Pango Wants to Fire the Humans Haunting Your Returns Portal

The YC S26 startup is building an agentic operating system for the purgatory after checkout. Its numbers are loud. The hard part begins when the package goes missing.

August 05, 2026

In partnership with

There is a particular hour in e-commerce when the glamour dies.

The ad has converted. The influencer has smiled. The customer has clicked the bright little button, and somewhere a Shopify chime has rung like a casino machine paying out. Then the parcel enters the swamp: carriers, warehouse systems, tracking pages, return labels, refund rules, exchange incentives, customs forms, support tickets, and the customer’s increasingly feral question—where is my stuff?

In that fluorescent back room, humans become middleware. They copy tracking numbers between tabs. They approve the same return for the six-hundredth time. They explain to an angry customer that the carrier says “delivered,” while the porch says otherwise. Pango, a five-person Y Combinator Summer 2026 company from Stockholm, would like to send an autonomous digital employee into this administrative bog with a machete.

You Already Have a Take on What AI Does Next

OpenAI or Anthropic? Which model leads the next benchmark? Which company ships the next major breakthrough?

If you follow AI closely, you already have opinions on where the industry is headed. Kalshi lets you trade on real-world AI and technology events, with markets that move as models launch, benchmarks drop, and announcements happen.

The people who follow this space most closely often see the story developing before everyone else. Put that knowledge to work and trade what you think happens next.

Bonus credit varies from $15 to $500. Terms apply.

The pitch is simple enough to fit on a shipping label: connect Pango to the merchant’s stack, describe the rules for deliveries, tracking, returns, exchanges, refunds, and claims, and let agents execute the routine work while humans judge exceptions. The company’s website now stretches the ambition across a broad logistics command center—transportation management, checkout delivery promises, order tracking, returns, pick-and-pack, data analysis, and more than 100 integrations.

This is not merely another chatbot taped to a dashboard. Or at least, that is the wager.

The seven-dashboard séance

Pango says its team spent 12 months embedded with 30 e-commerce brands and found each running a home-brewed contraption of five to seven tools. The merchant’s staff supplied the missing integration layer with nerves and fingertips. Pango’s YC launch post claims the system has since automated 99% of returns and shipping operations, cut operating costs by 20%, and signed 30 global brands.

Those are company claims, not audited figures, and the denominator matters enormously. “Automated 99%” might mean thousands of clean, repetitive approvals with the truly radioactive one percent still consuming most of the skilled labor. “Signed” can cover territory from a pilot to a meaningful contract. Pango does not publish revenue, retention, transaction volume, or customer-level case studies sufficient to test the claims publicly.

Still, there are signs of a machine with metal in it. Pango’s site names Understatement, Eytys, Kukinos, GLAS, and Switch Nails among customers and quotes Understatement saying the platform puts deliveries, tracking, and returns “on autopilot.” Nordic Innovation House reported that Pango raised 5 million Swedish kronor and described the company as a TINC Fall 2025 alumnus. HEARTFELT_ announced an investment alongside SSE Business Lab, Wave Ventures, Karaoke Club, and angel investors. That does not prove the 99%. It does prove that the circus has sold some tickets.

The founders at the loading dock

Steve Rahimi is the commercial animal in this two-founder story. YC says he has been an e-commerce entrepreneur since 17, with seven years spent building and scaling in e-commerce and technology. Nordic Innovation House adds a cinematic detail: Rahimi arrived in Sweden a decade ago with $13 in his pocket. Today his LinkedIn bio describes the mission as building a digital e-commerce employee.

Lukasz Reszczynski brings the delightfully deranged résumé startups manufacture when ordinary careers prove insufficient. YC says he was lead developer at Partfiniti, where he created a patented product configurator that powered more than $1.5 billion in quotes and the sales process of Accu-Tech, a company YC characterizes as having $300 million in revenue. He has acted as CTO for several companies, led delivery infrastructure work for a UK restaurant chain, earned a master’s in astrophysics from Liverpool John Moores University, pivoted into a molecular-biology PhD at the University of Vienna, and dropped out to build Pango.

YC lists Gustaf Alstromer as Pango’s primary partner. The YC bio links LinkedIn profiles for both founders. It lists no Twitter/X link for either founder.

A crowded returns counter

The enemy is not emptiness. It is incumbency.

AfterShip spans tracking, returns, shipping, and delivery promises. Loop has built a formidable returns-and-exchanges product. Ingrid attacks delivery experience and checkout logistics. Narvar, parcelLab, Gorgias, ShipStation, nShift, and an army of Shopify apps all occupy pieces of the same post-purchase acreage. Pango names AfterShip, Loop, and Ingrid directly, arguing that those products remain isolated tools while its AI-first system shares data and coordinates workflows across the whole operation.

That is the right strategic line and a dangerous one. A startup can win by unifying a fragmented stack; it can also become the newest dashboard in the stack, waving an “agentic OS” banner while the warehouse manager quietly keeps the old tabs open. The sales cycle grows teeth as Pango moves upstream. Logistics touches money, customer trust, tax, fraud, carrier contracts, warehouse throughput, and regulated personal data. Every integration is a fresh place for reality to kick the software in the ribs.

The exception is the product

Returns are full of adversarial ambiguity. Is the dress unworn? Did the parcel vanish, or did a neighbor take it? Should the customer receive an instant refund? Is the buyer a loyal whale, a first-time shopper, or a serial abuser? A machine acting across refunds, labels, inventory, and customer messages must be fast without becoming a wonderfully efficient fraud subsidy.

Pango says its agents escalate edge cases for human review. That boundary—what the agent may do, what evidence it sees, how it explains a decision, how quickly a person can reverse it—is more important than the sparkle dust around the word agent. The company lists GDPR and the EU AI Act on its site, while ISO 27001 is explicitly marked “in progress.” Enterprise buyers will want the unromantic goods: permissions, audit logs, uptime, rollback, data residency, incident response, and proof that an autonomous workflow cannot turn a bad configuration into ten thousand bad refunds before breakfast.

The other risk is margin. Pango must ingest messy events from carriers and commerce systems, run models, maintain integrations, support merchant-specific rules, and remain accountable when downstream systems lie. That can become a services business wearing software’s leather jacket. The real test is whether each new brand makes the platform smarter and easier to deploy—or merely adds another handcrafted maze.

Why this matters

The loud story of AI commerce happens before checkout: synthetic product photos, personalized ads, shopping agents, conversational discovery. But the durable advantage may be built after the sale, in the operations customers notice only when something breaks.

If Pango works, a mid-market brand could wield logistics capabilities once reserved for retailers with armies of operators and custom software. The returns portal stops being a bureaucratic trapdoor and becomes a place to preserve revenue with exchanges, smarter routing, and policy tuned to the customer and the item. The company gets one operational memory instead of seven mutually suspicious dashboards.

And if it does not work, the post-purchase swamp will accept another dashboard without complaint. Swamps are generous that way.

Pango has chosen a valuable, ugly problem and made claims bold enough to be falsifiable. That is more interesting than another agent that summarizes meetings. Now it has to survive the lost parcels, lying APIs, fraudulent returns, carrier outages, warehouse improvisations, and customers pounding the “where is my order?” button at 2:13 a.m.

That is where the digital employee either earns its badge—or gets locked in the returns cage with the rest of the software.

Sources