IngramSpark Assist FAQs (Metadata Assistant)

1. What does IngramSpark Assist do?

IngramSpark Assist recommends metadata based on your title and description text, helping you quickly populate fields such as Audience, BISAC, Keywords, Thema, Short Summary, and (when applicable) Qualifying and Regional Subjects.

Here's what the IngramSpark Assist feature looks like during title setup:

2. How does the recommendation process work?

The AI converts your Title and Description into numerical representations to understand the content and suggest relevant metadata. 

3. Will my book be used to train the IngramSpark Assist AI model? 

We do not use your content files (your book’s interior and cover) to train the AI model. However, if you use IngramSpark Assist, we may use the generated metadata to improve our understanding and enhance future recommendations.  

If you choose not to use IngramSpark Assist, we do not process or use your metadata in any way related to this feature. 

4. Which metadata fields are always populated?

IngramSpark Assist always recommends Audience, BISAC, Keywords, Thema, and Short Summary. We only suggest a Regional Subject when your text includes an apparent geographic reference. We also suggest Thema Qualifiers when applicable. 

5. Which language is IngramSpark Assist available in?

IngramSpark Assist processes and returns all metadata in American English. Although the model can interpret title descriptions in other languages, the output will always be in American English. Please note that quality may vary and may not be accurate when the input is provided in languages other than American English.

6. Do I have to use the recommendations provided by IngramSpark Assist?

No. You can remove or replace any recommendation at any time. IngramSpark Assist gives you suggestions to speed up your setup, but you control all final metadata choices.

7. Can I regenerate recommendations by changing my metadata?

Yes. When you update your Title or Full Description, you can manually rerun IngramSpark Assist to generate new recommendations. Because the model is additive, we recommend deleting any existing metadata you do not want to reuse before retrying. This ensures the new suggestions are based only on your latest content and reflect your intended direction.

8. Can I use IngramSpark Assist for titles that are already available in my account (or does it only work for new title setups)?

No. Currently, IngramSpark Assist works only during the setup of new titles. It does not apply to existing titles in your account.

9. How can I give feedback if the recommendations provided by IngramSpark Assist are not relevant?

During the beta period, we will invite users to share feedback. This helps inform improvements to the accuracy and usefulness of future recommendations.