Shopping Agents Know How to Buy. I Built WebMCP for What Happens Next.
Shopping agents are getting surprisingly good at the part of commerce that happens before checkout.
They can search.
They can compare.
They can recommend.
They can help us buy.
But commerce does not end when the payment succeeds.
Sometimes the camera arrives damaged.
Sometimes the wrong variant shows up.
Sometimes the package is late.
And suddenly the experience falls back to:
Find your order number.
Open a support page.
Choose a category.
Fill out another form.
Explain everything again.
That gap became the idea behind Post-Purchase Resolution.
What if a customer could simply describe what went wrong, and their agent could understand the website well enough to help — without silently becoming the customer or the merchant?
That is what I built with WebMCP + Shopify.
🎥 Demo: https://youtu.be/nbCvtMemO8k
🚀 Live app: https://post-purchase-resolution.vercel.app/
💻 Source: https://github.com/jpablortiz96/post-purchase-resolution
The idea
Imagine this purchase exists in your Shopify account:
Portable 4K Action Camera — $149
It was paid, fulfilled, and delivered.
But it arrived damaged and will not turn on.
Instead of asking the customer for an order number, I wanted the interaction to look like this:
“The action camera I bought recently arrived damaged and won't turn on. Can you find the purchase and tell me what I can do? Don't submit anything.”
That sentence contains no order ID.
No Shopify terminology.
No API language.
No instructions about which tool to use.
And that is exactly the point.
The website exposes structured capabilities through WebMCP, so the agent can discover what the site actually knows how to do.
The final WebMCP surface is intentionally small:
text
find_order
get_order
prepare_resolution
