The return request arrives. Someone on the team reviews it, generates a shipping label, updates the inventory system, processes the refund. If you're lucky, and if your operations staff catches it early enough, you salvage an exchange. Repeat a few hundred times per day, and you begin to understand why the post-purchase experience has become a quiet profit drain for direct-to-consumer brands.
Pango, a five-person team based in Stockholm, is betting that most of this work shouldn't require human hands at all. The startup—part of Y Combinator's Summer 2026 batch—launched what it's calling an "Agentic OS" for e-commerce on July 24, deploying AI agents to autonomously manage returns, delivery exceptions, tracking queries, and customer service. No humans required, at least in theory.
Whether that works at scale remains an open question. But the early results suggest something more than incremental improvement.
The Pitch: One System, All the Chaos
Pango's approach is disarmingly simple. Connect your e-commerce stack—Shopify, WooCommerce, whatever you're running—describe in plain language how you want returns and shipments handled, and the platform generates the workflows, customer portals, and AI agents to execute them. The system integrates with over 100 carriers, normalizes tracking data across all of them, and automates the kind of decision-making that typically sits with an operations manager pulling double shifts.
The product tries to cover the entire post-purchase lifecycle, which is more ambitious than it sounds. It manages delivery options at checkout using real-time carrier rates. Tracks orders with predictive ETAs. Handles return requests through a branded portal that looks like it belongs to your company, not some generic returns processor. Runs customer service conversations when someone asks "where is my order" for the third time.
According to the company's July launch materials, it automates 99% of returns and shipping operations and cuts operational costs by 20% for customers. Bold claims, and the kind that usually don't survive contact with reality—except Pango has at least one data point that backs them up.
The Switch Nails Case: What Actually Happened
Switch Nails, a beauty brand, provides the clearest view into how this works in production. Pango published a case study showing the company has tracked north of 150,000 deliveries through the platform. More revealing: Switch Nails converted 19% of return requests into exchanges and achieved 99% self-service resolution. Translation: nearly every post-purchase interaction happened without a human in the loop.
That 19% exchange rate deserves a second look. Most brands see returns as pure loss—you ship the product, the customer sends it back, you refund the money, and maybe you can resell the item if you're lucky. Converting one in five returns into an exchange means you're keeping revenue that would otherwise walk out the door. For a DTC brand operating on thin margins, that's not a rounding error.
The company claims to have signed 30 global brands since launching, though most haven't been named publicly. Shopify App Store reviews from March and April reference customers including SunDaze Surf and an unnamed racing shop. Understatement Underwear shows up in testimonials on Pango's site. The company maintains a 5.0-star rating, though the sample size is small enough that one bad month could change that.
Astrophysics, Molecular Biology, and the Problem of Returns
CEO Steve Rahimi started in e-commerce at 17—he's now spent seven years in the space, which means he's still young enough that this kind of operational tedium probably feels solvable rather than inevitable. His co-founder, Lukasz Reszczynski, took a different route: astrophysics, then molecular biology PhD studies he dropped out of, then lead developer at Partfiniti where he built a product configurator that powered $1.5 billion in quotes.
Before starting Pango in 2025, Reszczynski built delivery infrastructure for a UK restaurant chain, which helps explain why they think in terms of systems rather than point solutions. The company raised 5 million Swedish kronor from HEARTFELT VC, SSE Business Lab, Wave Ventures, and Karaoke Club, plus individual investors including Johan Licke von Sydow, Marc Verschueren, and Anders Signell. They operate primarily out of Stockholm, though older job postings list a San Francisco presence.
The funding round is modest by Silicon Valley standards, but for a team of five building enterprise software, it's enough to figure out whether this actually works.
The Crowded Field They're Walking Into
Returns management isn't exactly virgin territory. Narvar, Loop Returns, AfterShip, parcelLab—the incumbents are well-funded and deeply entrenched. Most offer returns management or order tracking as distinct products. Pango's argument is that post-purchase operations shouldn't live in five different tools that require custom integration work to talk to each other.
The company positions directly against Loop and ReturnGo in its marketing, arguing those platforms handle returns but leave checkout delivery options, live carrier rates, and tracking as someone else's problem. Whether merchants prefer an integrated system over best-of-breed tools depends on how much they value specialization versus simplicity. And whether Pango's unified approach delivers enough in each category to compete with dedicated solutions.
The timing, at least, makes sense. Salesforce launched Agentforce Commerce in July 2026, pushing AI agents that check inventory, resolve service issues, and supposedly close sales autonomously. The vocabulary Pango uses—agentic OS, autonomous operations—reflects a broader industry shift from chatbots that answer questions to systems that actually do things. Whether that shift is real or just rebranded automation is a question the next 18 months will answer.
What Happens When the Agents Break
For DTC brands buried in post-purchase support tickets, Pango offers a plausible way out. The Switch Nails numbers look good on paper—19% return-to-exchange conversion, 99% automation—and if they hold across a broader customer base, the platform could materially improve unit economics. Smaller brands with simpler needs might be fine with Shopify's native tools and a virtual assistant. But for anyone doing meaningful volume, the promise of autonomous operations is hard to ignore.
The real test comes when Pango scales beyond 30 customers and starts encountering the edge cases AI agents can't resolve autonomously. Post-purchase operations are inherently messy, full of judgment calls that don't fit neat rules. A customer says the product arrived damaged but the photo is blurry. Someone requests a return after the 30-day window but swears they've been traveling. A carrier marks a package delivered, the customer says it never showed up, and now you're in he-said-she-said territory.
How well Pango's agents handle ambiguity—and how gracefully they escalate genuine exceptions—will determine whether this is a marginal improvement or a genuine infrastructure shift. The company's betting that AI has crossed a threshold where it can reliably handle not just structured tasks but the kind of fuzzy judgment calls that have always required humans.
Maybe it has. Or maybe we're about to learn, again, where the boundaries are.
