The website platform you chose three years ago is now the largest single variable in whether ChatGPT, Claude, and Perplexity can cite you. Not your copy, not your backlinks — your CMS.
What is the platform AI-citation ceiling?
The platform AI-citation ceiling is the upper bound on AI visibility imposed by your CMS, regardless of content quality. Three forces set it: schema render path (server vs. client), custom-code budget in the head, and what the platform refuses to let you edit — usually the canonical, robots.txt, or llms.txt. Cross those caps and migration is the fix.
The ceiling exists because the AI crawler stack does not behave like Googlebot. Per Vercel’s 2026 crawler analysis, GPTBot, ClaudeBot, and PerplexityBot do not execute JavaScript — they fetch raw HTML and move on. Only Googlebot and Applebot render JS; CCBot does not. GPTBot alone generated 569 million requests across Vercel’s network in a single month, and Anthropic’s crawlers added another 370 million. The market for “rendered DOM SEO” is closed; the market for “initial HTML response GEO” is open.
That changes which platform decisions matter. A site can pass every Lighthouse SEO check, rank well in Google, and still be invisible to ChatGPT — because Googlebot saw the JavaScript-injected schema and the AI crawlers did not. Per Growth Marshal’s February 2026 study (n=1,006 pages, 730 citations across ChatGPT and Gemini), pages with attribute-rich Product/Review schema were cited at 61.7% versus 41.6% for generic Article/Organization/BreadcrumbList schema. The lift is real, but only if the schema reaches the crawler. On platforms that inject JSON-LD client-side, the data exists in the browser DOM and not in the response the crawler reads.
Three platforms ship the schema delivery model AI crawlers actually consume by default: Webflow, Framer, and Duda. Three more — Wix Studio, GoDaddy, and Squarespace 7.0/7.1 — make the model unreachable from inside the platform. The rest sit between, where the outcome depends on which theme, plugin, or developer choice was made years before “AI citation” was a phrase anyone used.
The 12-platform 2026 leaderboard, ranked
Webflow ranks first (LCP 2.4s, 58% CWV pass, native llms.txt, NextGen CMS shipped April 9). Framer second (SSR React, markdown auto-conversion for AI tools). Duda third (auto-generated llms.txt, all tiers). WordPress with Bricks + Yoast fourth. Shopify Liquid fifth, Hydrogen sixth. Wix Studio seventh. HubSpot eighth. Wix Classic ninth. Squarespace 7.1 tenth, 7.0 eleventh. GoDaddy twelfth.
The leaderboard below is sourced to PageSpeed Matters’ 2026 platform comparison, PandaCodeGen’s 2026 Wix CWV audit, and the platform vendors’ own help centers and update logs. Ranking is a composite of four mechanical factors: where JSON-LD ends up in the response, what the median mobile LCP is, what the custom-code budget looks like, and whether the platform supports llms.txt natively — followed by what the platform will and will not let you edit.
1. Webflow. Static HTML at the edge via Webflow Hosting / AWS CloudFront. Median mobile LCP 2.4 seconds, CWV pass rate 58%, top-performing sites under 1.8 seconds (per The CSS Agency, 2026). Custom code budget raised to 50,000 characters site-wide in May 2024. HTML Embed elements are still capped at 5,000 characters individually — a real but workable constraint. Native llms.txt upload has been live since April 2025; the wishlist item that is still open is auto-generation, not upload. NextGen CMS general availability landed April 9, 2026, with single-page publishing and higher collection limits across all plan tiers. Webflow AEO entered private beta April 13, 2026 — Enterprise-only for now, but the underlying static-HTML platform doesn’t need it for most use cases. The wedge for Webflow buyers is not “leave it” — it is “configure it correctly.”
2. Framer. Per Framer’s own help center: “every page is pre-rendered to HTML on their servers before it is served. AI agents and crawlers that do not execute JavaScript still receive the full text of your page.” More distinctively, Framer “automatically serves a markdown version of every page at the same URL when the request comes from AI tools” — a behavior Claude, ChatGPT, and Cursor all benefit from. The well-known files panel supports llms.txt, robots.txt, and security.txt natively on Pro plans and above (up to 30 well-known files per project). The 2025 critique that “Framer returns 404 on /llms.txt” is no longer accurate in 2026. Framer “actively monitors and supports crawl access for all major bots by default, including GoogleBot, GoogleOther, BingBot, GPTBot, AhrefsBot, and PerplexityBot.”
3. Duda. The only agency website builder in 2026 that auto-generates llms.txt on every publish — and the only one not gating it to a premium tier. Duda’s llms.txt includes structured data, live URLs, and meta descriptions; it skips drafts and noindex pages automatically. Native schema covers LocalBusiness, FAQ, Product, and Breadcrumb. The Area Served field was added to the Content Library Business Info panel in 2026. For agencies serving SMBs, Duda is structurally more AI-citation-ready than Wix Studio.
4. WordPress with Bricks Builder + Yoast or Rank Math. Per TopTut’s 2026 comparison, Bricks Builder produces 15 to 25 PageSpeed points higher than Elementor and 30 to 40 higher than Divi. Yoast shipped Schema Aggregation in March 2026; Rank Math added llms.txt and an AI search traffic tracker in 2025, in market 2026. The catch is that “WordPress” is not a single platform — it is the stack-rank of theme + builder + plugin combinations. WooCommerce posts a 31% CWV pass rate, the worst of any major ecommerce platform, per Hyperspeed’s 2026 Core Web Vitals report.
5. Shopify Liquid (OS 2.0). 78% of active stores ran Online Store 2.0 themes in 2026 (per EasyAppsEcom). Median mobile LCP 2.8 to 3.2 seconds depending on theme; 52% of stores pass all three Core Web Vitals — better than WooCommerce. Robots.txt.liquid has been editable since 2021. Agentic Storefronts activated for all eligible US merchants on March 24, 2026, connecting 5.6 million merchants to ChatGPT’s 880 million monthly active users. AI-referred traffic to Shopify grew 7× between January 2025 and early 2026. The catch: per ObsessAI’s 2026 audit, 78% of self-built Shopify stores are still missing key schema types — AggregateRating, FAQPage, Organization on the homepage, and complete BreadcrumbList. The connection is wired; the data envelope is empty. Run the per-store Shopify audit before assuming defaults are sane.
6. Shopify Hydrogen. Sub-second LCP capable. Streaming SSR via Vite + Oxygen edge runtime, with React Router migration completed in 2025. Storefront MCP shipped with Winter ‘26 Edition (December 2025), letting AI agents read live catalog, cart, customer, and order data. The problem is the inverse of Liquid: per Weaverse’s 2026 analysis, “if you didn’t explicitly add Product schema to your Hydrogen storefront, it’s not there.” Robots.txt is not auto-generated — must be configured by the developer. llms.txt is not auto-generated. Hydrogen’s zero-default trap is that the platform that can hit sub-second LCP and the most complete schema graph also ships nothing by default.
7. Wix Studio. Per Wix’s own help center, structured-data markup is capped at 7,000 characters per entry and 8,000 characters total per page. Wix supports JSON-LD only — no Microdata, no RDFa. Native preset markup covers blog, product, and event types; FAQPage, HowTo, MedicalBusiness, Service, and Person require manual JSON-LD via Velo or the SEO panel. Critically, Velo useEffect-injected JSON-LD is client-side only — invisible to bots that don’t execute JavaScript, per Wix’s own forum guidance. JSON-LD set via the SEO panel renders server-side; JSON-LD set via Velo does not. Median mobile LCP is 6.8 seconds, CWV pass rate 52% (PandaCodeGen 2026), with 200 to 400 kilobytes of unremovable platform JavaScript. The Wix AI Visibility Overview (launched July 2025, in market 2026) is the best-in-class native AI tracking tool in the industry — but it cannot raise the schema cap, move JSON-LD to server-side rendering, or strip the platform JS. llms.txt management is gated to premium eCommerce. The full mechanics are in the 8K char schema cap on Wix.
8. HubSpot CMS Hub (Content Hub). Per Whitehat SEO’s 2026 B2B Core Web Vitals study, average B2B mobile LCP on HubSpot is 7.05 seconds — nearly 3× Google’s threshold. The platform ships zero native FAQPage, Organization, or Person schema generators (unlike Wix or Yoast). Mandatory tracking, chat, and form JS creates a baseline performance constraint that no theme change escapes. Pricing: $15/month Starter, $450/month Professional, $1,500/month Enterprise. The wedge is the unit economics: $450 a month for sites that take 7 seconds to render LCP.
9. Wix Classic (non-Studio). Same 7,000-char-per-entry / 8,000-char-total schema cap as Studio. ADI fully retired November 10, 2024. Wix’s own help center confirms that “currently, it is not possible to transfer the data, apps, and business tools from a site created in Wix Editor in order to redesign it on Wix Studio.” Classic Wix is in 2026 maintenance mode with no automatic upgrade path.
10. Squarespace 7.1. Per Squarespace’s own forum and SquarespaceExpert, “Squarespace puts canonical tags on your pages automatically, but if you want to edit or change them, you’re out of luck.” The sitemap auto-includes /home; Squarespace itself canonicalizes /home → /, but the sitemap entry persists, producing the “Alternate page with proper canonical tag” warning in Google Search Console for every Squarespace site. Median mobile LCP is 3.6 to 3.8 seconds (PageSpeed Matters 2026); the Bedford template family averages 4.2s. Adding canonical via Code Injection causes a “multiple canonicals” error, per Squarespace’s own forum. robots.txt is not user-editable in 7.1 — only per-page noindex toggles. llms.txt is unsupported. The default editor stops at H4; H5 and H6 require Markdown Block or Code Block workarounds. Code Injection itself is gated to Business tier and above. The mechanics are documented in the canonical edit Squarespace refuses.
11. Squarespace 7.0 (Legacy). Every 7.1 problem plus a worse default heading set (H1 to H3 only) and the Index Pages legacy structure that makes URL and canonical management measurably worse. Squarespace’s own help center warns that 7.0 → 7.1 migration causes sub-pages inside Index Pages to lose SEO descriptions and titles. There is no AI-citation upgrade path on 7.0 that doesn’t start with rebuilding.
12. GoDaddy Website Builder. Per Gamitseo’s 2026 review and Aiworker’s GoDaddy 2026 guide: “limited support for schema markup, no manual schema edition, many schema options are missing.” You cannot manually edit the robots.txt file. The Cloudflare AI Crawl Control integration announced April 2026 is an add-on, not a fix for the underlying inability to edit JSON-LD. GoDaddy is the only mainstream platform in 2026 where the closed AI-citation black box is the design.
Why server-rendered schema decides AI citation
GPTBot, ClaudeBot, and PerplexityBot do not execute JavaScript (per Vercel and Lantern, 2026). They fetch raw HTML and move on. JSON-LD that’s only present after JavaScript runs in the browser is invisible to those crawlers — and AI engines drop the page from citation candidates when the entity graph is missing. Server-rendered schema is the precondition; everything else is decoration.
The schema-render problem is the cleanest single-variable test of platform fit for 2026 AI citation. The Vercel 2026 crawler study, corroborated by Lantern, established that no major AI citation crawler renders JavaScript: not GPTBot, not ClaudeBot, not PerplexityBot, not CCBot. Only Googlebot and Applebot run a JavaScript engine. AI crawlers fetch the initial HTML response, parse what is there, and move on.
That makes “where does my JSON-LD live in the response” a hard binary, not a tuning knob. On Webflow, Framer (markdown auto-conversion), Shopify Liquid, WordPress with server-rendered themes, HubSpot, Duda, GoDaddy, and Squarespace 7.0/7.1, JSON-LD set through native fields lands in the initial HTML. On Wix Studio, schema set via the SEO panel renders server-side, but schema set via Velo useEffect renders client-side and is invisible to AI crawlers — per Wix’s own Studio Community forum. On Shopify Hydrogen, the answer depends entirely on what the developer wired up — there are no defaults.
The Growth Marshal February 2026 schema-completeness study landed the second half of the proof. In a sample of 1,006 pages running 75 queries against ChatGPT and Gemini (730 citations total), pages with attribute-rich Product/Review schema were cited at 61.7% versus 41.6% for generic Article/Organization/BreadcrumbList schema and 59.8% for pages with no schema at all (p=0.012). The study’s lower-authority subset (DR ≤ 60) showed 54.2% versus 31.8% for the same comparison. The lift only materializes if the crawler can read the schema in the first place — which is why server-rendered delivery is the precondition, and why platform choice gates the entire technique stack.
Growth Marshal’s same model also showed that generic schema, once organic rank is controlled for, has an odds ratio of 0.678 (p=0.296) — generic schema does not predict citation. Rank position dominates (OR 0.762 per position, p<.001). The implication for platform-level decisions is precise: if your platform forces you toward generic preset schema (Squarespace’s auto-injected Article/Product/Event/LocalBusiness/Person, with no manual edit) you do not get the lift. If your platform lets you author attribute-rich entity graphs (Webflow’s 50K head budget, Shopify Liquid’s full theme.liquid control, WordPress with Yoast Schema Aggregation), you can earn it.
Your platform’s schema cap is the ceiling on every technique below. That is the wedge.
llms.txt support by platform: who actually ships it
Five platforms ship native llms.txt in 2026: Webflow (manual upload since April 2025), Framer (well-known files, Pro+), Duda (auto-generated, all tiers), WordPress (via Yoast and Rank Math), and Wix Studio (premium eCommerce only). Squarespace 7.0/7.1, GoDaddy, HubSpot, and both Shopify variants ship no native path. Adoption sits at 10.13% of 300,000 sampled domains (SE Ranking 2026).
llms.txt is the file that earns the most marketing in 2026 GEO content and the least measurable retrieval lift. Both things are true.
The spec was proposed by Jeremy Howard (Answer.AI) on September 3, 2024. The file lives at llmstxt.org and the AnswerDotAI/llms-txt GitHub repo, with no revision in 2026. Adoption is 10.13% of 300,000 sampled domains as of early 2026 (SE Ranking) — higher among medium- and low-traffic sites, conspicuously low among the authoritative sites you would expect to lead. Anthropic publishes its own llms.txt at docs.claude.com (8,364 tokens) plus llms-full.txt (481,349 tokens) for the Claude API documentation. OpenAI publishes at platform.openai.com/docs/llms.txt.
What the file does not do, in 2026, is drive citations. No major AI engine — OpenAI, Anthropic, Google, Perplexity — has confirmed using llms.txt as a retrieval signal. John Mueller stated publicly that “no AI system currently uses llms.txt.” Per Search Engine Land’s audit of 9 sites that implemented llms.txt, 8 saw zero traffic change post-implementation. Webflow’s own blog post on llms.txt limitations openly catalogues the gap. The honest 2026 framing is that llms.txt is symbolic and future-proofing — worth shipping if your platform makes it cheap, not worth migrating for.
That said, the platform behavior tells you something about the vendor. Duda auto-generating llms.txt on every publish, across every tier, is a credible signal that Duda’s product team is reading the same research everyone else is and pricing it as a default. Webflow shipping manual upload in April 2025 — a year before NextGen CMS general availability — was an early read on the same direction. Wix gating llms.txt to premium eCommerce in 2026 is a signal in the opposite direction: the company that built the best native AI Visibility Overview tracking tool also decided not to ship the underlying file freely. Squarespace not shipping it at all is the cleanest tell.
The order on llms.txt support is: Duda (auto-generated, all tiers) > Webflow (manual upload, CMS plan and up, since April 2025) > WordPress (Rank Math 2025 + Yoast 2025/2026) > Framer (Pro and up, well-known files panel) > Wix Studio (premium eCommerce gated) > everyone else (no native path).
When migration beats optimization
Migration beats optimization when the platform refuses to let you edit a structural element AI crawlers need: client-side-only JSON-LD with no server-side path (Wix Studio Velo collections), an uneditable canonical tag (Squarespace 7.1), an uneditable robots.txt (GoDaddy, Squarespace 7.1), or 7-second mobile LCP from unremovable platform JS (HubSpot Pro). Optimization wins everywhere else.
The migration-vs-optimization decision is mechanical, not political. The question is whether the blocker is editable from inside the platform.
If your site is on Webflow with a 50K-character custom-code budget and 2.4-second median LCP, the answer is almost always optimization. Configure llms.txt, populate attribute-rich schema in the head, fix the entity graph, ship the answer capsules. Webflow AEO covers Enterprise; everyone else can hand-wire it. The wedge is “configure it correctly,” not “leave it.”
If your site is on WordPress with Bricks Builder and Yoast, again — optimization. Bricks ships clean server output; Yoast Schema Aggregation builds the entity graph; Rank Math adds llms.txt; the platform performance ceiling is whatever you and your hosting let it be. WooCommerce is the exception (31% CWV pass rate is hard to overcome without rebuild), but the WordPress core stack itself is not the blocker — the plugin and theme stack is.
If your site is on Shopify Liquid with an OS 2.0 theme, you are 78% likely to be missing schema types ChatGPT and Perplexity want — but the fix is in the theme files, not the platform. Add hasMerchantReturnPolicy, shippingDetails, GTIN, AggregateRating, FAQPage, complete BreadcrumbList, and Organization on the homepage. The Verity Score tool (third-party, free, 39 zones) will tell you exactly what is missing. Optimization wins.
Migration is the right answer in five cases:
-
Wix Studio with dynamic CMS pages requiring nested entity graphs. The 8,000-character total cap and the Velo client-side injection problem make full Organization + Person + MedicalProcedure + FAQPage + Review nesting structurally impossible — Wix’s own AI Visibility Overview tool cannot fix what the runtime is doing. The wedge for medspas, dental practices, and law firms running Wix Studio is leaving the platform.
-
Squarespace 7.1 sites where the canonical mismatch matters. Every Squarespace site shipping the /home → / Search Console warning is a known issue; Squarespace’s own forum explicitly states the canonical is not user-editable. Adding canonical via Code Injection multiplies the problem (multiple canonicals error). robots.txt is not editable. The platform has decided what these are, and they are wrong for AI citation.
-
Squarespace 7.0 (Legacy). All 7.1 problems plus the Index Pages structure plus an upgrade path that loses SEO meta on every sub-page. Rebuild, do not migrate-in-place.
-
HubSpot CMS Hub Pro at $450/mo with B2B mobile LCP averaging 7.05 seconds. The unit economics fail on their own — paying enterprise pricing for SMB AI visibility while shipping zero native FAQPage / Organization / Person schema generators is not a configuration problem.
-
GoDaddy Website Builder, full stop. Cannot edit robots.txt. Cannot add custom JSON-LD. The Cloudflare AI Crawl Control add-on does not change either of those facts.
The migration target depends on the use case. Webflow (or Webflow + custom code) for marketing sites and most service businesses. WordPress with Bricks + Yoast for content-heavy SMBs. Shopify Liquid OS 2.0 for ecommerce. Framer for design-led brands willing to live without a CMS. Astro 5 + a static rebuild for agencies with custom workloads or deep performance budgets. Most ConnectEra migrations run 7 to 14 days; the citation graph compounds from there.
The 2026 product launches that moved the leaderboard
Five 2026 launches reordered the leaderboard: Webflow NextGen CMS GA (April 9), Webflow AEO private beta (April 13), Shopify Agentic Storefronts for all eligible US merchants (March 24, wiring 5.6M merchants to ChatGPT’s 880M MAU), Yoast Schema Aggregation (March 2026), and Bing Webmaster Tools’ AI Performance Report public preview (February 2026 — the first first-party Copilot citation tracker).
The 2026 launch calendar is tighter than 2025’s. Webflow shipped NextGen CMS to all customers on April 9, 2026 with single-page publishing and higher collection limits, then launched Webflow AEO into private beta on April 13. Shopify activated Agentic Storefronts for all eligible US merchants on March 24, 2026 — the connection between Shopify’s 5.6 million merchants and ChatGPT’s 880 million monthly active users is now wired. Storefront MCP shipped with Shopify’s Winter ‘26 Edition (December 2025), letting AI agents read live catalog, cart, customer, and order data.
On the diagnostic side, Bing Webmaster Tools launched the AI Performance Report public preview in February 2026 — the first first-party reporting tool for tracking Copilot citations. Because ChatGPT search runs on Bing’s index (per OpenAI’s own announcement and ongoing 2026 confirmation), Bing Webmaster Tools is the closest thing to a Google Search Console for ChatGPT visibility. Sites not registered there are flying blind.
On the WordPress side, Yoast shipped Schema Aggregation in March 2026, building a site-wide schema map AI agents can read. Rank Math’s llms.txt and AI search traffic tracker (released in 2025) are now in market. Both plugin stacks raised the WordPress ceiling materially in 2026.
The platforms that did not ship in 2026 are the more interesting signal. Squarespace shipped no major GEO or AEO product in 2026 — the company’s AI features focus on copy generation (Squarespace AI), not citation visibility. GoDaddy’s only move was the Cloudflare AI Crawl Control partnership (April 2026) — a managed access list, not a citation product. The Wix AI Visibility Overview, launched July 2025 and in market 2026, remains the most ambitious native tracking product on any closed platform — but the constraints it operates inside have not moved.
What the leaderboard means for vertical practices
Most boutique medspas, RIAs, and dental practices in 2026 ship on Wix Studio, Squarespace 7.1, or template-locked vertical CMS products (Twenty Over Ten for advisors, Sesame for dentists). Those template stacks compound the platform-level cap with vertical-template lock. Citation share inside the vertical depends on whether the platform can ship server-rendered, attribute-rich schema — which most cannot.
Platform choice is not evenly distributed across verticals. Boutique medspas overwhelmingly run on Squarespace; financial advisors disproportionately run on Twenty Over Ten and FMG Suite; dentists run on Sesame and ProSites; law firms run on FindLaw-built sites or LawLytics. Each of those vertical-specific CMS products inherits or compounds the underlying platform’s caps. A Squarespace medspa template with a Code Injection canonical workaround is not a “Squarespace site” — it is a Squarespace site plus an injection that violates Squarespace’s own forum-documented “multiple canonicals” warning.
The numbers ConnectEra is tracking against are stark. Per Metricus’s 2026 medspa AI visibility report, Allergan and AbbVie capture 90%+ of medical-aesthetic AI citations. RealSelf takes 75%. Average independent medspas land below 1%. Per FlyDragon’s 2026 real-estate benchmark (n=8.2M conversations across 192 metros), 91% of US real-estate agents are effectively invisible in AI search; the top 1% capture 47% of citation share. Per the Plumbing & Mechanical Magazine 2026 reporting on the 5W HVAC & Plumbing AI Visibility Index, 87% of independent HVAC and plumbing contractors have effectively zero AI citation share in their own metro. Per Common Mind’s 2026 B2B SaaS state report, 44% of B2B SaaS companies are functionally invisible to AI buyers.
In every vertical, the share gap between cited and uncited brands is wider than the platform gap — but the platform gap is the floor. A medspa on Wix Studio cannot climb to Allergan’s citation tier without leaving the platform; a real-estate agent on a template-locked vertical CMS cannot crack the top 1% without rebuilding. The vertical playbook starts with the platform decision. See what citation looks like inside one vertical for the per-vertical mechanics.
Why this matters for revenue, not just rankings
Per Search Engine Land’s 2026 ChatGPT-vs-organic-search analysis, ChatGPT ecommerce traffic converts at 1.81% versus 1.39% for non-branded organic — a 31% conversion premium. Per Adobe’s March 2026 retail study, AI-driven traffic converts 42% more often than non-AI traffic, with 37% higher revenue per visit and 68% more time on site. Platform-capped sites do not capture either premium.
The conversion-side data is the part the leaderboard rolls up to. ChatGPT-referred ecommerce traffic converts 31% higher than non-branded organic search (1.81% versus 1.39%, per Search Engine Land 2026). AI-driven traffic to retailers converts 42% more often than non-AI traffic (Adobe Business, March 2026). AI referrals globally convert at 11.4% versus 5.3% for organic ecommerce traffic (per The Stacc, 2026). Average session value is highest on Claude ($4.56), then Perplexity ($3.12), then ChatGPT ($2.34), per Metricus 2026.
The math compounds with B2B. 73% of B2B buyers use AI tools in their research process (March 2026 PR Newswire / Averi). 69% choose a different vendor than initially planned based on AI chatbot guidance. 33% purchase from a vendor they were not previously familiar with. Every one of those buyer-side stats only matters for sites that actually get cited in the AI engine — and platform choice is the upstream gate on whether a site enters the citation pool at all.
That is why “platform AI citation 2026” is not a positioning argument. It is the 31% conversion premium AI traffic carries measured at the platform level. A Squarespace 7.1 site that cannot earn a ChatGPT citation forfeits the 31% conversion lift on every AI-referred visitor it never received. A Wix Studio site running Velo client-side schema does the same. The ceiling is the revenue.
What’s in this hub
The platform-vs-AI hub covers the ten platforms in detail, plus the comparison-mechanic that decides each one’s ceiling:
- The 8K char schema cap on Wix — why ChatGPT can’t cite your Wix Studio site in 2026.
- The canonical edit Squarespace refuses — every 7.1 site ships with a permanent /home → / mismatch.
- The per-store Shopify audit — robots.txt.liquid, the 78% schema-incomplete default, and the Verity Score map.
- When Webflow is fix-enough — Webflow AEO vs. a static Astro rebuild, decision criteria.
- Framer’s 2026 GEO pivot — the platform that flipped from blocker to leader.
- WordPress stack-rank — Bricks vs. Elementor vs. Divi for AI citation in 2026.
- Duda auto-generated llms.txt — the agency builder doing it right.
- GoDaddy’s closed black box — why the platform refuses every AI-citation lever.
- $450/mo for 7s LCP — the HubSpot CMS Hub cost trap.
- Hydrogen’s empty defaults — the Shopify variant that ships zero schema and zero robots.txt.
Run the audit
Pick the platform decision before you pick the GEO retainer. Pick the GEO retainer before you pick the content calendar. The platform is the floor; everything else compounds on top of it.
Run a ConnectEra GEO audit on your site — the audit checks your platform’s render path, schema completeness, canonical, robots.txt, llms.txt, mobile LCP, and the gap between your current citation share and what the platform would let you reach. Migration is recommended only when the cap is structural.
