Consulting and delivery
AI modules for your online store, built with your team
You want AI inside your own systems instead of one more tool? I bring the AI know-how and the e-commerce know-how. Your developers build with me and can carry on by themselves afterwards.

Joffrey Petit
Founder of ecomDNAI. Built the platform in five months and runs it.
The proof
ecomDNAI is live. This is what we build with you.
5
months from the first commit to today's platform
250
backend routes for AI modules, accounts and billing
450
automated checks run before every release
8
languages in the interface and in the results
What you get
AI know-how and e-commerce know-how from one source
AI know-how
The right model for each task
Text, image, video and voice: weighing quality against price.
Cost per AI action under control
Every action has a price and is booked on the server.
Prompts treated like code
Versioned, tested and with fixed rules for the output.
Developing with AI agents
Developers steer agents, checks secure every release.
E-commerce know-how
What retailers really need
Modules that follow the daily work, no tech demos.
The rules of each channel
Knowing the requirements of marketplaces, Google and Meta.
Data from your systems
Connecting store, ad accounts and CRM so the AI works with facts.
Brand and language
Brand profile and glossary keep texts and images consistent.
How I work
AI agents, run like a company
A human gives the assignments and approves every change. The agents have fixed roles, coordinate directly and hand over in writing.
Step 1
Departments
One agent per field: advertising, video, pricing, SEO. Each one builds in a separate copy.
Step 2
Reviewers
Independent agents review code, rules and data protection. Whoever builds does not review.
Step 3
Gatekeeper
Automated checks run before every release. If one is red, nothing moves on.
Step 4
Release
Only one manager releases. First a preview, then live after your approval.
The connections
Google, Bing, Shopify and every tool with an API
Your staff sign in to their account and grant access. The AI reads the data and works with facts. Writing to an account only happens after explicit approval.
For open AI assistants I use MCP, the standard that lets an AI operate tools by itself.
- Ads, Analytics, Search Console, Tag Manager
- Microsoft
- Advertising (Bing)
- Meta, LinkedIn
- Ad accounts and campaigns
- Shopify
- Own app: products, orders, SEO, subscription
- CRM, payment
- Klaviyo, HubSpot, Emarsys, Stripe
- All others
- Every tool with an API or REST API, such as ERP, PIM, inventory management, customer service
Working together
Two ways, three steps
Option A
With your developers
I guide, your team builds. The knowledge stays in-house from day one.
Option B
I build for you
I deliver the modules with my AI team. Your developers take over operations afterwards.
Step 1
Select
Identify the two or three modules with the greatest benefit and calculate the cost per action.
Step 2
Build
Build the first module: first in a preview, then live after your approval.
Step 3
Hand over
Your team takes over. I review, advise on models and costs and bring in new ideas.
The starting point
AI workshop: two days, one clear plan
I look at your store, ad accounts and processes. At the end you have a list of the three modules with the greatest benefit, with effort and cost per AI action.
- Day 1: capture systems, data and processes
- Day 2: select modules, calculate, present the plan
- Result: a written plan your team can implement even without me
Your request
Request an intro call
30 minutes, free of charge. You tell me what should improve, and I tell you openly whether and how AI can help.
Prefer to look at the tool first? Try ecomDNAI
The request form loads with JavaScript. Without JavaScript you can reach us here: