Amrita Bhasin, Co-founder and CEO of Sotira, on how AI is leapfrogging an old-school supply chain and turning unsold inventory from a cost into revenue.
Reverse logistics, the world of overstock, unsold inventory, and returns, is roughly an $800 billion industry. And according to Amrita Bhasin, a huge chunk of it still runs on pen and paper. Freight gets staged by hand. Compliance documents get shuffled by someone on a plastic folding chair. A lot of her customers have never heard of ChatGPT. Some don’t have websites.
I’ll be honest, this was an industry I knew basically nothing about before this episode. I walked away thinking it might be one of the best places anywhere to deploy AI right now.
Amrita is the co-founder and CEO of Sotira, which helps retailers and brands discreetly offload and monetize unsold inventory. She started out in mom-and-pop liquidation, learned how the business actually works, realized there was no streamlined software for it, and went and built it. Her work has been featured in Forbes, Fortune, Bloomberg, and CNBC.
How are AI agents changing supply chain and logistics?
Amrita calls supply chain a “leapfrog industry.” It’s skipping the incremental steps other industries went through and jumping straight from paper to AI.
“We’re literally going from a pen and paper process of I’m staging freight, figuring out what pallets, what cases go on what trucks. That’s all done on pen and paper. It’s largely done by hand.”
— Amrita Bhasin, Co-founder and CEO, Sotira
At Sotira, AI agents fill in compliance paperwork automatically, generate freight documents before the truck arrives, handle billing and scheduling, and stage loads. Give the system a full SKU list and it tells you it needs six trucks and how to pack each one most efficiently. That used to be hours of manual math in Excel.
“You can effectively have one person with AI doing the work of three to four people previously.”
— Amrita Bhasin, Co-founder and CEO, Sotira
Sotira is seeing 10+ hours saved per week from faster documents and automated billing and scheduling. That’s a big deal in an industry where time is money and margins are thin. It reminded me of what Amazon has done internally for years, except now it’s available to a regional distributor that doesn’t have Amazon’s engineering budget. That’s what democratizing AI actually looks like.
Here’s the part I found most interesting. Amrita argues that bringing data and automation to industries like this has a bigger ROI than building yet another tool for people who are already tech-savvy. The gap is just so much wider.
What is the $800 billion reverse logistics opportunity?
The reverse logistics market includes pure overstock (brand-new items, tags attached, still in original packaging) plus returns. Everything the primary market doesn’t consume flows into a secondary market of discount grocers, off-price stores, surplus emporiums, and dead stock stores.
And the volume keeps growing. Amrita pointed to “Summerween,” Halloween candy and costumes showing up on shelves in June and July, as a sign of how much pressure retail is under to push seasons earlier and keep inventory turning. Then add demand that’s nearly impossible to forecast. A brand goes viral on TikTok Shop for one day, stocks up to meet it, and is left holding the bag when sales return to normal. Super Bowl Sunday alone drives a quarter’s worth of planning in food and beverage, and demand falls off a cliff the week after.
“Instead of paying to destroy your product, you are capturing a recovery rate. We see it as being able to bring our customers from some level of negative, because they’re paying to destroy, up to a level of positive.”
— Amrita Bhasin, Co-founder and CEO, Sotira
That’s the reframe. Most brands today discard, dump, or incinerate overstock. Amrita’s pitch is that it’s working capital sitting in a warehouse, and the accounting team can see the impact directly on the bottom line.
She also made a comparison I hadn’t thought about: Europe has reverse logistics much more figured out than the US, partly because of sustainability legislation and partly because of a cultural expectation not to waste. In the US, incinerating product is still the status quo.
Can data and AI solve demand forecasting in retail?
Everyone in retail wants the same thing, and Amrita hears it on every conference stage she’s on.
“The number one question is how can we get better data and more consistent, accurate data?”
— Amrita Bhasin, Co-founder and CEO, Sotira
But she’s clear-eyed that better data runs straight into the privacy debate. She brought up the famous Target story, where the retailer’s data predicted a pregnancy before the customer had told anyone, and the lawsuits around period tracking apps.
This is where I jumped in, because I have strong feelings about it. There’s a line between creepy and useful, and brands need to be good stewards. But in my experience, far more people want relevant advertising than don’t. You’re going to see ads either way. Why would you want irrelevant ones?
Amrita actually tested this herself. She turned off location for Instagram and Facebook ads while moving around over the summer. Meta’s last known location for her was DC, so while she was living in New York she got nonstop ads for DC restaurants. She turned location back on. As she put it, if you’re going to be in the system anyway, you might as well get something useful out of it.
How can brands monetize unsold inventory without hurting their brand?
This is the catch-22 that explains why so much product gets destroyed. Dump overstock on Amazon at a discount and you’ve just taught every customer never to pay full price again.
“If you walk into a store, you’re buying based on what the discount is at that point. It’s that treasure hunt bargain vibe versus if you type the product in and it shows up on Amazon for a lower price, that’s really cannibalizing.”
— Amrita Bhasin, Co-founder and CEO, Sotira
So Sotira routes goods mostly to brick-and-mortar secondary market sellers, not online marketplaces. She described overstock Hokas selling for $20 at a store in Maine, with bargain hunters driving up from New York to grab them. That treasure hunt culture is pulling a lot of younger people into a traditionally old-school industry.
Amrita also connected overstock back to marketing data. Newer brands like Vuori and Alo are outcompeting Lululemon partly because they’re so good at using TikTok, Instagram, and influencer data to reach the right buyers directly. Better targeting means tighter production, which means less overstock in the first place.
“The brands and retailers that are able to use data are frankly the ones that are going to be able to win because they’re going to be able to reach customers faster.”
— Amrita Bhasin, Co-founder and CEO, Sotira
How are tariffs affecting the supply chain?
A listener asked, and Amrita’s answer was that everything is running day by day. Aluminum tariffs hit sparkling beverage and energy drink companies on a nine-month production cycle. Furniture from China has its own story. For Sotira, it’s meant more overstock to process, and some fascinating arbitrage: sneakers carry higher import duties than slippers, so add a thin layer of fuzz to the sole and the classification can change.
The Bigger Picture
What struck me most was Amrita’s point about the word “agent.” Even technical people can’t agree on what it means, and for her customers it’s a barrier, not a selling point. So Sotira doesn’t lead with AI at all. Customers just see faster deals and less paperwork.
That’s the lesson for all of us selling data and AI. Nobody buys the technology. They buy the outcome. I remember pitching VCs on data a decade ago and hearing that it was hard to invest in something you can’t see. Logistics has the same problem today. But when the backbone of the economy still runs on folding chairs and filing cabinets, the upside for anyone who brings clean data to it is enormous.
Connect with Amrita Bhasin on LinkedIn or check out Sotira.
This article was adapted from an episode of Deconstructing Data, BDEX’s weekly podcast on data-driven marketing. Tune in live every Thursday at 4:15 PM Eastern on LinkedIn.
About BDEX: For companies that need clean identity data to power their products, BDEX offers unmatched quality and execution. Visit bdex.com and click “Talk to an Expert” to get started.
Video
Watch the full episode on YouTube: AI Agents Supply Chain and the Future of Retail Logistics
