Returns are a nightmare.
Two Boxes just raised $3.2 million to fix that problem with AI, and honestly, it’s about time someone took this seriously. The Denver startup landed funding from Assembly Ventures to scale its AI-powered returns processing platform, targeting retailers and third-party logistics providers who are drowning in returned merchandise.
Why This Actually Matters
Look, I’ve tested dozens of AI tools that claim to solve business problems. Most are solutions looking for problems. But returns? That’s a real pain point with real money attached. Every retailer knows the drill: products come back, someone has to inspect them, decide if they’re resellable, figure out restocking, and somehow not lose their mind in the process.
Two Boxes is building AI to automate that mess. The funding will help them advance their product roadmap and engage more aggressively with 3PLs and enterprises. Translation: they’re going after the big players who process thousands of returns daily.
The Honest Assessment
Here’s what I want to know that the press releases don’t tell us: What’s the actual accuracy rate? How many false positives are we talking about when the AI decides whether a returned item is damaged or resellable? Because if this thing flags perfectly good products as unsellable, or worse, lets damaged goods slip through, retailers will dump it faster than a customer returns a wrong-sized sweater.
The $3.2 million raise is solid but not spectacular. It’s enough to build and iterate, but not enough to suggest investors think this is the next unicorn. That’s actually a good sign. It means Two Boxes is probably focused on solving a specific problem well rather than trying to be everything to everyone.
What They’re Up Against
Returns processing isn’t sexy, but it’s expensive. The logistics space is notoriously resistant to new technology because margins are thin and mistakes are costly. Two Boxes needs to prove their AI can handle the chaos of real-world returns: items without tags, products in wrong boxes, customers who claim something is defective when they just changed their mind.
The fact that Assembly Ventures led this round matters. They’re not throwing money at every AI startup with a pitch deck. They typically back B2B software that solves operational problems, which suggests they see real potential in automating returns workflows.
The Reality Check
AI in logistics and fulfillment is tricky. I’ve seen plenty of computer vision systems that work great in controlled environments and fall apart when faced with warehouse lighting, damaged packaging, or the infinite variety of products that get returned. Two Boxes needs to prove their system can handle that variability.
The plan to engage more aggressively with 3PLs is smart. These companies process returns for multiple retailers, so if Two Boxes can prove ROI there, they’ve got a repeatable sales model. But 3PLs are also notoriously price-sensitive and risk-averse. The AI better work, and it better save them real money.
What I’m Watching For
I want to see case studies with actual numbers. Not “improved efficiency” but “reduced processing time by X minutes per item” or “decreased mis-categorization by Y percent.” I want to know how long implementation takes and what kind of training data they need to get the system working in a new environment.
The returns problem is real, the market is huge, and AI is probably the right tool for the job. But execution is everything. Two Boxes has the funding to prove their technology works at scale. Now they need to deliver results that justify the investment and convince skeptical logistics operators to trust their AI with real inventory decisions.
I’ll be watching to see if they can turn this funding into a product that actually makes returns less painful. Because if they can, they’ve got a real business. If they can’t, well, $3.2 million buys you enough runway to figure that out before the money runs dry.
🕒 Published: