Client feedback  analysis

Analysing feedback to understand clients' sentiment is not new. However, providing insights beyond simple "positive, neutral and negative sentiment" analysis provides a far more powerful tool. Our consultant developed AI Large Language Model (LLM)-based Proof of Concept that could carry out this analysis.

In consultation with the business, two types of solution were developed: i) tailored software that could support multiple LLMs (including OpenAI's ChatGPT, Anthropic's Claude, Google Gemini, and local LLM) and that using agent-based AI.

Multiple iterations were developed, with business feedback improving the outcome of each iteration.

Not only can the solutions analyse a client's feedback for multiple sentiments, but they can also categorise the responses - either using AI to suggest categories, or for the responses to be matched against user-defined categories. Finally, the AI solutions can create draft email responses for relevant business personnel to review.

Both Proofs of Concepts were very well-received and are currently being considered for implementation.