Interview Ludovic Genest, Cloud Software Manager at Orisha Commerce France

In the fall of 2025, we interviewed Ludovic Genest, Cloud Software Manager at Orisha Commerce France, to understand how APIs are transforming the implementation of recommerce in retail.

In this conversation, he discusses the key challenges of architecture, interoperability, and governance that enable retailers to seamlessly connect their buyback, workshop, or resale modules to their information systems.

Summary:

• The principles of a composable architecture
• The essential endpoints for recommerce
• API security and governance
• Quick wins and real-world feedback
• The 3-year outlook for connected retail

Introduction: framing the topic

RE-commerce involves a range of activities including selling, repairing, renting, and reselling. We’re introducing a new product line that manages these actions through modular systems. These modules are designed to interface seamlessly with the existing products of our subsidiaries (Ginkoia, Openbravo, Fastmag), maintaining independence from the ERP or POS. APIs are the best way to facilitate communication between these modules and the ERP or POS.

At Ginkoia, we’ve adopted a new philosophy by providing not just APIs but also a developer portal through the Ginkoia API Center to support our clients better.

Leveraging our RE-commerce functionalities through APIs can enable IT systems to create flexible solutions. Nevertheless, working with APIs demands technical expertise not only during the initial implementation but also over the long term. Grasping the use and functionality of APIs is a significant challenge that must not be overlooked.

Architecture & interoperability

It’s crucial to clearly delineate the front end from the back end, ensuring the user interface utilizes APIs that might originate from different sources. Intermediate APIs closer to the functional level should connect to lower-level technical APIs. This facilitates the use of various APIs in the lower layers deriving from different ERP or POS systems, without disrupting interfaces and functionality.

First, verify that your ERP or POS encompasses adequate APIs and ensure that all functional actions are executable through these APIs.

In my workshop, I generate a workshop order, which helps manage tasks, technicians, and statuses. The order can also handle checkout payments and automatically update its status post-payment. This entire workflow is managed through API/webhook flows.

Nowadays, tools like service buses prevent direct API calls. Instead, we utilize service buses and webhooks to support asynchronous and persistent services.

When developing APIs, we often emphasize CRUD APIs (Create, Read, Update, Delete) and may believe the task is complete when we’ve provided these for all managed objects. However, crucial APIs are often overlooked, such as functional APIs, multi-criteria search APIs, and batch processing APIs.

Endpoints and critical data

Employ CRUD and GraphQL APIs for swift searches. 

Crucially, maintain abstraction layers between data layer APIs and business layer APIs. Expose business layer APIs to ensure that modifications in data layers don’t affect business layers. 

Classify APIs based on data types, like MasterData versus TransactionData. 

MasterData APIs should excel in searches, supply increments, and facilitate large data reads (using cache). 

TransactionData APIs should emphasize high-frequency creation (using queues).

Data consistency is a central challenge and difficult to tackle, requiring proper constraints. For example, analyzing sales over the past three months may tolerate missing the last quarter-hour of data. However, stock analysis, management control, and inventory demand high data precision. Employ consistency control tools regularly, such as running automatic checks to identify discrepancies and scheduling global synchronizations during less busy times.

Security, governance, and compliance

Implement robust authorization systems and tokens, ensuring that no code contains sensitive information (URL, login, API key); instead, use vault mechanisms for storage. 

API key management for third-party apps should also adhere to standards. Numerous solutions allow delegation of these tasks, negating the need for custom development, managed by the API Manager.

APIs have no unique GDPR requirements; follow the same principles as direct database access. 

However, cleanup APIs that remove personal data when no longer needed are often overlooked. 

Be cautious about logs that may record details of API calls containing personal data and manage them carefully.

Utilize an API manager to manage all functionalities of an API engine, avoiding custom coding for these mechanisms. 

For instance, an API call quota should be managed natively by the API manager, not coded directly in the API.

ROI, deployment & “quick wins”

Invest time in choosing the right integrator or ensuring your technical teams understand the philosophy behind the proposed APIs for efficient integration. 

It’s essential to have a robust technical and functional vision to implement reliable and scalable solutions.

At Orisha Passy, we consistently offer presentation workshops, saving valuable time for everyone involved. This helps the integrator avoid heading in the wrong direction.

We track KPIs like load times and the number of queries, which are standard but crucial. 

As volume increases, consider optimizations such as implementing Redis Cache across various APIs that share data to minimize database access and improve response times. While this technique is complex and necessitates a robust Azure infrastructure, it becomes essential as volume grows.

Without naming the brand, I remember an API designed for creating customer orders for sales (order placement, customer creation, product storage, invoice issuance). 

An integrator used this API for rentals, leading to ski equipment sales in the ERP at daily rental rates with storage. This underscores the importance of clear communication between functional teams, technical teams, clients, integrators, and API providers.

Don’t view APIs as a magic solution for all issues. 

• Recommendation 1: Utilizing APIs doesn’t eliminate the need to design your project. 
• Recommendation 2: Plan and define processes for managing changes (like version updates) with your teams or integrator to anticipate rather than react to them. 
• Recommendation 3: Deploy Mock testing for APIs to prevent conducting tests in a production environment.

Vision & perspectives

The importance of APIs will continue to grow, requiring tools for effective data flow tracking to ensure consistency and coherence. 

In the future, every project will involve APIs. Despite the potential stagnation in standards, technologies and tools will evolve rapidly.

We need to break down business functions instead of relying on a single, all-encompassing solution. Creating complex systems is challenging and requires moving away from the idea that one software can handle everything. We need more architects and fewer buyers of complete solutions.

Define your objects clearly and establish data-driven mapping between external and internal objects.

Bonus:

“Sure, my provider has all the APIs, so the project will be easy.” But if you’re using APIs from two different vendors, how have you planned for them to communicate with each other?

It’s more about methodology: prioritize defining the architecture and conducting design work before any development begins.

Load and performance testing are often overlooked. I remember a situation where a tool was used to send tickets for BI. After modifying my POS to send tickets upon creation, it worked. However, during the first Saturday of sales, the ticket creation time ballooned from 2 seconds to 45 seconds due to slow ticket dispatch and numerous retries caused by unresponsive APIs. Proper load testing would have highlighted the need to delegate ticket dispatch to a different system from my POS.