AI agent platform for an e-commerce brand in premium travel products
The marketing team at ONYX Journey was already using AI extensively, but everyone worked on their own island. Vpact built one platform where all data comes together and AI agents with full business context run analyses, fill dashboards, and proactively surface signals.
The client
ONYX Journey (part of Qumax Brands) is building a premium brand in suitcases, bags, and travel accessories. A small, fast team that does everything themselves: from product development and creative strategy to media buying and supply chain.
They sell through their own webshop and marketplaces in twelve European countries, with the Netherlands and Belgium as their largest markets.
The starting point
The ONYX marketing team was already quite far along with AI. They were working extensively with Claude and had even connected data sources. But in practice, they kept running into the same walls.
Weekly reports were assembled by hand from seven different sources. Creative analysis and briefings cost the creative strategist around twenty hours per week. And for attribution and margin insight, an expensive external analytics tool was running that the team got too little out of for what it cost.
External connectors dropped out regularly and had to be reconnected time and again.
Reports crashed halfway through, credits ran out before the work was done. Off-the-shelf AI tools from external parties burned through their credits in two weeks.
Every team member had their own setup, prompts, and way of working. There was no shared environment where data and context came together.
The solution: one AI agent platform
For Qumax Brands we built the same thing we run internally: one platform where all data comes together and where the team uses AI agents to run analyses, build dashboards, and receive proactive signals about their (marketing) strategy.
Direct API connections with Shopify, Google Ads, Meta, TikTok, and their email platform, among others. No more shaky intermediate layers, but a central data layer that syncs continuously and holds historical data too.
An agent that knows the company's KPIs, goals, and language and builds on that in every analysis. The team asks questions across all channels at once and gets substantiated answers based on the real numbers.
Every Monday morning the weekly report is ready: revenue, orders, ROAS, and spend per channel, side by side. The team starts the week analyzing instead of collecting data.
The result
Hard return figures are difficult to give in the first months, but what the team reports back is clear.
Nobody assembles dashboards or reports by hand anymore. The Monday morning report is simply there, and according to the team it delivers “really good action points” for the weekly strategy meeting.
Webshop, advertising platforms, and email live in one environment, giving the team better insight than before.
Because the model has the full context of the business, analyses are directly applicable instead of generic.
No more credits running out halfway through the month, no vendor lock-in. External tools that promised the same are being phased out.
And it doesn't stop there
A platform like this is not a one-off project. We recognize patterns in how the team uses the agents and then build new dashboards and features for them. There is already a list of next steps ready, from automatic creative performance lists for the media buyers to an AI assistant on the webshop itself.
That is exactly how we like to work together: not delivering one tool, but growing the platform alongside the team and the business.
More cases
All cases
Become a leader in your industry with AI.
Tell us your challenge.