Minimal AI (YC S25) ads

A few weeks ago, Niek and I got on a plane to Stockholm to meet our first Swedish customer in person 🇸🇪 It started with a few targeted cold calls to test the Nordic e-commerce market. The team at NUDIENT was one of the first I spoke to, and after the famous 6-week stretch where Sweden basically shuts down for summer, we started working together. I enjoyed it so much that I suggested to Niek we fly over to meet them. The plan was a lunch with just Pontus and Vidar. To our surprise, they invited some friends who run great e-commerce brands in Sweden, and we ended up at a large table just outside Gamla stan, having a great time. Since then, we've felt a real pull from Sweden. We've met a lot of amazing people in e-commerce, and a few more brands from that table have joined us. Everyone said the same thing about Swedish e-commerce: meet people in person (same advice Y Combinator gives). So that's what we're doing. First stop: XP Awards in Stockholm this Thursday (where coincidentally Nudient is nominated for Innovator of the Year 😅). Pontus and Vidar, thanks for being great friends and partners! And if you want to see some unique phone cases, backpacks or suitcases, go check out NUDIENT! See you again soon, Stockholm 🇸🇪

XXL Nutrition automates 83% of customer support tickets, saving 70 hours a week and growing their Trustpilot from 4.2 to 4.5 in 2 months. Not a chatbot. Full end-to-end resolution: reshipping orders, investigating parcels, resetting accounts, all handled automatically across WhatsApp and email. This is what real AI support automation looks like. → https://lnkd.in/emsZwj2M

Minimal AI (YC S25) got featured by Forbes as one of the most promising startups of the Y Combinator S25 batch. With just a 0.6% acceptance rate, this was the most competitive batch ever (!) It's great to be on this list alongside some of the smartest people I have ever met, including fellow Europeans Pietro Zullo and Luigi Pederzani from mcp-use (YC S25), Aditya Jha and Neil Nie from Verne Robotics and our friends Alec Howard, Carlos J. and Michael Mason from Cacao. Thanks Dasha Shunina for sharing our story! Link to full article in comments.

Excited to share our $3.6M seed round for Minimal AI (YC S25), backed by Y Combinator, Rebel Fund (from Reddit's CEO Steve Huffman), Zeno Partners, Formosa Capital, Sunshine Lake and amazing angels including 15 (!) of our customers. We built Minimal AI from one belief: setting up and managing AI customer support agents should be easy. Sounds obvious. It's not. Every eCommerce brand has unique systems, processes and policies that change constantly, and most teams are too busy running the business to become AI experts. That's why we introduced the AI Manager. Just tell it what you need in natural language and it instantly updates your agents, configures integrations and asks clarifying questions when needed. It even analyses thousands of conversations to improve on its own. Some highlights of the last 8 months: • Crossed 7-figures in ARR with a team of just three • Customers achieving 90%+ ticket resolution rates • Relocated to San Francisco for YC S25 thanks to Brad Flora • Raised our seed round in a week, before Demo Day even started None of this would have been possible without our customers who believed in us early. And of course, my co-founder Titus Ex and the rest of the team who worked relentlessly to make this a reality. Now back to work!

XXL Nutrition now automates 83% of customer support tickets. That saves them >70 hours every week, lifted their Trustpilot score from 4.2 to 4.5 in just 2 months, and significantly increased CSAT. And we're not talking chatbot responses, we're talking full end-to-end resolution of thousands of complex issues across WhatsApp and Email, including: • Reshipping missing products or damaged packages • Lost/damaged parcel investigations with Sendcloud • Account resets for forgotten passwords • Registering product issues and complaints ..and much more. Most people don't realise that true automation beyond ~50% requires AI Agents that actually take action, like the examples above. But getting those actions to work reliably is hard with most tools. The agent needs to know exactly when to act, what to do, and no two systems are the same. That's where our AI Manager made the difference. Instead of months of complex implementation, Bram Van de Loo simply talked to our AI Manager, who had each action up and running in minutes. A huge thank you to Bram Van de Loo, Stan Thomassen, Guus van Nunen & Robin Akker for this great collaboration! Link to case study in the comments!

This week we're rolling out self-learning for all our AI Agents. We've found that the hardest part of building and maintaining AI Support Agents is not the LLMs (these are good enough), but guiding the people in charge of the AI to set it up properly. Most of our customers are non-technical and have worked years in a support role. With our AI Manager, we've already made it much easier to set up and evolve the AI's integrations, knowledge, processes, and more. Now we're taking it a step further. The AI Manager analyses every single ticket, clusters the gaps, and transforms them into tasks: concrete steps to improve your AI Agents, optimising for automation rate and CSAT. These tasks are so good, they're the single best thing you can do to improve your AI Agents. No ambiguity. Super excited to lead the way in self-improving AI Agents. This is just the start. Eventually, they'll be fully self-optimising.