Spring naar content
Opinie artikel

The Death of Manual Feedback Analysis:

How AI is Revolutionizing Customer Feedback Processing

How to improve your customer feedback loop – it’s easier than you think!

Are you struggling to process and make sense of the customer feedback your organisation receives? Do you have a backlog of thousands of review comments and don’t know where to start? If so, you’re not alone. Our international client, active in 13 European countries, faced a similar problem and reached out to us because they wanted to get in control of the high customer demand.

The problem
The client receives too much customer feedback. This leads to not being able to process the feedback. I acknowledge how paradoxical this is in itself. But with almost a thousand written emails and reviews per month, containing very valuable information, our client did not have enough people or time or procedures in place to process them or make it measurable.

The solution
During one of our innovation sessions, an internal employee suggested that ChatGPT could be used to create order from chaos. We investigated the possibility. After some tweaking, chain prompting, and template examples, we came up with the following prompt:

  • Choose the top 10 most mentioned issues (separated by [separator]). Then for every issue give 2 examples of the following feedback
  • {feedback} [separator] {feedback} (…)

The result was exactly what we had asked for. The examples the AI chatbot returned could be found back in the collection of actual given reviews, and were not made up by the AI. Furthermore, AI doesn’t care about writing errors, language, or broken English. It actually understands what a person is trying to say.

We used the ChatGPT API and created a small program that exports the written feedback from a monthly generated Excel sheet. Instead of ploughing through walls of text, the client feedback employee (yes, that is an existing position) now has to run a script once a month and use the generated metrics however they want.
You can imagine the impact of these metrics for discovering trends, comparison to actions/promotions, IT focus etc.

Read more: