I keep hearing that Google is changing the way they handle optimization and site ranking…and AEO and GEO are the future, which embraces AI. I know Elements has meta-tags and alternative tags, etc…available. Will Elements keep up with the new trend?
@dan @ben Feel free to edit this or update if there is better instruction or tuts available.
There are no secret tricks. The truth is if you are good and knowledgeable at what you do the playing field is more fair than ever. But the search giants and the AI-LLMs will be hard to win unless you establish your site as an authority. That means having on your website a lot of relevant data that is well researched, cited, and organized. The following is an AI generated summary from my personal notes on the subject.
Gemini Summary: Elements gives you the control and freedom to do the work, but it will not do the thinking for you. If you know exactly how to assemble those four pillars, you can build a site that is perfectly optimized for the AI era.
To optimize for AI search (AEO/GEO), a webmaster using Elements cannot just drop text on a page. You must manually manage semantic HTML hierarchy, inject proper metadata, manually connect a tight web of internal links, and write or generate the hidden JSON-LD schema code yourself.
The Four Pillars Explained
To successfully prepare an Elements site for modern AI and search engines, these four elements must be handled with intent:
1. Semantic Markup
Elements provides an inspector panel that lets you change generic containers (like <div>) into semantic tags like <article>, <section>, and <aside>. It also lets you explicitly assign heading ranks (<h1>, <h2>, <h3>).
- What you must do: You have to intentionally group your essays into these tags. AI scrapers read this layout to understand the difference between the actual article body, a sidebar, and a footer. Without your manual assignment, it’s just a flat wall of text to the machine.
2. Meta Data
Elements includes a dedicated inspector panel for browser titles, page descriptions, and Open Graph tags (the data used when a link is shared on social platforms).
- What you must do: You have to write unique, compelling descriptions for every single page. AI and search engines use this metadata as a quick summary to judge whether your page matches a searcher’s intent before they even crawl the rest of your text.
3. Logically Structured Internal Link Data
Because Elements creates flat, static pages rather than dynamically relating them in a single database grid, it doesn’t automatically understand how your articles relate to one another.
How Google and AI Methods Work Now
Modern search engines rely heavily on two core mechanics to determine what content gets served to users:
-
Website Intent vs. Searcher Intent: AI models use semantic search, which looks at the meaning behind words, not just literal keywords. When a user types a prompt, they have a specific intent (e.g., they want an exhaustive tutorial, a quick fact, or a deep philosophical essay). The AI reads your webpage to analyze your “Website Intent.” If your content layout is chaotic or fails to state its purpose early, the AI flags it as an “intent mismatch” and moves on.
-
E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness): Because AI models synthesize answers, they are terrified of quoting fake or low-quality information. They look for deep, interconnected “topic clusters” on a site to prove you are a genuine expert. They also look for hidden JSON-LD Schema code blocks—standardized, machine-readable notes embedded in a webpage that explicitly tell the AI: “This text is an Article, written by an expert Author, published on this exact Date.”
-
What you must do: You must manually link your essays together using descriptive anchor text (e.g., linking a chapter on Character to a supportive essay on Delegation). This builds “topic clusters.” When an AI crawler sees an orderly web of internal links, it recognizes that your site has deep, organized authority on the broader subject.
4. JSON-LD Schema
This is the hidden machine language that explicitly tells an AI search bot exactly what your content represents (e.g., that it is an independent article with a specific author and publish date). Elements does not generate this automatically for your text layouts.
-
What you must do: You must use an external tool to write or generate the JSON-LD script, copy it, and manually paste it into the custom header or code panels within Elements for that specific page. See the minimum requirements below assuming at least on image on a page:
-
<script type="application/ld+json"> { "@context": "https://schema.org", "@type": "BlogPosting", "mainEntityOfPage": { "@type": "WebPage", "@id": "https://yourwebsite.com/blog/your-post-url" }, "headline": "The Exact Title of Your Blog Post", "description": "A short, one-to-two sentence summary of what the article is about.", "image": "https://yourwebsite.com/images/featured-image.jpg", "datePublished": "2026-06-08T08:00:00+02:00", "author": { "@type": "Person", "name": "Your Name", "url": "https://yourwebsite.com/about" }, "publisher": { "@type": "Organization", "name": "Your Publication Name", "logo": { "@type": "ImageObject", "url": "https://yourwebsite.com/images/logo.png" } } } </script>
The Bottom Line
Elements is a professional design canvas. It won’t hold your hand or automate the data engineering in the background, but it leaves the doors wide open so that an educated webmaster can build a flawless, lightning-fast foundation for the future of search.
To see a practical demonstration of how the application layout handles site assets, global configurations, and template structures under the hood, this RapidWeaver Elements Project and App Overview walks through the workspace environment where you manage these specific custom code panels.
Thank you…I knew there’d be a ton to digest. It just seems that the SEO playing field is taking dramatic left and right turns via AI. I see good and bad in these changes…where as before websites relied on meta-info to tell search algorithms what was worth looking at and relaying…which was exactly what any particular site was conveying in the first place. I appreciate the info…I’ll get reading!
Thank you for sharing this insight!
Any name or links to these tools?
Does #4 need to be done for every page?
If yes - for me - that’s 100+ custom html components - a little much…
@dan Should this new computer code naming json be embedded into elements?
Just a thought - if this is the new way forward…
@Pegasus added a tool below, or as it falls in this thread, just above. ![]()
I guess that depends on the results you want. Is your content something you want in search results and AI referrals? If so the short answer is yes. If your website is a brochure site, the search engines aren’t going to use your info for anything other than “near me” for the most part.
What Are You Using Elements For?
A simple list of website types and how search and AI traffic works in 2026
Most search advice online is written for one kind of website. The author rarely says which kind. A forum member reads a guide about topic clusters or keyword maps, applies it to a site that needs none of it, and loses weeks of work for nothing.
So the first job of any webmaster is to know which kind of site you are building. That one choice decides whether search and AI traffic matter to you at all.
Ask this about your site.
Can a machine give your visitor what they came for without the visitor ever arriving?
If the answer is yes, your site is exposed. AI answers on the results page are taking your clicks. Pew Research found that when an AI Overview shows up, only about 8 in 100 searches lead to a click on any site. Without the AI Overview, that number is 15 in 100.
If the answer is no — if your value needs the person to show up, buy, book, hire, or belong — your site is safe. Search and AI are doorways to you. They cannot stand in for you.
Three buckets follow from that question.
Bucket 1 — The answer is the product
These sites hand over information that a machine can read and repeat:
- Quick-fact and definition pages
- Generic how-to articles and listicles
- Recipe pages with no brand or following behind them
- Commodity news and aggregation
What changed. The click is going away. Semrush reported that food and recipe sites have seen traffic drops of 50 to 70 percent as full recipes now show inside AI answers. A page that only states a fact gives that fact to the AI for free.
What still works for these sites:
- First-hand experience and original data a machine cannot copy from somewhere else
- Real depth on a narrow subject, so you become the source the AI quotes
- An email list and a brand, so people come back by choice and by name
The goal here has moved. Chasing the click is a losing game now. Earn the citation. Own the relationship.
Bucket 2 — The visit or transaction is the product
A machine can describe these sites. It cannot do them for the visitor:
- Local businesses: restaurants, salons, trades, clinics, shops with a front door
- Online stores and product pages
- Booking, appointment, and quote pages
- Web apps, tools, and software
- Membership sites, forums, and communities
Why they hold up. An AI can tell someone about your pizza place. It cannot seat them, cook for them, or take the order. Several 2026 studies put real estate and e-commerce among the safest site types for organic clicks, because Google still needs the product page and the live transaction to answer the search.
What drives traffic here:
- For local sites, your Google Business Profile is the largest lever you control. It carries roughly a third of local ranking weight, and your main business category is the top match signal. Reviews do double duty — they rank you and they sell you. A one-star rating increase has been tied to a 5 to 9 percent revenue gain, and restaurants at 4.4 stars and up see far higher weekly sales than those below 3 stars.
- For stores, clear product pages, honest reviews, accurate stock counts, and a clean path to checkout.
- For apps and communities, word of mouth and brand searches carry most of the load. Search is a side door, and that is fine.
Bucket 3 — The site is a calling card
Their audience already knows the name. People arrive by referral, a business card, a link in a bio, or by typing the address:
- Portfolios and brochure sites
- Freelancer, consultant, and studio sites
- Artist, musician, and author pages
- Church, club, and nonprofit sites
- Event and campaign one-pagers
These sites were never built to win “best [thing]” searches. They do not need that traffic to do their job. Their whole purpose is to greet someone who already heard about you and turn that visit into a call, a booking, or a follow.
What matters for them:
- Show up cleanly when someone searches your name
- Load fast on a phone
- Make the contact, the booking, or the next step plain and quick to find
- Keep your name, address, and details the same everywhere
Spending weeks on keyword research for a calling-card site is wasted effort. The work is to be clear, fast, and easy to act on.
The shared toolkit, for when you do want traffic
Some pieces help any site that wants to be found. These are the ones worth your time:
- Semantic structure. Mark your page so a machine can tell the article from the sidebar from the footer.
- Metadata. Write a plain, honest page title and description for every page.
- Schema (JSON-LD). The hidden code that tells an AI what a page is, who wrote it, and when. A Carnegie Mellon study placed schema among the top signals tied to whether an AI quotes a page.
- Internal links. Connect related pages with clear link text, so a machine sees an organized body of work.
- Speed on a phone. A slow page loses rank and loses people. Most search now happens on mobile.
- Author identity. A real name, a real bio, a real About page. AI leans on who stands behind the words.
It’s become a huge mindset shift in 2026
The click is no longer the prize. Across the major studies, AI Overviews have cut clicks to outside sites by somewhere between 38 and 61 percent on the searches where they appear. Two things gain value in its place:
- Being the source an AI names. Cited brands earn more of the clicks that remain.
- Owning your audience. An email list, a following, and people who search your name on purpose. That traffic cannot be taken from you on a results page.
Where Elements fits
Elements is a design canvas. It hands you direct control of every piece in the shared toolkit — the semantic tags, the meta panel, the code panel for your schema. It builds fast, clean, static pages, which is the exact technical floor that helps every site above.
It will not do the thinking for you. It will not pick your bucket, write your reviews, or set your strategy. It gives you sharp tools and a clean workbench. The plan is yours.
Summary: Start with defining your intent and need.
Looks complicated and time consuming …?
How to Be Found in 2026
Search clicks from Google are drying up. So for most site owners the real question is no longer “how do I rank?” It is “how do people find me at all?” The doors that still work in 2026 all reward the same thing — a real person with real experience being useful in public.
Which doors fit you depends on the kind of site you run, so sort that out first. Then work through the ones below, grouped by the job each one does.
Channels worth your time
Email — the one channel you own. This is the anchor under everything else. No algorithm sits between you and the reader, and no results page can take your audience away. Every other channel below should pour into this one. If you do only two things, make it one publishing channel and an email list.
Communities where your people already gather. Reddit has become a heavyweight. It now accounts for somewhere between 20 and 40 percent of all AI citations and ranks at the top across the major models. Niche forums, Facebook groups, and Discord servers do the same job at smaller scale. The rule here is strict. These communities and the AI systems both reward genuine help, and they push back hard on anything that reads like a sales pitch. One honest, useful answer outperforms fifty planted ones. This door is for people willing to show up as a real, helpful member.
Podcasts — your own, or as a guest. Audio builds a loyal audience that comes back. The spoken content gets transcribed and indexed, and guesting on an established show borrows someone else’s audience for an hour. A good fit for anyone who likes talking more than being on camera.
YouTube has become one of the most-cited sources in AI answers. Across the major AI tools, its share of citations sits around 20 percent, and it ranks at or near the top of the sources those systems trust. Google owns it, so it feeds Google’s AI answers directly. A video also carries a real face and a real voice, which is the trust signal that wins now.
Other people’s audiences. Guest posts, interviews, collaborations, and cross-promotion with a business that complements yours. This does two jobs at once. It borrows reach today, and being mentioned and quoted by other sites is one of the signals AI engines weigh when they decide whom to name.
Review and listing platforms. For local and product sites, this is the main way you get found at all — your Google Business Profile, Yelp, and the directories for your trade. Treat it as a real channel and keep it current.
Being quoted by the AI engines themselves. This is your own site doing the work. Write clear, specific answers that come from real experience, give them a named author, and keep the structure clean. Then when ChatGPT, Perplexity, or Google’s AI builds a reply, your page can land inside it. These engines favor real experience and specific detail. They pass over polished sales pages. It pays off even at lower volume — visitors who arrive from an AI answer have been measured converting far higher than ordinary search traffic, with one study putting it near 14 percent against under 3 percent.
The Common Thread
Every traffic funnel/channel rewards the same thing: a real person with real experience being useful in public. That is the author and trust signal at the center of search now. In 2026 it stopped being a box to check and became the whole game.
The 80/20
Pick one channel where you publish — video if you are comfortable on camera, audio if you like to talk, writing if that is your strength. Pair it with an email list you own, and build the list from that channel and your website. Once those two run on their own, add communities and collaborations on top. Anyone who tries to run all of them at once will do all of them badly.
That’s why marketing is a whole industry. It’s huge and it takes time and patience. The most recent data I could find real quick was for 2024. See below:
2024 advertising/marketing spend — major countries (USD)
| Country | 2024 estimated ad/marketing spend (USD) | Notes & citations |
|---|---|---|
| United States | $291 billion (total ad spend, ~digital $223B) | Estimates from US total ad market and digital ad spend reports. |
| China | $135 billion (digital-led ad market) | Estimated China digital & total ad market 2024. |
| United Kingdom | $38 billion (total ad spend; digital ~$27B) | UK advertising market size and IAB/PwC figures for digital. |
| Japan | $40 billion (total ad spend; digital ~$22B) | Japan ad market 2024 estimates (TV + digital split). |
| Germany | $24 billion (total ad spend; digital ~$15B) | Germany 2024 ad spend estimates. |
| India | $21 billion (advertising market; digital ~$7–9B) | India ad market 2024 with strong digital growth. |
| Canada | $8.5 billion (total ad spend; digital ~$6B) | Canada 2024 advertising spend estimates. |
| Australia | $6.8 billion (total ad spend; digital ~$5B) | Australia ad market 2024 figures. |
| France | $12.5 billion (total ad spend; digital ~$7–8B) | France advertising market 2024 estimates. |
| Brazil | $10.2 billion (total ad spend; digital ~$6B) | Brazil ad market 2024 estimates. |
Trying to automate it with Claude…
From Claude:
I don’t have direct access to your Elements CMS to push changes live — that requires your MCP connection, which would need to be active in this conversation for me to call content_update_item or similar commands.
What I can do right now:
If you tell me which blog post (or give me the slug/entry number), I can generate the fully populated JSON-LD block with your real data filled in — TrumpetStudio.com URLs, Michael Droste as author, Windy Town as publisher, correct dates, etc. — ready to paste into the appropriate field in Elements.
If Elements has a global header/footer injection field (or a custom code block in your page template), dropping it there once would apply it site-wide automatically, which is the smarter move anyway. That way every new post gets it without manually adding it to each entry.
Do you know if Elements v2 has a global code injection area in the site settings?
Reading the elements manual, watching videos, this forum, CMS, Google SEO … fixing and optimizing photos and writing texts, there is no time left to build websites ..
You seem a lot more informed than the average web designer…is this something you can assist with on a pro basis? Probably asking for a lot of people viewing this topic. I’d be interested to talk.
I cannot offer technical advice or help with Elements directly, Michael. I am an infopreneur. I understand the principles and I tried to make them as generic as possible for an Elements user to use. My history goes all the way back to the first version of RapidWeaver, second release. I was a white hat SEO guy back in the day. Nearly ALL of that is dead now.
SEO is dead, long live SEO. 
Dan let me know that Elements was not and is not going to be useful for my project types: long-form articles with a lot of internal linking and detailed navigation. So I have moved on to another platform. Do not ask, do not DM me; I will not say how or expand on that.
The RapidWeaver community helped me grow in my early days. I still feel the desire to give back. That is all this is. Honest truth that is simply our reality in 2026. For me, an online presence is actually more fun and easier than it has ever been.
Maybe. But let me be matter-of-fact. I can show someone the road, read the signs, point out many of the potholes, give marching instructions; however, I will not do the work. You can DM me and we can discuss some details, but I make no promises. I love what I do. I love helping others. But I have no patience for excuses or babysitting. I am respectful, as much as I can, but I do have my boundaries.
Yep, you do need to find balance. Look at the type of website, pick your path, apply the 80/20 rule. That’s it!
My take on this is that Elements is missing something: semi-auto creation of a JSON-LD block.
The whole “publisher” section of a JSON-LD is likely the same for an entire site, it needs to get defined Project Settings. The NewsArticle portion could come be shown in the Inspector for the page and filled in as necessary. The BlogPost possibility could be automatically generated from the CMS. DatePublished could come from the Publish push or the CMS.