TL;DR: Bing Webmaster Tools has a new AI Performance report. You can use its Grounding Queries to get cited more in all AI systems.
People are spending a lot of money on paid tools to see how their sites are getting surfaced in LLMs.
Bing Webmaster Tools (the Bing equivalent of Google Search Console) just introduced this for free.
They have a new tab called AI Performance – still in beta.
Most actionable is it shows you grouped keywords AI is using to surface your content.
For those that don’t know:
When you ask an LLM a question where the user would want up-to-date information, this often triggers a web search.
Rule of thumb – if your prompt depends on “what’s true right now”, the web search is triggered. If your prompt is about “how things work” or “what something means”, there’s no web search.
Product/service recommendations, for example, typically use web searches.
You can even see how ChatGPT breaks your prompt into web searches which it then synthesizes.
Grounding Queries in Bing Webmaster Tools - what they are
Connect your site to Bing Webmaster Tools.
Go to AI Performance. Scroll down to Grounding Queries.
These show grouped representations of the searches LLMs are doing to surface and cite your brand.
Something you may notice is some of the language AI is using to search is very different from what you intended with your content.
Or things are phrased very differently from how you phrased them.
Or you just didn’t cover this grouped representation of keywords thoroughly in your content.
You likely don’t need to create new pages – more pages aren’t always better.
How to use Bing Grounding Query data to get more LLM traffic
Start by searching the grounding query on Bing. If you’re not in positions 1 – 3, check for:
- Are there any individual words in the grounding query that are not used high on your page or on your page at all?
- Are you covering the grounding query in depth on your page?
If you answer no to either of these, fix it. It can be very, very easy to take a word or two from that grounding query and use it in a natural way higher on your page. Or to create a section for that grounding query.
You may see that the grounding query is very different from your page’s subject matter.
Only then, if you can’t naturally fit it into your page, it may be better to create a new page for it.
In that case, use the grounding query, written in natural language, in your:
- Page Title
- URL slug
- H1
- Beginning of your page’s first sentence
- Meta description
Again, only do this if the grounding query is too different from the page you have ranking for it in Bing.
If you’re targeting the grounding query properly, but still not in positions 1 – 3 of Bing, that means you’re lacking authority:
- Take some pages of yours that are ranking in positions 1 – 5 of Bing, internal link to the page you want to boost. 2 – 3 internal links will be enough. Don’t use your grounding query for the anchor text of all internal links – instead use various natural language phrasing for it. If the grounding query is in natural language, use it in one of the internal link anchor texts.
- Build backlinks to the specific page or to your site as a whole if it’s a new site. Here’s the anchor text ratio to use for external links.
How this helps you across all LLMs
Bing’s new AI Performance report shows data for:
- Microsoft Copilot
- AI-generated summaries in Bing (the equivalent of Google’s AI Overviews)
- “Select partner AI integrations”
The way that I look at Bing data is as a small representation of Google data. The biggest LLMs are doing their searches on Google.
But:
- Bing results are often (not always, but often) close to Google’s.
- Microsoft’s LLMs are closely modeled off of the others, which are honestly all close to each other; overall outputs are often broadly similar.
- If you’re getting surfaced a lot in one major LLM, good chance you’re getting surfaced well in the others.
So by using Bing’s AI Performance to identify and fix weak spots in your Microsoft AI coverage, you’re really fixing weak spots in all your AI coverage.
Get even more citations
My purchase-intent SEO method, Compact Keywords, works phenomenally for getting surfaced in AI.
Compact Keywords isn’t about creating informational blog posts that answer questions with no customer-intent.
It’s about creating conversion-based landing pages targeting scenarios where searchers are looking right now to use something, purchase something, or make a discovery call.
The searchers know what they’re looking for, they’re just looking for a brand to give that to them.
And by explaining on the page why you are that brand that they should trust… by explaining how your brand is the best choice for their high-intent scenario… you also explain to LLMs why they should recommend you when your pages show up for the searches LLMs are doing.
Compact Keywords are not about making best X for Y listicles which are getting destroyed right now:
Compact Keywords are about nabbing even higher-intent keywords, which are often under-targeted, like:
- voice note recorder for dentists
- handmade artisan ceramic mugs
- emergency lawn drain repair boston
These are 400-500 word pages explaining why you are a great choice for this thing the searcher is looking to achieve right now.
Compact Keywords shows you:
- How to find these keywords
- How to structure your site for these keywords
- How to make and rank these conversion-based landing pages
- How to audit your site to maximize its surface-potential
- How to do link building so you build SEO authority while also using language that signals to LLMs you’re an authority in your niche
- Lots more
There’s never been a better time to save years learning how to do SEO that gets customers.
Get the course now at https://edwardsturm.com/compact-keywords/
