AICommon Sense AILarge Language Models (LLM)

Automating Marketing Material Classification in Pharma

Click to navigate to the LinkedIn page of Jean-Jo Adjizian.
Jean-Jo Adjizian
Managing Partner - Common Sense AI

Challenge – 

A leading pharmaceutical company faced significant challenges in managing its extensive library of marketing materials. These materials needed to be carefully classified into specific categories based on their content and intended messaging to ensure efficient reuse and compliance with communication guidelines. However, the classification process was entirely manual, making it a labor-intensive and time-consuming task.

The manual approach often led to inconsistencies, errors, and unclassified materials, which resulted in wasted resources and missed opportunities to leverage valuable assets. Teams struggled to locate relevant materials quickly, delaying marketing campaigns and impacting overall efficiency. As the volume of materials grew, the limitations of the manual system became increasingly evident, highlighting the urgent need for a streamlined, automated solution to improve categorisation and facilitate seamless reuse.

Solution

To make the classification process more efficient Algorhythm Group developed a pipeline, processing the marketing materials and submitting their contents to a large language model (LLM), to automatically propose classifications.

The pipeline was engineered to fulfil the requirements:

  • Robustly and efficiently handle large requests (batches >1000 materials).
  • Deterministic LLM response format.
  • Automatically detect a wrongly formatted response -> engage in a conversation with the LLM to attempt a correction.

Impact

The solution is able to automatically provide classifications for any desired set of materials.

Thanks to our intelligent automation solution, the team members can now concentrate their time on other and more important tasks.

Delen

Continue reading