Design daily FAQ maker with AI.

In this tutorial, the daily FAQ with ranking will be explained.

Step 1 : Gathering all data

In this step, it is mandatory to store all answers and questions in the database. We need to store all answers to each question to gather all data. It is much better to use docker and PostgreSQL to execute this process daily. Also, this process could be a Cron job to do every day.

Step 2: Clean up data

In step 2, it is necessary to remove outlier data, for instance, advertisement data and changing department and postpone the problems to future.

Step 3: category “guidance” , “ask for more details”

In this step, answers are classified into two groups: the first seeks more detailed questions from the client, and the second gives the client guidance. For this action, it is possible to find the answers using special words or prompt engineering to find the category.

Stats

After these three steps, it is possible to make stats for the situation.

Step4: cluster questions

In this step first, the question is summarized by the prompt engineer and will cluster by K-means and will find common topics.

Step 5 : cluster answers

In this step first, the answers will cluster by K-means and will find common topics.

step 6 : clean questions and answer

Sometimes, personal data or unnecessary data exists in questions and answers, and it is mandatory to remove them. We can use a prompt engineer or manually clean up the data.

Step 7 : final review FAQs

In this step manually check the FAQs to figure out everything will be ok.

Step 8 : make dataset for FAQs

This is a last part and export csv for the result.

Visit us at DataDrivenInvestor.com

Subscribe to DDIntel here.

Design daily FAQ maker with AI.