At SurveySensum, we aim to understand what the customers of our clients are saying and share instant reports with you. These reports will assist organization on what things they need to work or improve on.
So, how does this work?
- First, you upload all your data
- Then, you or our team train us by tagging verbatims for a sample of data.
- Later, the machine automatically predicts - eliminating the manual tagging process!
- At last, you can share reports with the product team and other stakeholders.
Predictions are what the machine makes after getting its due training.
For example, if the data uploaded by you has 100 texts, we would recommend you to teach the machine by creating tags and tagging at least 50 texts to train the model.
After training the machine, when Predictions are turned on, the system will take some time and will present its predictions to you.
The AI model will read the texts and try to understand what the customer will be saying and post that, the tag will be suggested and predictions will be made.
If the accuracy of the machine is not good or unsatisfactory, you can retag the predicted tag and then run the predictions to get better accuracy. The idea is to do the training so well that the machine does not make inaccurate predictions.
Note - As of now, the machine will work on all data except for chat data.
Pro Tip - We also give the users the option to give sample data to the machine to test out the predictions.
This works when the customers run the predictions and see the data, it is after they have run the predictions that they get the first value of TA Upload.