10 min read
GEO (generative engine optimisation) makes your site citable by AI assistants (ChatGPT, Perplexity, and Google’s AI Overview). When someone asks one of them “who in [your area] does [what you do]?”, GEO decides whether your site is in the answer.
SEO is the long-standing practice of making your site rank in a normal Google search.
The Xponent approach runs both from the same WordPress install:
llms.txt file, AI-visibility monitoring, and faster indexing.They are configured once as a pair and then run quietly. This guide walks through that configuration step by step, either so you can do it yourself, or so you understand exactly what we do when you ask us to.
Honest expectation-setting: GEO makes you a candidate for AI answers; it does not guarantee a ranking, and AI mentions build slowly. Treat
llms.txtas cheap, sensible insurance, not a magic switch. We never promise a ChatGPT ranking.
You need:
Loose on the tools? That is exactly where Xponent can stand in for you. See the “Have us do it” boxes throughout: each one is the honest point, not a hard sell.
| You can do this | Xponent can do it for you |
|---|---|
| Follow this guide on your own site | Request help from the team: a specialist configures it and verifies it works |
Rank Math must be the single source of your site’s SEO output. If both Rank Math and Squirrly try to write titles, sitemaps, and schema at once, they conflict and neither works cleanly.
In Rank Math → Titles & Meta, set a title format like %title% | Your Brand once. Squirrly is no longer overriding titles, so this will now actually apply.
Open your live homepage and, in your browser’s view source, you should see one page title and one description. If there are two of each, stop here and flag it: that is the clash we are about to resolve on the Squirrly side. (See the Troubleshooting section at the end.)
Have us do it: If confirming modules and output feels uncertain, this is a 20-minute job for our team and it is the single most important part to get right.
This is the step people miss, and it is the whole trick: Squirrly must write NO SEO output of its own. It hands everything to Rank Math and keeps only the GEO features (the llms.txt file, AI-visible monitoring, indexing).
The “No Configuration” option is not in any of Squirrly’s normal menus after first setup. It lives on a hidden onboarding screen. In your browser, go to:
yourdomain.com/wp-admin/admin.php?page=sq_onboarding&tab=step1
Login if prompted, and you should see the Squirrly onboarding screen.
On that screen, click “No SEO Configuration” (sometimes labelled None or No SEO config). It saves immediately. This single switch tells Squirrly to stop writing metadata, sitemaps, and social tags; Rank Math owns all of that now.
Important side effect: switching this way resets Squirrly’s local settings, and it turns off the two GEO features you actually want (
llms.txtand indexing). That is expected. You re-enable them in the very next step; do not skip it or GEO silently stays off.
yourdomain.com/llms.txt).llms.txt file lists #The llms.txt file is your site’s “table of contents” for AI. Keep it short and list your best pages first:
/wp-admin/, /wp-login.php, /wp-json/, /xmlrpc.php, and (on a store) cart, checkout, and account pages.Take a moment to review the list Squirrly generates and trim it to what an AI should read.
Have us do it: The “No Configuration” switch is the point most self-managers stumble, because it is hidden and it resets settings. If you do not see the onboarding screen, or re-enabling the GEO layer looks off, hand it to us: it is exactly the kind of thing that is far cheaper to get right once than to debug later.
WordPress caches pages, and a cached page will keep showing the old SEO tags even after you change the settings. This makes setup look like it failed when it actually worked.
After the cache is purged, confirm all five checks pass. If you are comfortable in a terminal (or a browser tool), you can check the live site directly:
SITE="https://yourdomain.com"
# 1. Sitemap is owned by Rank Math
curl -s "$SITE/sitemap_index.xml" | head -3 # look for "Rank Math"
Pass criteria (all should be true):
1. The sitemap is generated by Rank Math (not “Squirrly”).
2. Your homepage has exactly one meta description.
3. Your homepage has exactly one canonical link.
4. Your tag has a single prefix attribute (a duplicate means both plugins are still writing).
5. llms.txt returns a 200 and lists your chosen pages.
Not a terminal person? Open
yourdomain.com/llms.txtin a browser (a 200 / a readable list on the screen means it is working), and use an online page-source viewer to count titles/descriptions. Or simply ask us to run the check.Have us do it: Verification is the other spot that looks simple and trips people up (cache masking, duplicate tags from a third plugin). We run this as a standard final check on every setup.
GEO setup is the same for everyone; the next step depends on what kind of business you run.
If in doubt, run this paragraph by the Xponent team; we’ll tell you which depth fits before you spend time on the wrong one.
llms.txt stays live and is crisp: that is the baseline; it means AI can systematically read your site.No magic, no promises of a ChatGPT ranking: you are now a candidate AI can quote, and you can measure whether the mentions grow. That visibility is the whole point.
| Practice | Why |
|---|---|
| One SEO output owner | Rank Math writes titles, sitemaps, and schema; Squirrly writes none of it. No conflicts. |
| GEO as a layer on top | Squirrly adds llms.txt, AI visibility, and faster indexing without touching your SEO output. |
| Verified, not just configured | Every setup is checked against five pass criteria before it counts. |
| Honest about AI | You become citable and we measure it; we never promise a ranking. |
| You can, or we can | Every hard step has a clean “ask the team” path; no lock-in, no shame. |
| Symptom | Likely cause | Fix |
|---|---|---|
| Two titles / two descriptions on your homepage | Squirrly still writing SEO output | Re-run Part 2 (No Configuration) and make sure the sitemap/social toggles are off; purge cache |
| Sitemap still says “Squirrly” | Squirrly’s sitemap toggle is still on | In Squirrly → Tweaks & Sitemaps, set XML Sitemap OFF; do not try to fix it with a redirect |
llms.txt returns 404 |
GEO layer not re-enabled after No Configuration | Re-run Part 2.3 (LLMs File ON) and purge cache |
| Duplicate description only on the homepage | Elementor’s own site settings | In Elementor → Site Settings → Settings → Meta Description, clear the field and let Rank Math own it |
| Squirrly features turn back on after an update | Squirrly updates can re-enable output | After any Squirrly update, re-check Part 2 settings and purge cache |
If you hit a snag and the table does not cover it, that is a good time to ask the team. Most self-managers who get stuck land on one of the obscure ones above, and having an expert clear it in minutes beats an afternoon of trial and error.
