thank you so much — I'll make sure all those links are in the show notes. I'm absolutely blown away by this conversation. I very much look forward to doing a part two — really digging into predictions and customer acquisition, we could almost do an entire episode just on that. Thank you so much for everything you do for the industry, for coming on today, and for the partnership we have with CB/I Digital. I look forward to continuing that.Mike Le: Thank you so much, Joe. It's been so much fun.Joe Fox: I really appreciate it, thank you Mike. Audience, thank you so much for tuning into this episode of Retain Grow Thrive. Please connect with Mike on LinkedIn, shoot him an email, check out CB/I Digital and Conative AI — all those links are in the show notes. If there's a particular guest you'd like to see us have a conversation with, mention them in the comments and we'll do our best to get them on. Talk to you soon.
