This article explains how to use AgenticCatalog to automate the push of your product feeds to AI systems. AgenticCatalog is currently available in beta.
Overview
The AgenticCatalog solution automates pushing your product feeds to AI systems. This transforms your product data into an AI-ready catalog, enabling AI agents to easily discover your brand. By merging Botify's data with your existing merchant feed, you’ll help your brand be found by AI agents, a trend that's becoming increasingly important as they shape the consumer shopping experience. Read more in our blog post.
⚙️ About the AgenticCatalog beta
AgenticCatalog is currently in beta. This article explains the functionality available at the start of the beta period; it will evolve rapidly, and this documentation may not reflect all the latest features.
Beta limitations
At the start of the beta period, the following limitations apply:
We will pull your feed, convert it into an ACP-compliant feed format, and then push it to OpenAI. UCP-compliant feeds for other AI platforms and marketplaces will be available when AgenticCatalog is generally available later this year.
You can manually upload a sample feed file to test the enrichment feature for up to 100 products, but we will not push enrichments to OpenAI during the beta period.
Monitoring your agentic commerce performance is out of scope for beta, but will be included when AgenticCatalog is released for general availability.
👀 Not in the beta program? Contact your Account Manager to learn more.
Requirements
The following are required to use AgenticCatalog in the beta phase:
A US-based retail or e-commerce website.
Acceptance into OpenAI’s Merchant program.
A Google Merchant Center-compliant feed shared via:
S3 bucket (self-hosted)
GCS bucket (self-hosted)
URL
Your OpenAI credentials for us to push the converted feed.
An analytics source integrated with your Botify project.
🇺🇸 Only US-based websites can use AgenticCatalog feeds during the beta period, as ACP feeds are not yet available outside the US. Non-US sites can still benefit from AgenticCatalog to optimize their field for Google Shopping and to future-proof product data for AI search and agentic commerce.
How it works
Botify’s crawl data is merged with an existing merchant feed to generate ACP-compliant product feeds. AgenticCatalog automatically enriches and populates the fields that LLMs rely on, such as reviews, Q&A, and product highlights, using contextual data from PDPs. Pushing this structured product data ensures your products appear frequently and accurately when AI agents search for and recommend products. Ultimately, this will directly inform the purchases completed autonomously by AI agents.
The AgenticCatalog steps include:
You or Botify create a source to identify the product feed.
Botify pulls your GMC feed.
You optionally run the sample enrichment (which will not be pushed to OpenAI in beta).
You review the feed.
Botify converts the feed into an ACP-compliant feed and pushes it to OpenAI daily.
Creating an AgenticCatalog feed source
Access AgenticCatalog from the Welcome Page; or
Navigate to Activation > AgenticCatalog.
Click Create source.
Identify a name to identify the feed and the seller’s name.
Optionally fill the remaining fields to further define your feed, then click Create source. Botify will associate the feed you shared with the newly created source in AgenticCatalog.
Enriching the feed
After creating the feed source in AgenticCatalog, you can test the enrichment feature with a sample feed.
On the AgenticCatalog landing page, click the title for the feed source you created.
In the Sample Feed section, click into the field to select a local feed file, then click Upload sample.
After the file is uploaded, you can view the original feed, the OpenAI-compliant feed, and the GMC-compliant feed in AgenticCatalog:
To enrich up to 100 products from your Botify crawl data, in the Enrich Sample section:
Identify the number of products to enrich.
Select a sampling strategy: Either the first number of products in the feed, or every number of products identified in the Max products field. We recommend using the “First N products” strategy to test different product types.
Click Run enrichment. This process may take several minutes, depending on the feed size. You can navigate away from the page while the enrichment runs.
When enrichment is complete, you can view the enriched feed and the platform compliant feeds.
About product enrichment
We enrich product details with content discovered during Botify crawls to enhance product listings and boost visibility. This includes adding structured data, such as Q&As and product reviews, and new attributes based on the product page details. The Enrichment Details block explains the recommendation to help you understand the potential value of including the product details in your product feed.
Exploring feeds
To view feeds, including enriched product attributes:
Either click one of the enriched views:
or select a version from the version history:
Apply the “Enriched products” filter to view only enriched products in the feed.
Click into a row to see all attributes, including enriched attributes. The enriched attributes are identified in the left column with the “optimized label”, and the details are shown on the right in the Enrichment details box.
Click Download enrichment report to get a CSV file of all enrichments made to the feed.
Identifying errors in feeds
When viewing a feed:
Apply the “With errors” filter to find the issues to validate before we submit the feed to OpenAI.
Click a product row to view details. The validation errors are shown at the top of the page, including the attribute names and their expected values:
Exporting feeds
To export AgenticCatalog feeds from one of the original or enriched views, click Download feed to export the full feed:
To export only a subset of the feed, such as enriched products or those with errors, apply the corresponding filter and then click Download feed:
Alternatively, download a portion of the feed’s current version in your preferred format from the Source page.
About Feed Versions
For each compliant feed generated and pushed to OpenAI or Google, AgenticCatalog creates a new feed version. The feed tile shows the current version, the number of products included in the feed, and the date and time when it was pushed to the platform.
Click the tile to open the feed details and access the full version history:
👉 Feeds successfully ingested in Botify are labeled “normalized”.
After we push a compliant feed to a platform, it’s added to the Delivery History section. Click the button next to a version to access the compliant feed:
Delivery History Status
The delivery history can have any of the following status types:
Scheduled: When the version is scheduled to be pushed to the platform.
Delivered: When the version was successfully pushed to the platform.
Failed: When we attempted to push the version to the platform, but the platform ingestion was unsuccessful.
Deactivating a feed source
A feed source is activated when you create it. Deactivate a feed source if you no longer want to use it.
👀 There is no way to delete a feed source; deactivate the feed sources you no longer need.
To deactivate a feed source:
Click the source card on the AgenticCatalog landing page.
Deselect the Active box.
Click Save changes at the bottom of the page.
Getting help during beta
If you need help working in AgenticCatalog or encounter issues during the beta period, contact your Account Manager.
FAQs
Can I use AgenticCatalog if I don't have an existing feed?
Can I use AgenticCatalog if I don't have an existing feed?
No, AgenticCatalog requires a feed as an entry point.
Do I need a full website crawl to use AgenticCatalog?
Do I need a full website crawl to use AgenticCatalog?
No, a a full crawl of your PDP pages is sufficient.
Will I see enrichments in the URL Explorer when AgenticCatalog is in GA, so I can compare metrics on enriched pages vs. non-enriched pages?
Will I see enrichments in the URL Explorer when AgenticCatalog is in GA, so I can compare metrics on enriched pages vs. non-enriched pages?
This is not planned, but we are investigating it.
How do AgenticCatalog feeds work with LiveUpdate?
How do AgenticCatalog feeds work with LiveUpdate?
LiveUpdate surfaces product availability and pricing, not enriched product data. We are currently exploring how feeds enriched by AgenticCatalog will work with LiveUpdate.