AI model training

AI model training

Learn how to train your AI model to properly interpret data and provide accurate results
Last Updated February 14, 2023
Table of Contents

An AI model is used to get predictions on new queries for which you do not know their answers. When you train AI, you’re teaching it to properly interpret data and learn from it in order to perform this task with accuracy. Just like with humans, this takes time and patience.

How to train the AI model

1. Feed

There is a list of articles that can guide you on how to perform good AI training. You need three ingredients to train AI well: high-quality data, accurate data annotation and a sense of experimentation.

Once you’ve added/edited/deleted entities, intents, and examples, you’re ready to make your bot learn!

2. Train

AI models are automatically trained every 2 hours. If you want to manually train your AI model to speed up your testing:

  • Go to your Artificial Intelligence > Models.
  • Then press on the Train icon if the status is ‘Pending’ or ‘Failed’.
  • Training will start. A progress bar will appear until the status changes to ‘Success’.
  • If the status become Failed, try solving AI training commmon failure.

3. Evaluate

Use the AI training analytics section to evaluate your model and make any changes if needed.

ℹ️ Any change in the Tactful Admin dashboard profile configurations takes about 5 minutes to reflect on the profile

ℹ️ The conversation model must have at least 2 valid intents to work for each profile-supported language.

Want to learn more?
Was this article helpful?