Skip to main content
Configure knowledge-base FAQ refresh jobs in the preezie CMS to keep answers synchronised with selected pages on your website.
This guide covers the conversational knowledge base in Brand manager. To generate product-page FAQs and highlights for the product catalogue, see Configure FAQ and Highlights jobs.

​
Before you begin

Select the tenant you want to manage, then open Brand manager and choose the Knowledge base (FAQs) tab. The automatic job refreshes existing FAQ answers. It does not create questions that are newly discovered on a source page. Run an initial website scrape or add the FAQs manually before enabling automatic updates.

​
Configure source URLs

  1. Select Manage Scraping.
  2. Select Auto-update Settings.
  3. Under URLs to scrape, select Add URL.
  4. Enter the complete URL of a page containing FAQ content, such as https://example.com/help/shipping.
  5. Enter a descriptive label, such as Shipping FAQs, and select Save.
  6. Repeat these steps for every page whose existing FAQ answers should be refreshed.
At least one source URL is required. Both the URL and label are required for each entry. Use the edit and delete controls to maintain the source list.
Add the most specific pages that contain the source answers. A focused source list reduces unnecessary scraping and makes it easier to identify where an answer originated.

​
Set the update schedule

In the Update schedule section:
  1. Enter a positive interval value after Every.
  2. Select days or months as the interval unit.
  3. Turn on Schedule active.
  4. Select Save settings, then confirm the change.
For example, use 7 days for a weekly refresh or 1 month for a monthly refresh. Month intervals must be between 1 and 12. Saving an active schedule with at least one URL creates or updates the recurring job. The time at which you save becomes the schedule anchor. The settings panel displays Next scrape after the schedule has been saved.
The recurring job runs in UTC. The CMS formats the next-run timestamp for display in your local time zone.

​
Pause or change a job

Open Manage Scraping → Auto-update Settings again to change its URLs or interval.
  • Turn off Schedule active and save to remove the recurring job while retaining the configured URLs and interval.
  • Turn the schedule back on and save to create the job again with a new schedule anchor.
  • Saving changed URLs replaces the job’s complete source list. All URLs shown in the panel are processed by the next scheduled refresh.

​
What a scheduled refresh changes

During each run, preezie scrapes every configured URL and tries to match the results to existing FAQ records using the question title and source URL.
  • Only the answer content of a matching FAQ is updated.
  • Answers marked as manually edited are not overwritten.
  • Newly discovered questions are skipped.
  • Unchanged answers are left as they are.
  • If the schedule is inactive or contains no URLs, the refresh performs no work.
This behaviour lets you protect curated answers while keeping imported answers aligned with their source pages.

​
Run a full scrape manually

Use Manage Scraping → Run Now when you need to scrape the tenant’s entire configured website and discover new FAQ content. Confirm the warning to enqueue the import.
Run Now is different from the scheduled answer refresh. It can scan pages such as shipping, policy, and returns, create a significant number of FAQs, and incur additional charges. Review whether targeted manual updates are more appropriate before starting it.
While the full scrape is queued or running, the CMS displays Scraping in progress and prevents another manual scrape from being started. The FAQ list refreshes after the job leaves its active state.

​
Troubleshooting

If an expected answer is not updated after a scheduled run, check that:
  • Schedule active is on and Next scrape is displayed.
  • The source page is still present in URLs to scrape and can be accessed without signing in.
  • The existing FAQ title and source URL correspond to the content on the source page.
  • The FAQ was not manually edited, because manually edited answers are protected.
  • The source answer actually changed; unchanged content is not rewritten.
To import a newly added question, use Run Now, upload it, or add it manually. The recurring refresh deliberately does not create new questions.