Built for a Catalogue of One: Agentic Commerce and the Reseller Economy

Most conversations about agentic commerce start with a tidy little scenario. A shopper asks their AI assistant to reorder washing-up liquid, the agent finds the usual brand, checks the price, pays, and a box arrives on Thursday. Lovely. It also happens to be the easiest version of the problem, because washing-up liquid has a barcode, a fixed price and a few million identical siblings.

The reseller economy runs on something far messier and far more interesting: items that exist exactly once. A 1970s Denby teapot with a hairline chip. A Barbour jacket with one replaced toggle. A box of mixed Lego that may or may not contain a complete Millennium Falcon. For anyone selling pre-loved goods, every item is its own product launch. That's precisely why agentic commerce matters more here than almost anywhere else.

So what is agentic commerce?

Agentic commerce describes AI agents that take action on someone's behalf across the buying and selling process. Earlier AI tools answered questions and suggested things. Agents go further and do the work: they look, decide, publish, adjust and follow up, with a human setting the goals and stepping in for the judgement calls.

For second-hand sellers, I'd define it like this:

“AI that acts on a seller's behalf to turn one-off, unrepeatable items into sold stock, handling the identifying, describing, pricing, listing and selling that used to depend on a person's time and expertise.”

The listing tax

Ask any reseller where their week goes and you'll hear the same story. Sourcing is the fun part, with the car boot at dawn, the charity shop rummage and the thrill of spotting a sleeper. Then comes the listing tax. Every item needs photographing, researching, measuring, describing and pricing, and every one of those steps asks for a little expertise and a lot of patience.

Multiply that across hundreds of items and the maths gets brutal. Stock piles up in spare rooms and back rooms. Good pieces get underpriced because nobody had time to check the sold listings. Great pieces get a lazy title and sit unseen for months. The bottleneck in second-hand selling has always been the gap between having an item and having a listing, and that gap is exactly where agents earn their keep.

What a seller's agent actually does

A useful way to picture it is three layers of work, each of which used to need a person.

Understanding the item. The agent looks at a photo and works out what it is, who made it, roughly when, and what condition it's in. This used to be the domain of the seasoned dealer or the volunteer who just knows her Royal Doulton.

Deciding what happens next. The agent checks live market data and recent sold prices, recommends a price, and chooses the right home for the item, whether that's eBay, another marketplace, the shop floor or the bargain bin. Then it writes a listing that both people and search engines can read.

Running the sale. The agent publishes the listing, keeps an eye on how it performs, adjusts the price when an item sits too long, handles questions and offers, and flags anything that needs a human decision.

The seller stays firmly in charge. She sets the strategy, approves what matters and applies her eye for the unusual. The agent absorbs the repetition, which is the part nobody got into reselling for.

Meanwhile, the buyers are sending robots

Here's the twist that should make every reseller sit up. Agentic commerce is arriving on the buy side too. Shoppers are beginning to hand their searches to AI assistants: find me a size 10 waxed jacket, good condition, under £60. Those assistants read structured data. They look for accurate brands, sizes, materials, conditions and prices. A beautifully photographed item with the title "lovely vintage coat!!" is charming to a human browser and close to invisible to an agent.

As more buying flows through AI, the sellers whose listings are detailed, accurate and machine-readable will be the ones who get found. Sell-side agents become the thing that keeps one-of-a-kind stock discoverable in an agent-mediated market. That makes this a question for right now, well before it becomes a question for everyone.

What stays human

Agents are excellent at repetition, research and consistency. They're less good at walking into a house clearance and sensing which box holds the treasure, building a loyal following on Instagram, or deciding that a particular piece deserves a story and a premium. The reseller's eye, taste and relationships remain the heart of the business. Agentic commerce simply hands that person back the hours the listing tax used to take.

For charity retailers, the stakes are even higher. Every donated item that sits unlisted is income a cause never receives. When an agent can identify, price and list stock quickly, more donations reach the right buyer at the right price, and volunteers get to spend their time on the shop floor and with the community rather than wrestling with product descriptions.

Getting ready

If you sell second-hand, a few habits will serve you well as agents become part of the workflow:

• Take good, consistent photos. Agents see what you show them, so clear shots of labels, makers' marks and flaws pay off immediately.
• Keep your pricing honest and data-led. The market rewards sellers who price against real sold data, and agents make that effortless.
• Think in structured detail. Brand, size, material, era and condition are the language buying agents speak.
• Choose tools that close the loop. The biggest gains come from agents that carry an item all the way from photo to live listing, rather than handling one step and handing you back the rest.

The catalogue of one, at catalogue speed

Big-box retail spent decades building systems for selling a million of the same thing. The second-hand economy never had that luxury, and it's been paying for it in time, missed value and stock that never saw daylight. Agentic commerce finally gives one-of-a-kind items the treatment catalogue products have always enjoyed: fast, accurate, well-priced and findable.

That's the future we're building at ListAid for charity shops and resellers, with an AI web app that identifies items from a photo, prices them against live market data and gets them listed on eBay in moments. The stock was always valuable. Now it can move at the speed it deserves.