The problem field of catalogue integration
By Andreas Skorski, founder and CEO
At The List, we built the <GENESYS> API to solve the pain points of the luxury retail industry. Catalogue integration is a big one. Here are some of the ways <GENESYS> addresses these pain points.
—
1. 𝐌𝐞𝐫𝐜𝐡𝐚𝐧𝐭𝐬 𝐝𝐨𝐧’𝐭 𝐧𝐞𝐞𝐝 𝐭𝐨 𝐜𝐡𝐚𝐧𝐠𝐞 𝐚𝐧𝐲𝐭𝐡𝐢𝐧𝐠 𝐚𝐛𝐨𝐮𝐭 𝐭𝐡𝐞𝐢𝐫 𝐩𝐫𝐨𝐝𝐮𝐜𝐭 𝐝𝐚𝐭𝐚 𝐬𝐭𝐫𝐮𝐜𝐭𝐮𝐫𝐞. ✔
The computer vision LLM of our API reads images in order to classify products and maps it to our marketplace’s category tree structure.
It also norms the attributes. So the 18 versions of green in the fashion space are not a problem for us. 🍏
With this approach, the product description can even be in Italian, or have no description at all.
Based on an image alone, we can create 120 attribute families for a product. 🖼
To get set up on THE LIST takes days. Other marketplaces require months.
2. 𝐓𝐡𝐞𝐲 𝐝𝐨𝐧’𝐭 𝐧𝐞𝐞𝐝 𝐭𝐨 𝐛𝐮𝐢𝐥𝐝 𝐚𝐧𝐲𝐭𝐡𝐢𝐧𝐠 𝐚𝐭 𝐚𝐥𝐥, 𝐚𝐜𝐭𝐮𝐚𝐥𝐥𝐲 💆💆♀️
We already have 60+ APIs to common ERP systems, but in case we don’t, our integration team builds the connection to our partner’s ERP system.
3. 𝐎𝐮𝐫 𝐦𝐨𝐝𝐞𝐥 𝐢𝐬 𝐨𝐧𝐞 𝐨𝐟 𝐭𝐡𝐞 𝐦𝐨𝐬𝐭 𝐚𝐜𝐜𝐮𝐫𝐚𝐭𝐞 𝐢𝐧 𝐥𝐮𝐱𝐮𝐫𝐲. 🎯
A lot of people ask: why can’t we just install Gemini, OpenAI, Llama or other LLM models and be up and running?
The problem is commercial viability. Which can only be achieved with:
🔢 A lot of data to train the models on
🕦 Time — years of tuning and tweaking, as the fashion industry has complex cycles
We now have one of the highest accuracy rates in the industry — with 98.9% accuracy we detect, translate and enrich the right product data.
How do we know we are not wrong, given such a large number of SKUs? We have multiple models controlling each other, trained on millions of catalog data points.
