Heading 1
Heading 2
Heading 3
Heading 4
Heading 5
Heading 6
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.
Block quote
Ordered list
- Item 1
- Item 2
- Item 3
Unordered list
- Item A
- Item B
- Item C
Bold text
Emphasis
Superscript
Subscript
The decision on the table is how to bring generative AI into your content operation, and under whose control. You can connect a large language model yourself, use what Salesforce offers natively, or combine the two. Each route shifts who owns accuracy, brand voice and cost.
What generative AI does inside Page Designer
Page Designer is the visual editor in Salesforce B2C Commerce where merchandisers build content pages and components without a developer. Adding generative AI to it means a content manager can ask for a first draft of a banner text, a category introduction or a product story from inside the component they are editing. The draft appears in the field, ready to edit. Page Designer serves SFRA storefronts and, through its content APIs, headless ones too, the route explained in headless commerce with the Composable Storefront.
FORWARD built such an integration on Commerce Cloud (Agentforce Commerce), connecting OpenAI models to Page Designer through custom editors that plug into existing components. The same module works in Business Manager, the back office, where it writes missing SEO titles and descriptions for products and categories. It runs on single items or in batches, so a prompt that works once can be applied to a full category.
SEO metadata is where the time saving is largest
Missing page titles and meta descriptions are a common gap on large catalogues, because nobody has time to write thousands of them by hand. Generative AI closes that gap quickly. With product attributes as input, a model can draft a title and description per product that is specific enough to be useful and short enough to fit the search result.
The quality depends on what goes in. A product with complete attributes gets a precise description. A product with a name and a price gets a vague one. The same discipline that makes a good product page applies here, and it is set out in product detail page best practices.
Trust, but verify every output
Language models generate text that sounds right, and sometimes it is wrong. A description can claim a material the product does not contain or a feature it does not have. On a product page, that turns into returns, complaints and in some categories a legal problem.
The answer is a review step that nobody skips. Generated content lands as a draft, a named person checks facts against the product data, and only then does it go live. Clear prompts that tell the model to use only the supplied attributes cut the error rate. They do not remove the need for a human to read what is published.
