Preference Learning is an opt-in, per-user feature that stores small fragments of a user’s review interactions to personalize future review suggestions. Data is not used to train models across customers — it’s stored only to personalize that individual user’s experience.

1. Click on the Spellbook icon, and scroll down to "Preference Learning." Toggle it on or off accordingly.

Users can delete their stored memories/preferences from the same settings area at any time. Deleting removes previously stored fragments.

2. You can then run a Review with this feature running

3. Alternatively, the first time you run a review without the feature, a card will appear at the top of the Reviews list to Enable or Dismiss — this explains the feature and links to Settings.

4. Smart Suggestions are generated by comparing previously accepted review text to target paragraphs in new review suggestions — matches can be surfaced as “smart suggestions.”

Q: Is my data used to train models shared across customers?
A: No. Data is used only to personalize the individual user’s experience.
​

Q: Can I opt out later?
A: Yes — toggle the feature off in Settings. No new data will be stored after it’s off.
​

Q: Can I remove previously stored preferences?
A: Yes — you can delete stored memories/preferences from Settings at any time.
​

Q: How long is the data stored?
A: Fragments are retained for 1 year after generation.
​

Q: What prevents collection?
A: Either the user turning the feature off or by informing the Spellbook Success Team at success@spellbook.com