3 September 2024
Here is a description of some of the updates since the last product news.
PIM
Maintain filters directly from the widget form
You can now create or modify data set filters directly within the widget form, simplifying the process of managing filters. When setting a filter, users are redirected to the PLP (product list page), where they can preview the filtered list. Once the filter is saved, it will automatically apply to the widget form.
See more information about dashboards here.
Better navigation in mass edit
To make it as quick and easy as possible to navigate between different fields in mass edit, we have improved arrow key navigation. Users can now seamlessly move between fields using the arrow keys, reducing the need for mouse interaction. This enhancement streamlines mass editing workflows and speeds up the editing process, particularly when working with large data sets.
This update will significantly improve efficiency and reduce the time spent on repetitive tasks.
See more information here.
Improved import through the UI
When importing CSV data via the UI, you can now run enrichers as part of the process.
You can also download source files for jobs that require a file for import, such as CSV imports, making it easier to review previous imports and debug any issues.
Ecommerce Search
Several improvements to suggestions have been made to Ecommerce Search:
- Suggestions now have improved support for elisions, such as recognising "L'Oréal" when users search for "Loreal".
- Suggestions that do not result in a product or content match are now filtered out, ensuring the response only includes relevant options.
- Improved suggestion relevance, providing more accurate and contextually appropriate suggestions.
- Enhanced product relevance, preventing multiple matches on the same word from creating an unfair advantage.
There have also been several performance improvements to both search functionality and data ingestion processes.
For developers
BAIA custom functions
Previously, BAIA (Bizzkit AI Assistant) focused on the user interface, offering features such as natural language interactions, user recipes, system recipes, and semantic functions (e.g., in PIM).
However, BAIA has now introduced a new capability that allows you to interact directly with it via the API. This feature is known as custom functions.
Custom functions are designed for developers building AI functionalities, such as:
- Pre-selecting optimal values in forms within a backend administration site.
- Drafting responses to customer support inquiries.
- Providing chat support to end customers, for instance, answering questions about opening hours or return policies.
While it’s possible to interact directly with services like OpenAI or Gemini, the advantages of using BAIA custom functions are clear:
- You don't need to write complex boilerplate code to integrate AI – just use a few simple endpoints in the BAIA API.
- You don’t need to manage new subscriptions or API keys – you can rely on your existing BAIA licence.
- You don’t need to worry about security, GDPR, or compliance – BAIA handles these concerns for you.
- You don't need to monitor model updates or decide on upgrades – BAIA takes care of this automatically.
BAIA custom functions are currently in preview and available via the preview API for you to explore.
PIM
The endpoint POST /api/resolved-views/attribute-field-values/-resolve
now accepts segmented PIM types. This will make it easier for consumers of the POST /api/resolved-views/{entity}/-search
endpoints to resolve referential attribute values since you no longer need to "translate" segmented types into their corresponding base types.
DAM
There is a few enhancements to the API:
- The endpoint
/api/{culture}/attributes-values/search
now include height, width and image size in bytes. - The endpoint
/api/_/cultures
allows you to list content languages / cultures.
CMS
We have created a new NPM package that includes a signIn
method, allowing plugins to easily obtain a bearer token for interacting with other Bizzkit products. This simplifies integration with the CMS when using TypeScript. For more details, see the documentation page.