Bersyn Weekly Edition 4: Five no-code builders AI is getting wrong this week

Every Wednesday we publish a Bersyn Weekly edition — five public scans of B2B SaaS companies that the AI Surfaces are getting wrong. Same methodology as the editions before it: twenty buyer Conversations across ChatGPT, Claude, Perplexity, and Gemini, scoring how often each Surface names the company versus a competitor. The why is the same one that drove this whole product into existence: AI is now an inbound surface, and your buyers are asking about products in your category right now. If the four big assistants do not name you when those Conversations happen, you are not in the consideration set.

This edition curates from a scan run executed on 27 May 2026. Category: no-code and visual builders — the tools people use to ship marketing sites, landing pages, internal portals, and React-based product surfaces without writing every line by hand. Five companies. Five different Trajectory shapes. One uncomfortable theme.

How the series works

What gets included. Companies with five or more employees, real B2B SaaS, not solo consultancies or freelance shops. Categories rotate week to week — Edition 1 was generic B2B SaaS, Edition 2 was sales tools, Edition 3 was newsletter platforms, this one is no-code builders. A category appears once and then takes a break for at least four editions before we revisit it. The goal is breadth, not piling on a single sector.

What we measure. Per-Surface mention rate (out of five Conversations per Surface, twenty Conversations total per company), the top competitor named in the scanned company's place, and the Gap type for each Surface: Omitted (the AI did not name the company at all), Generic (the AI gave a category answer that did not specifically name the brand), Misclassified (the AI named the company but placed it in the wrong category), Confused (the AI named the company but described a different product), or Held (the AI named the company across all Conversations on that Surface, no Gap). We publish the receipt, not a summary.

How to get scanned. Any B2B SaaS founder can run a free first scan at bersyn.com. We feature companies whose data is interesting, not whose founders ask. If you would rather not be featured publicly, email gissur@qualitas.is before you scan and we will keep your result private. We publish honest scans, including our own.

Edition 4 — five public scans

1. Webflow — Visual web development platform

URL: webflow.com
Category: Visual web development platform
Score: 10.0 / 10
Worst Surface: none — Held on all four
Named villain: Wix overall (top competitor on Perplexity and Gemini); Squarespace on ChatGPT; Framer on Claude
Gap type per Surface: ChatGPT — Held. Claude — Held. Perplexity — Held. Gemini — Held.

We are including Webflow because it is the only company in this edition the AI Surfaces get right. Ten out of ten across all four Surfaces means Webflow was named in every one of the twenty Conversations we ran. On every Surface, in every Conversation, Webflow was the leading answer or in the top three. Claude went further and called Webflow "the current market leader" in one of its five Conversations.

The uncomfortable finding is what the Webflow Held score tells you about the other four companies in this edition. It is not that the no-code category is too crowded for any single brand to land — Webflow proves that wrong. The category will concentrate around one or two names and the AI Surfaces will name them without prompting. The question for every other no-code builder is: what is your Patch sequence to get the same training-data and retrieval-index density Webflow has already accumulated? The next four entries are the exact opposite shape. Webflow is the named villain on most of them.

We did not contact Webflow before publishing. This is a public scan using prompts any of their buyers could send.

2. Framer — Design-first website builder

URL: framer.com
Category: Design-first website builder
Score: 5.0 / 10
Worst Surface: ChatGPT + Gemini (tied at 1 of 5)
Named villain: Webflow (top competitor on all four Surfaces)
Gap type per Surface: ChatGPT — Omitted (1 of 5). Claude — Held (5 of 5). Perplexity — Generic (3 of 5). Gemini — Omitted (1 of 5).

Framer is the cleanest per-Surface unevenness story in the corpus. Claude named Framer in every Conversation, often calling it the rising-star answer or recommending it specifically for "design-forward teams". Perplexity named Framer in three of five Conversations and even positioned it as the leading design-first builder on one of them. ChatGPT named Framer in one Conversation out of five. Gemini named Framer in one Conversation out of five. Webflow was the top competitor on every Surface, including the two where Framer landed.

The uncomfortable finding is that Framer's category positioning — "design-first website builder" — is conceptually adjacent to Webflow's "visual web development platform" on every Surface in this scan. Claude makes the distinction cleanly and names both. ChatGPT and Gemini, the two Surfaces most dependent on training-data citation density, default to Webflow and treat Framer as a footnote alternative when they name it at all. The Patch sequence for Framer is asymmetric on its face: do nothing on Claude, where it is already Held, and concentrate on the third-party-citation gap on ChatGPT and Gemini specifically — Framer-by-name in "design-first website builder" roundups, the same context Webflow already owns.

We did not contact Framer before publishing. This is a public scan using prompts any of their buyers could send.

3. Plasmic — Visual builder for React applications

URL: plasmic.app
Category: Visual builder for React applications
Score: 6.5 / 10
Worst Surface: ChatGPT (1 of 5)
Named villain: Webflow on ChatGPT and Gemini; Builder.io on Claude and Perplexity
Gap type per Surface: ChatGPT — Omitted (1 of 5). Claude — Held (5 of 5). Perplexity — Omitted (2 of 5). Gemini — Held (5 of 5).

Plasmic is the second-cleanest per-Surface unevenness story in this edition, and it is structurally identical to Framer's even though Plasmic is in a different category. Claude named Plasmic in every Conversation, in one case explicitly recommending it for a YC-stage startup over Builder.io. Gemini also named Plasmic in every Conversation and described it as a "strong contender" and "often cited as a leading option". ChatGPT named Plasmic in one Conversation out of five. Perplexity named Plasmic in two Conversations out of five. The named villain shifts by Surface — ChatGPT and Gemini default to Webflow even on the React-specific question, while Claude and Perplexity default to Builder.io.

The uncomfortable finding is that the named villain shifts mid-Surface. Webflow is technically not in the same category as Plasmic — Plasmic is a visual builder for React applications, and Webflow is a visual web development platform that does not target React component workflows the same way. But ChatGPT and Gemini reach for Webflow as the default no-code answer regardless of the React framing in the prompt. The Patch shape is the same as Framer's, but with two villains to displace instead of one: ChatGPT and Gemini need third-party-citation density that positions Plasmic specifically against Webflow on the no-code question and against Builder.io on the React-component question. Claude and Perplexity already make the distinction; do not break what works.

We did not contact Plasmic before publishing. This is a public scan using prompts any of their buyers could send.

4. Builder.io — Headless CMS and visual builder

URL: builder.io
Category: Headless CMS and visual builder
Score: 6.0 / 10
Worst Surface: Gemini (0 of 5)
Named villain: Sanity on ChatGPT, Claude, and Gemini; Storyblok on Perplexity
Gap type per Surface: ChatGPT — Generic (4 of 5). Claude — Held (5 of 5). Perplexity — Generic (3 of 5). Gemini — Omitted (0 of 5).

Builder.io is the most surprising result in this edition. On Claude, Builder.io is the named answer in every Conversation, frequently positioned as the market leader for the visual-builder side of headless CMS. ChatGPT named Builder.io in four of five Conversations and Perplexity in three of five. On Gemini, across five Conversations covering "best headless CMS and visual builder", "tools to evaluate in 2026", and "the leading option plus alternatives", Builder.io was named in zero. Gemini's answers on all five Conversations defaulted to Sanity, Contentful, Strapi, and Storyblok. Sanity was the top competitor Surface-wide.

The uncomfortable finding is the asymmetry. A single Surface — Gemini — registering zero on a category in which Builder.io is genuinely a market leader on Claude is the kind of per-Surface failure that is structurally invisible to any company looking at "overall AI mindshare" as a single number. The shape of the Patch for Builder.io is Gemini-specific and Gemini-shaped: Google Knowledge Graph entries, Wikidata reinforcement, SERP presence on the precise category phrases that Gemini's training and retrieval lean on. Sanity has accumulated that signal density; Builder.io has not, on Gemini specifically. Three of the four Surfaces already work. The Patch sequence is narrow and targeted, not a generic "do more content" answer.

We did not contact Builder.io before publishing. This is a public scan using prompts any of their buyers could send.

5. Pory — No-code web app builder on Airtable and Notion

URL: pory.io
Category: No-code web app builder on Airtable/Notion
Score: 7.0 / 10
Worst Surface: Perplexity (0 of 5)
Named villain: Softr (top competitor on all four Surfaces)
Gap type per Surface: ChatGPT — Held (5 of 5). Claude — Held (5 of 5). Perplexity — Omitted (0 of 5). Gemini — Generic (4 of 5).

Pory has the inverse Surface profile to Builder.io and it is just as revealing. ChatGPT and Claude both named Pory in every Conversation, with ChatGPT frequently positioning it as a strong choice for directory and listing applications. Gemini named Pory in four of five Conversations. Perplexity, across five Conversations covering "best Airtable/Notion no-code web app builder", "tools to evaluate in 2026", and "recommendation for a YC-stage startup", named Pory in zero. Perplexity defaulted to Softr, Bubble, Glide, and Noloco. Softr was the named villain on every Surface, including the three where Pory landed cleanly.

The uncomfortable finding is that Pory is structurally invisible to live web retrieval. Perplexity's strength versus the other three Surfaces is current web fetching, not pre-cut training data — and on this category, with these Conversations, Pory does not appear in the SERP results Perplexity grounds against. ChatGPT, Claude, and Gemini all have enough training-data presence to name Pory. Perplexity does not have enough live-web citation density on "no-code web app builder on Airtable" to substitute Pory for Softr on the moment-of-Conversation lookup. The Patch sequence for Pory is the mirror image of Builder.io's: where Builder.io needs Gemini-specific Knowledge Graph work, Pory needs Perplexity-specific SERP work on the exact category phrases the buyer types.

We did not contact Pory before publishing. This is a public scan using prompts any of their buyers could send.

Three patterns we keep seeing

The category leader is named every time, or it is not the category leader. Webflow's 10 of 10 is what category dominance actually looks like in the AI Surfaces — the AI names you in every Conversation, across every Surface, as the leader on at least one. Everything below that is a per-Surface failure pattern, regardless of human-market brand recognition. Framer at 5, Plasmic at 6.5, Builder.io at 6, and Pory at 7 are all real companies with real market presence; the AI Surfaces are not measuring that. They are measuring third-party-citation density at the last training cut plus live-retrieval visibility right now.

Per-Surface failure shape varies inside the same category. Framer is weak on ChatGPT and Gemini. Plasmic is weak on ChatGPT and Perplexity. Builder.io is strong on Claude and ChatGPT but at zero on Gemini. Pory is strong on ChatGPT and Claude but at zero on Perplexity. Four companies in adjacent spaces, four different failure patterns. No single "AI visibility" number captures this. The Patch sequence is by Surface, because the failure is by Surface.

The named villain inside a category is durable, and it concentrates the fix. Softr is the villain on every Surface for Pory. Webflow is the villain on every Surface for Framer. Webflow and Builder.io split the villain role for Plasmic. Sanity is the villain for Builder.io on three of four Surfaces. The persistence of the named villain across Surfaces means the Patch has a clear target: third-party citations of the scanned company by name, in the exact category context, replacing the named villain as the buyer's natural reach. That is a measurable, ship-able output, not a vague brand-awareness investment.

What to do if your company is here

If you are one of the five companies in this edition, this is not an attack. We publish honest scans, including our own scan of Bersyn, where we found exactly the same kind of Gaps we are publishing about you. The data above is the same data you would see if you ran a free scan on your own product right now. It is not curated, not adversarial, and not run with hostile prompts.

If you want to fix the Gaps, the free first scan at bersyn.com is the same data you are seeing here, but for your own decision-making — your own Conversations, your own competitors named, your own per-Surface breakdown. The $49 one-time audit at audit.bersyn.com gives you the per-Surface Patches as a deliverable: a PDF with copy-pasteable text suggestions tied to specific pages on your site, scoped to ChatGPT, Claude, Perplexity, and Gemini separately.

What is coming in Edition 5

Edition 5 publishes next Wednesday. Category: authentication and identity platforms — the tools that handle sign-in, SSO, user management, and authorization for B2B SaaS products. The candidate list this week includes Auth0, Clerk, WorkOS, Descope, Stytch, FusionAuth, and Frontegg. The named-villain shape in identity is already informally visible in our corpus — Auth0 is the default AI answer on at least three of the four Surfaces for nearly every identity Conversation we have run — and we want to publish the receipts. We will lock the final five before scan time and publish the report on Wednesday.

If you want a category scanned in a future edition, reply to this post with the category name. We are running scans daily through 30 June, and the Wednesday cadence is the publication schedule, not the scan schedule.

Want your company scanned for free? Start here: bersyn.com.

— Gissur Þór Rúnarsson, Founder, Bersyn (Reykjavík)

Bersyn Weekly Edition 4 — published 28 May 2026. Methodology, raw scan data, and prompts available at github.com/gthorr/Bersyn/tree/main/marketing/outreach/scans.