AI Product data, structured. About Mavitra AI · Commerce software since 2007 · AI in production since 2016 We're building the infrastructure that makes product data usable.
Retailers hold the facts about every product they sell — in images, descriptions, supplier sheets. Almost none of it is structured, so filters miss, ads mismatch and marketplace listings fail. Mavitra AI turns that stranded information into the data your store, your feeds and eBay can actually use. We have been shipping AI inside commerce systems since 2016; this is the same discipline applied to the data layer every retailer depends on.
Product data was meant to sell. Mostly, it just sits there.
Every catalog we've looked at tells the same story. The drive size is in the description. The fitment is in the title. The pressure rating is on the spec sheet. None of it is in a field a filter, a feed or a marketplace can read, so the customer who searches for exactly what you stock never finds it. That's real revenue, earned by real products, going to waste.
Unstructured in. Structured, verified, written back.
Mavitra AI reads each product the way a good buyer would, extracts the attributes that matter in your category, scores each one with its evidence, and writes the approved values to your Shopify metafields. Filters work. Feeds match. Fitment exports. Nothing of yours is overwritten; where your own fields are blank, we fill them. The goal is infrastructure: data that moves freely from your catalog to every place it needs to be.
Commerce and enterprise software since 2007.
Mavitra has been building serious commerce and enterprise systems since 2007, from the Sunshine Coast to Dubai: retail platforms, data pipelines, payments and loyalty infrastructure for organisations that cannot afford to get it wrong. We have run AI in production since 2016, long before it was a headline, in retail search and product data. Mavitra AI applies that depth to a problem every retailer has and few have solved well. More about Mavitra →
Dean Harris
Founder & CEO
Mavitra was founded by Dean Harris, who has spent more than three decades architecting software, data and commercial systems for enterprises and government across Australia and globally. He builds at the intersection of commerce, data and financial systems — infrastructure that lets value move more freely between businesses and the people they serve.
Mavitra AI started in the automotive aisle. Dean races, and has spent years on the buying side of parts catalogs where "fits my car?" is the only question that matters and the answer is rarely in a field. Structured product data is, in his view, the most under-built layer in retail: everyone depends on it, almost no one owns it.
Read more about Dean →How we build.
Your data stays yours
Everything we derive is written to your store, under your account. Leave, and you keep it. Nothing is used to train models.
Evidence for every value
Each attribute carries a confidence score and the evidence it came from. Below threshold, a person decides. You can audit any value back to its source.
Built and hosted in Australia
Processed and stored in Azure Australia East, with the same security, isolation and audit discipline we bring to financial systems.
Operating from the Sunshine Coast, Australia and Dubai, UAE. Building for Australian retailers first.
See what your catalog is hiding
100 of your products, enriched, with a report on what's missing and what it's costing you. Free, no install.