Return on ad spend up 200%.
$450K in new annual revenue.

Every customer looked the same in a spreadsheet.

The client collects large volumes of customer data but was segmenting it with traditional, static data visualisation — methods that grouped customers into categories without showing who they actually were.

The goal was a sophisticated AI-powered tool that could process that volume of data and turn it into insight sharp enough to drive targeted marketing and forecast revenue, rather than just describe the past.

Three constraints were non-negotiable:

• Large volumes of raw customer data, still segmented manually
• Static data visualisation standing in for real insight
• No dynamic personas to make segments actionable

One engine, three ways of understanding the customer.

At the centre sat a RAG (Retrieval-Augmented Generation) inference engine, built to enable dynamic customer segmentation through AI-driven personas — replacing static charts with living representations of who each segment of customers actually was.

Feeding it was an automated ETL pipeline using unsupervised clustering, structuring targeted marketing strategy around patterns in the data rather than fixed rules. Alongside it, a Customer Lifetime Value (CLV) prediction model gave the business forward-looking revenue forecasting instead of historical reporting.

Together, the three systems turned a data-processing problem into a decision-making one: which customers to target, with what message, and what they were worth over time.

The segmentation paid for itself many times over.

RAG-powered segmentation didn’t just process more data — it changed what the business could do with it. Return on ad spend rose 200%, engagement climbed 50% on the back of AI-driven personas, and the platform helped drive a $450K increase in annual revenue. It also underpinned a $4M seed funding round, turning a segmentation project into the foundation for the company’s next stage of growth.