Content Inventories with AI

AI isn't magic, but it can help you be more effective at exploring, framing, deciding about, and taking action on your content

Spreadsheets are the currently most prevalent way that people interact with inventories. They give us a sense of direct control, and every content strategist knows how to use a spreadsheet.

Charts and reports are a way of distilling analysis for a wider audience and to see the forest for the trees. But using them does require more expertise.

AI provides a way to ask content analysis questions in a more natural way while still allowing the spreadsheet, chart, and reports views.

Chat with your inventory

What you can expect from the Beta:

  1. You will receive an access code and a link to a page to get started.
  2. You will enter a URL on that page.
  3. A crawl starts, for up to the first thousand URLs of your website
  4. As the crawl proceeds, Chimera will give some basic information about the site
  5. Once the crawl ends, you will be able to ask questions via chat about the crawl.

We plan on sending out Beta invites sometime in the summer.

Types of questions you can ask in chat*

  1. Single Answer questions. Example: "How many URLs are in this analysis?"
  2. Tables. Example: "Create a table with the columns URL, name, and links in. Each row is a URL. Show the top ten URLs by links_in."
  3. Charts. Examples: "Show the distribution of sections of the site, but count of content."

* Largely implemented already, and which we are honing for the Beta. Based on feedback from the Beta, we will be expanding this.

Add information to your inventory from LLM Prompts

To be effective in inventorying, results from prompts must be:

  • Normalized (for instance limiting the answer to just one or two words for a content type)
  • Repeatable
  • Based on your needs (for instance, you can upload a hiearchical taxonomy to tag your content to)
  • Efficient (Content Chimera asks an LLM all at once for multiple fields for better efficiency)

Still the same Content Chimera under the hood

Over the past seven years Chimera has evolved to a very powerful content analysis data warehouse. Notably, since Chimera already stores information in a structured manner, this can be leveraged with AI features. For instance, when asking questions about the data like the above, Chimera already knows a lot about the data since these questions aren't asked entirely "blind" (for instance, Chimera knows the most popular values per field, so that can be leveraged).

More AI is coming

We have already implemented key functionality like categorization of content via LLM, and we have already had initial success at Chimera automatically writing Chimera Expressions with plain English prompts.