Overview
Amfibija is a TCG specialty retailer in Serbia — trading card games, mostly. The category is unusual in a single respect that dominates everything else about the business: prices change daily, sometimes hourly, across hundreds of SKUs. That requirement is what no commerce platform supports without custom work — except oozmi, which Amfibija chose in 2021.
Five years on, the platform runs every part of the retail operation — ecommerce, CRM, ERP, OMS — and every catalog and business change since has shipped through the integrated AI chat. Not a single developer ticket.
Technical context
Trading card games are priced like a small secondary market. A new set release, a tournament result, a print-run rumour — each can move card prices within a day. The retailer either tracks those moves into the storefront or sells at the wrong price. Doing it manually doesn't scale; doing it natively requires the platform to treat per-SKU dynamic pricing as a first-class concept rather than a custom field on a generic product schema.
Today Amfibija runs amfibija.rs on oozmi end-to-end. Ecommerce, CRM, ERP, and OMS are all in production on a single instance. The pricing logic the team needs is built in, not bolted on.
Challenges
Every commerce platform Amfibija evaluated in 2021 could sell things — but none supported the way TCG items are actually sold. The team had two ways to make a generic platform fit, and both were unacceptable.
Path A: Salesforce. An enterprise-tier licence that would unlock the flexibility, at enterprise-tier cost a specialty retailer running on TCG margins couldn't justify.
Path B: Magento or WooCommerce, customised. Two separate agencies quoted around €100,000 in custom development to graft daily-changing pricing onto the standard product schema. That spend would be pure programmer hours, with more to follow every time something changed.
Neither path made sense. The third option was to pick a younger platform that already treated pricing as the thing being sold — that platform was oozmi.
Solution
Amfibija chose oozmi in 2021. The platform handled daily-changing TCG pricing natively from day one — the specific operational requirement no one else supported without custom work. The choice was a bet on a young platform; five years on, the bet has held.
What's more interesting is how the operation has evolved. The global TCG market has changed substantially since 2021 — supply chains, exclusive distribution, new product lines, secondary-market dynamics. Amfibija's catalog and business model changed with it. Every one of those adjustments was made by the team through oozmi's integrated AI chat. The team describes in business language what should change; the platform absorbs the change. No programmer in the chain.
That includes the things any other platform would have routed through engineering: mass updates to hundreds of SKUs at once, new product lines shipped end-to-end through the admin (parameters, filters, search, landing pages), catalog rule changes that ripple across the storefront in seconds.
Key Results
The headline number is the budget that wasn't spent. Amfibija avoided ~€100,000 in custom development — what a Magento or WooCommerce path would have cost on day one to do the same thing — and the equivalent of that spend every time something changed since. Pricing flexibility comparable to enterprise-tier suites, without the enterprise-tier licence Amfibija couldn't justify.
The team is the second result. Zero developers in the chain across five years of catalog and business changes. The merchandising and operations functions describe what should change in business language; the AI chat builds it. The work that would have been engineering tickets on Magento or WooCommerce — or licence-line items on Salesforce — sits inside one tool, on the schedule of the people who actually run the shop.
What's next
Amfibija doesn't have a roadmap of platform migrations — they already use the four modules they need. What's next is whatever the next TCG market shift demands. The point of the architecture isn't where it ends; it's that the next change ships the same way every change before it has: through the team, through the chat, without engineering in the path.
For anyone evaluating oozmi for a similar specialty operation — where the business has a hard requirement that generic platforms route through custom development — Amfibija is the five-year proof that the native-fit bet pays out.