AI for Business Leaders · 2026-03-17 · 12 min
Key moments - from our scoring
Substance score
58 / 100
Five dimensions, 20 points each
Ministry of Supply, a direct-to-consumer workwear brand founded by Aman Advani, has fundamentally restructured its operations around rapid experimentation and AI-driven decision-making. Facing simultaneous tariff pressures from manufacturing across Mexico, China, and New York, the company pivoted from traditional inventory holding to just-in-time production with partners like Taylor Industries, while building a culture where tests can move from concept to results in 10 days rather than months. The key organizational shifts include single-point ownership (eliminating committees) and sequential rather than parallel execution, compressing 20 hours of traditional work over six months into seven days. On the supply side, Ministry of Supply built its proprietary Ministry Production System using AI to place orders at the SKU level twice daily rather than twice yearly - solving what would otherwise be an impossible manual administrative task. On the consumer side, the company uses Lovable for landing page generation, Arido for audience building, Shopify and Klaviyo as core platforms, and AI-driven A/B testing for emails and UX optimization. Advani emphasizes problem-first thinking over tool-first adoption, and stresses that successful downstream implementation requires factory partners to adopt the right mindset (not skill set) for velocity - enabled by passing purchase orders in their native formats with two-week visibility into demand.
They built the Ministry Production System using AI to automate manual PO creation and placed purchase orders at the SKU level (two per day) rather than style level (twice yearly), with factories receiving orders in their native templates and two-week demand visibility to enable preparation.
Single-point ownership of decisions (eliminating committees) and sequential rather than parallel execution; 20 hours of work that traditionally spread over six months now compresses into seven days by batching tasks together.
Landing pages via Lovable, audience building with Arido, A/B testing for emails through AI tools, and dynamic UX optimizations - always starting with a specific problem, not the tool itself.
Start with the problem, not the tool; AI is one solution in the toolbox, not the only one, and success requires integration with existing platforms (Shopify, Klaviyo) and clear, short-term profitability metrics, not long-lead sentiment indicators.
As customer acquisition becomes harder and more expensive, retention and reacquisition of existing customers with quality products has become easier; the shift to closet-share (customers buying multiple pieces) has driven stickiness metrics up significantly in the last six months.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode delivers some concrete operational insights - particularly around rapid testing cycles (10 days), the shift to SKU-level frequent ordering, and problem-first AI adoption - but is diluted by repetitive framing, softball questions, and social pleasantries. The guest restates the 'problem first, not tools first' philosophy multiple times without deepening it. Useful for operators but not packed with novel thinking.
We've changed the system so we are allowed to test more things more frequently and learn faster.
How can we place rapid pos. Right. So we used AI to build this and create a, uh, we call it Ministry production system.
The 'problem-first, not tool-first' approach to AI is sensible but well-trodden advice in 2024. The rapid experimentation and 10-day testing cycle is moderately fresh for apparel, but the core ideas - agility as superpower, single decision ownership, sequential not parallel work - are familiar frameworks. No truly contrarian or first-principles thinking emerges.
And the second major rule is no shortcuts. So all the stages of implementation, execution, decision, all still happen. Negotiations, all in there, they just happen sequentially in rapid fire.
The big mistake we made, and I think a lot of people made with AI was just one. Just fire hose. Right. Take it all in, let's do everything.
Aman Advani is the founder and CEO of Ministry of Supply, a real apparel brand operating at scale with concrete supply chain and demand challenges (Mexico, China, NY production, 2,500 SKUs weekly). He is a practitioner, not a theorist, with hands-on experience navigating tariffs, inventory, and operational complexity. Solid caliber for a B2B operator audience, though not a household name or exceptionally high-profile founder.
My name is Iman. I'm the founder and CEO of a clothing brand, Ministry of Supply
We were fulfilling product out of Mexico, we make product in China, and we make a lot of product in New York. So we were facing every ounce of tariff unrest.
The episode includes specific operational metrics: 2,500 SKUs weekly, shift from 2 POs per year to potentially 2 per day, 10-day testing cycles, on-demand manufacturing with Taylor Industries, and named tools (Shopify, Klaviyo, Lovable, Arido). However, concrete financial impact is absent - no revenue figures, margin improvements, or LTV gains are quantified. The concrete details support the narrative but lack depth on business outcomes.
We have these customers who are not buying one piece, but buying the whole bag. That shift to closet share has been incredibly helpful.
Right now we're placing a PO for every style every week. 2,500 SKUs every week. We used to place them twice a year.
The host asks reasonable opening questions but rarely pushes back or probes deeper. Questions are mostly affirming ('That's incredible,' 'Awesome') rather than challenging. When the guest mentions major shifts (tariffs, on-demand manufacturing, AI systems), the host does not ask for specifics on what broke, costs, timelines, or failures. The conversation ends with off-topic small talk (wine tasting, shorts), and the host does not follow up on competitive advantage or sustainability claims.
Wow. Okay. That's incredible.
Right. And I remember that sort of just taking back in production, onshoring again, just back in the days when things were made in New York and la, especially apparel items. Right?
Computed from the transcript - who did the talking, and the words that came up most.
In this insightful interview, Iman from eTale shares how his apparel brand adapted to supply chain disruptions, leveraged AI for rapid testing and inventory management, and built customer loyalty through experimentation. Discover strategies for agility, AI integration, and maintaining innovation in a challenging market. Supply chain agility and on-demand inventory AI-driven rapid testing and decision-making Building customer loyalty and increasing LTV
Transcribed and scored by The B2B Podcast Index.
Speaker A: Welcome to AI for business leaders podcast. Hello, everyone. We have Iman from ETEL at Palm Springs right here. Really, really excited to have him. You know, big fan of the brand. So if you, if you don't mind,
Speaker B: give me a little background about yourself. Yeah, thanks for having me. Uh, my name is Iman. I'm the founder and CEO of a clothing brand, Ministry of Supply, and we're here in beautiful Palm Springs enjoying e tail to showcase the brand and tell people a little bit more about our clothing. I think soft, stretchy, machine washable, where to work, where to travel, where to conference. That's what we focus on.
Speaker A: Awesome. Awesome. Um, as some people know, I did talk a little bit about my, my family background as well. We actually started manufacturing sportswear like Nike in the 80s. That's kind of like my family's back to business. It's a very tough business, as you know, and the SKU counts everything else. But today it is about you. So I guess even just to start out, as you mentioned, you don't mind the hard questions. So in the past 12 months there's been a lot of changes, right? Tariffs, the consumer sentiment change. AI, how has it been going for you? And you know, what are you doing about it?
Speaker B: Yeah, man, it's been a wild 12 months for us. I mean, we have the perfect storm. We were fulfilling product out of Mexico, we make product in China, and we make a lot of product in New York. So we were facing every ounce of tariff unrest. We made a massive shift in our supply chain away from holding a bunch of inventory and moved to a very kind of on demand, just in time type of inventory. And that's been hugely beneficial. Shout out to a partner, Taylor uh Industries in New York, who's been an incredible factory backend for us to produce units of one, but short, uh, version is huge time of change. And major lesson learned is, as I've been saying, agility is the only remaining superpower. So just how quickly is our rate of change?
Speaker A: Right. And I remember that sort of just taking back in production, onshoring again, just back in the days when things were made in New York and la, especially apparel items. Right?
Speaker B: Yeah, we still do about half and half in New York and China and we have a Chinese factory that we've been using for years that the quality is incredible. So how can we allow both to coexist and still, you know, get through the tariffs and um, grow to see another innovation, you know?
Speaker A: Awesome. From a positive standpoint, what have you seen in the past 12 months in your business that you're just very bullish about.
Speaker B: Yeah. The loyalty of our customer base has always been wonderful, but it's kind of gone off the charts in the last, really six months, particularly as it's harder and harder to acquire new customers. The good news is that it's easier and easier if your product is good, to reacquire and retain your existing customers. So we've just seen that all of the metrics on stickiness continue to go up where we have these customers who are not buying one piece, but buying the whole bag. That shift to closet share has been incredibly helpful.
Speaker A: That's really awesome. And I think that's something that a lot of brands will wish that they. They, they can do or they have in terms of their loyalty of the customers. There are things that you learn that's worked out well as you guys sort of test it out. How do we, you know, increase the LTV at the end of the day, I guess.
Speaker B: Yeah, yeah. You know, what we've really done more than any specific win was created a culture of experimentation. Uh, it's actually what I'm here at ETEL talking about is how quickly can you test something? All right, I want to test out putting video or gifs on the homepage, or I want to test out a new product idea or a new widget or a new tool. How can we go from that moment to it worked or not within 10 days, and where that might have taken us three, six months in the old world.
Speaker A: Right.
Speaker B: We've changed the system so we are allowed to test more things more frequently and learn faster.
Speaker A: 10 days.
Speaker B: That's pretty incredible concept to end of test.
Speaker A: Wow. So I assume that also takes some kind of organizational change and also training for the team as well. How did you navigate that?
Speaker B: Yeah. So number one thing is I'd say single leadership. So making sure that one person owns the decision at every stage of any.
Speaker A: This is true street style.
Speaker B: I like it. So single ownership. So there's no committees. It's one person owns the final decision. So it's very clear. I want to take your input, but I'm going to make the final decision. And the second major rule is no shortcuts. So all the stages of implementation, execution, decision, all still happen. Negotiations, all in there, they just happen sequentially in rapid fire. So we're taking what used to happen in 20 hours of work over six months and making it 20 hours of work over seven days. And by clumping it together and making it sequential, we're able to rattle off more tests.
Speaker A: Wow. Okay. That's incredible. TK will pivot to. Everybody talks about AI, Obviously, AI is everywhere. AI is obviously developing really, really rapidly. I think what's really exciting for people who are in retail and commerce is that on one end, how does that affect your day to day sort of workflow for internal teams? Two, what are some of the things you're seeing that you can bring to the table for the consumers, our loyal customers?
Speaker B: Yeah, yeah. So I'll answer the first one first, which is that I think the big mistake we made, and I think a lot of people made with AI was just one. Just fire hose. Right. Take it all in, let's do everything. Second one is kind of being a little bit more discerning and not just saying yes to every AI tool. The second thing is actually starting with the problem, not the tool. And so for instance, how do we place. Right now we're placing a PO for every style every week. 2,500 SKUs every week. We used to place them twice a year. How can you do that? The only answer is AI. Right. And so now what we do is we built a system that allows us to place rapid pos. Right. So we used AI to build this and create a, uh, we call it Ministry production system. So one example of saying it's not AI, it's problem, and the only solution happens to be AI. So how do we scale? Well, let's use AI.
Speaker A: So, right.
Speaker B: Going back to problem, first thinking and then using AI as a choice.
Speaker A: Right.
Speaker B: In your toolbox, but not the only tool.
Speaker A: So when you're saying the velocity of these orders increase in utilizing AI, do you mean was the problem more so just ordering the right style and the right SKU on top of just the amount of the numbers and then just the process, administrative task of making orders.
Speaker B: You're spot on. So in our world, data is coming inbound continuously. We're getting orders every minute, right?
Speaker A: Sure.
Speaker B: Uh, on the outbound, we're placing two POS per style per year.
Speaker A: Right.
Speaker B: Wild. With the help of AI, we've built this system mps, which is allowing us to place. We could place two orders a day at a SKU level, not two orders a year at a style level. And that was a human constraint. We were unable to possibly produce that many POs. They're very difficult, very manual. There's ah, translation to the factory and their formats, then to the freight forwarder, then to flex port. There's so much required at any single PO that to do so in such volume is impossible without the help of AI, it isn't about ordering more or less. It's about frequency and SKU level detail. Not you know, a uh, discrete chunky POS going out twice a year.
Speaker A: Right, right. Because. So you can be much more sort
Speaker B: of focus specific and stocking rates go up, inventory on hand goes down, discount rates, obsolescence goes down.
Speaker A: Right.
Speaker B: And our workload is lighter. Right. Which is the beauty of AI. Right. Is if you're doing it right, it can be a win win.
Speaker A: It should be helping the humans, it's not replacing humans. That's right. That's what I think. How does that affect your vendors downstream though? Because I assume the velocity now ramps up on your end. They're going to have to uphold whether it's supply chain is complicated, but all the way to production as well. Yeah, you're going to need to basically
Speaker B: I love that we're just going on a supply chain. I'm into it. I love it. I love. So I know I'd say first of all they have to be in the right mindset. So it's not a skill set, it's a mindset.
Speaker A: Yeah.
Speaker B: They're capable of doing. Any factory can do this. It's a mindset. Right. Meaning it's a hassle. Right, Right. The other thing is, I'll say we're also making it easier for them. Right. So we're passing them POS in a style that they're much more comfortable ingesting. So there's no translation. Usually it was take RPO and put it in their format. Now we're giving it to them in their own language, quite literally in their own templates that they built for us to say this is how you can help us. So again by, by being able to connect our systems more intrinsically. They can see demand two weeks before we even place the po.
Speaker A: Right.
Speaker B: They know it's coming, they can prepare, they can batch as they need. But I will say the requisites to any system working is the players having the right mindset.
Speaker A: Right. Okay, that sounds good. Shifting a little bit to the, to the consumer side. Have you seen, have you guys tested anything that's AI based and what have you guys seen from there?
Speaker B: Yeah, so I'd say absolutely. I mean if you, if you go through our experience, a ton of what is happening there. UX optimizations, AI driven, uh, our landing pages are all now built by Lovable. Which is, you know right now, um, you know our site A B testing is all done through AI on the bar. Emails are tested using AI tools. So our Audiences are all built using Arido, which is an AI audience building tool.
Speaker A: Sure.
Speaker B: So you got to think every little piece of this is being augmented not by something we built, but by something we had to have the discernment to say we have a problem. In this case, let's say our email audiences are too manual.
Speaker A: Right.
Speaker B: And recency oriented. Let's find a solution. Are offering us a much more dynamic offer. So let's use their AI and test it out.
Speaker A: Right.
Speaker B: So in each of these cases, again, problem first. AI is a solution, but uh, not the only one.
Speaker A: I get that too. Just with the amount of talks that people have around AI, I think it's really easy to kind of get into the mindset that the candy store mindset, like everything is being shipped every single day just from all the AI tools and native AI tools. Other aspect where you feel like it's just, it's um, as sort of established
Speaker B: as it should be in, in, in the AI ecosystem.
Speaker A: Right.
Speaker B: You know, what's happening, at least in our world is we're finding that your backbone kind of defines your tool sets. Our backbone is Shopify Klaviyo. Not surprising, not uncommon. And that, you know, now it's you've got people who are only plugging into Klaviyo or you've got people who are only plugging into Shopify. And so it's a requisite to say now we've got the hub, one of the spokes and I think that's the first forcing factor to say it has to be easily implement or else there's no chance we can test it in a week or two. Right. So this has to be a plug and play. Right. Potential.
Speaker A: I have heard a lot of different sort of native AI, uh, companies also explaining how they're shifting the go to market strategy. They're saying that on the buyer side the expectations are much higher. You're going to need easily implement. Uh, the solution two is that nobody wants long sort of dev time. Three is the contract has to be flexible. You may have to do some proof of concept as well as bake offs, all these things.
Speaker B: Four is the success metric has to be clear and profitable.
Speaker A: Yeah. Right.
Speaker B: We're not looking for slow, long lead metrics of sentiment or you know, hey, did we make money with this improvement? Did it make our lives better in some way and allow us to keep on building something we love? Right. The premise is not that we, you know, are profit mongering, it's, it's how uh, do we make a better healthier business, grow it and get in more people's hands.
Speaker A: Awesome. What are you looking forward to for, uh, the most in the next 12 months?
Speaker B: Shorts.
Speaker A: Okay.
Speaker B: I know the question is probably premised on like software and AI, but we've got.
Speaker A: It's okay.
Speaker B: No, this is coming. We've got new shorts coming, we've got new T shirts coming. So I like, I'm a product junkie.
Speaker A: Yeah. Yeah.
Speaker B: Ah. And so we're like, spring summer collection is just incredible. And, uh, I'm like, I saw the fit model today on video wearing the final sample of the shorts. No, we did it. Yeah.
Speaker A: Nice, Nice. Everybody gotta go check it out for sure.
Speaker B: And then, uh, I'm also excited for the tariffs to hopefully lift after 150 days, which will make everyone's life easier. Right?
Speaker A: Yeah, I agree. I agree. One last question.
Speaker B: Yeah.
Speaker A: What's been the one most fun activity in Palm Springs so far for you, man?
Speaker B: I haven't taken enough advantage of it, but I'm going to give a forward looking one. We're doing a wine tasting tonight. Weather holds up. It's, I believe, outdoor. Uh, yeah. I don't think it rains much here. Right?
Speaker A: No.
Speaker B: And, uh, I'm so excited at the end of this wonderful conference. Yeah. A nice glass of wine and decompressed a little bit.
Speaker A: This sounds fantastic. Thanks for having me. Thank to everyone listening if you found this valuable. Subscribe to the podcast AI for business leaders on, um, Spotify, Apple podcast, YouTube and leave a comment. Please also share with others navigating the future of AI. Thanks for listening and we'll catch you on the next one.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.
321 | How to Build an AI App Without Coding: From Idea to App Store with Replit & Claude by Bryce Rattner KeithleyLeveraging AI · on Lovable80 / 100
MicroConf Tactics: Start a SaaS From $0 in 2026MicroConf On Air · on Lovable77 / 100
Summer Break Replay: How Dylan Jahraus Turned Etsy into a Multi-Six-Figure BusinessMOMetize Your Passion · on Shopify76 / 100
The Rising Cost of GrowthThe Fractional CFO Show with Adam Cooper · on Shopify75 / 100
Meta Ads: Leverage UCP and First-Party Data for Meta Ads Success with Yiqi Wu, AimerceKeep Optimising · on Shopify72 / 100
Define the Message and the Roadmap Follows with Jamie Domenici, CMO of KlaviyoThe B2B CMO Podcast with Jon Miller and Sydney Sloan · on Klaviyo70 / 100
Why ElevenLabs Built a Team Around Adoption74 / 100
Albert Chun, Founder/CEO of AI Circle, on Training a Frontier Model, and Why "Everyone's B+" Without Experience61 / 100
Gary Benerofe, General Partner at Mu Ventures, on the Future of Agentic Commerce and What VCs Really Look for in AI Founders61 / 100
The AI Maximalist Playbook: James Raybould on Building at the Speed of Ideas65 / 100
Charlie Ninegar, COO of DSS Games, on Getting People Off Their Phones and Growing Against the Grain75 / 100