Aizy is a performance marketing agency in Breda, Netherlands, that runs paid ads and SEO for hundreds of clients on Google, Meta and TikTok. Its own AI software handles campaign optimisation while human marketing experts set the direction and review results. The company started in 2024 and now works with clients across retail, e-commerce, health and B2B.
Every Aizy client gets a WhatsApp group with their account team in it, and that group is where briefs, approvals and performance questions all land.
Any agent could reply, so often nobody did
Aizy was already running on Respond.io, built on the WhatsApp Business API, which does not work with regular WhatsApp groups. So the client groups, where all the actual work happened, sat outside that setup completely.
12 people worked them through a single shared number.
Because no message belonged to anyone in particular, a client question could be picked up by any agent, which in practice meant it was often picked up by none of them. Requests sat in the group until the client came back and asked again.
Urgent messages carried no signal either. With groups active right through the day, a campaign problem and a casual thank-you looked identical in the list, and the only way to tell them apart was to read the whole thread.
For an agency whose entire promise is hands-off, a client having to chase is the one thing that undoes it.
Managers had no record of any of it
That same blind spot ran up to management. There were no response times to look at, no count of what was still open, and no record of how any individual account had been handled. When a client asked why something had slipped, the answer had to be reconstructed by scrolling.
Aizy wanted three things from a new setup: ticketing, analytics, and an AI agent that could take the repeat questions. All of this had to work inside the client WhatsApp groups.
Aizy runs the same client groups on the same number it always did. What sits on top of them is new.
With every message now carrying an assignee, account managers stopped re-reading threads to work out what was still open, and they now start the day on a ticket queue with owners and due dates.
When a ticket is raised, the person it belongs to is notified straight away, and the ticket stays attached to the conversation it came from, so the context travels with it.
The AI agent took the repeat questions off the team entirely. It answers what it can from the knowledge base, raises tickets for the real tasks, and escalates anything it cannot handle, which leaves the human experts on the accounts that actually need them.
Leadership gained something it did not have at all before. On a monthly client call, an Aizy manager can now answer how fast the team responded and what is still open on that account, without relying on memory.
