The University of Leeds and SR Mailing Limited KTP 24_25 R4

A mailing company is teaching a large language model to sift through its customer data and tell the sales team exactly what to sell to whom. The problem is straightforward: SR Mailing Limited, like many firms, sits on a mountain of transactional data but lacks the tools to turn it into actionable insight. Standard industry practice relies on broad-brush marketing campaigns that miss individual customer preferences. This project embeds an AI—similar to the engine behind ChatGPT—directly into the company’s data analysis workflow. The model will learn patterns from past purchases, returns, and browsing behaviour, then generate precise recommendations for sales and marketing teams. If it works, the company gains a substantial competitive advantage by disrupting standard industry practices. The immediate impact is commercial: more relevant offers, fewer wasted mailings, and higher conversion rates. But the approach has wider implications for any business that handles customer data—retailers, utilities, subscription services—where the quiet infrastructure of marketing automation could become far more personalised without requiring a data science team on staff. This is applied research with a clear commercial endpoint, not fundamental science.