No-code AI builders still fail the website test

The no-code AI category fails the test its own buyers apply to it. Jennifer Kerum, co-founder of Anchor Croatia, a luxury yacht charter brokerage, benchmarks AI builders against website builders — a transition from specialist skill to consumer product that happened roughly twenty years ago — and reports she still has to go through "five hundred loops" to build one intake platform. A serious build in this category currently needs two products. A website builder needs none.

What is the request, stated plainly?

Accessibility and easy integration, benchmarked against consumer software rather than developer tooling. Kerum's own framing:

"When you think about when you used to build a website twenty years ago, you needed to have a web developer, right? And now you have so many different programs that even if you didn't know anything about websites, you could just plug and play."

Her four stated requirements, given to Artificial Reality, the AI buyer-evidence podcast hosted by Galina Fendikevich: let a non-technical owner build a working AI application without hiring an engineer; integrate with an existing website and existing client systems; approach plug-and-play against the website-builder standard; and not require going through five hundred loops to reach a result.

Artificial Reality's additions: account for team training, which she names as a separate second step; produce something maintainable by the person who built it, since a workflow only the founder understands is a liability; and fail visibly rather than silently, because an intake agent quoting a wrong charter price does real commercial damage and this operator has already been burned by confidently wrong AI.

Why does removing the code not remove the barrier?

Because the concepts are what stop people, not the syntax. Website builders crossed over when the abstraction matched how a non-technical person already thinks: pages, blocks, images.

No-code AI builders still ask users to think in nodes, chains, prompts and connectors — a developer's mental model with the syntax removed. That is a different achievement from the one the comparison implies, and it is why an operator who can specify her product precisely still cannot ship it.

What is the tell that the category is not there yet?

That a single small-business build needs more than one vendor. Artificial Reality identified four horizontal no-code AI builders as apparently relevant to this request — none named or endorsed by the operator, none independently tested, and none vertical to charter:

• MindStudio — visual builder for AI agents and applications with access to 200+ models and 1,000+ integrations, deployable as standalone web apps or embeds on an existing site; $36M raised. Strongest on the site-embed path.
• Voiceflow — customer-facing AI agents across channels including web chat and voice; $35M raised, backed by Felicis and GV. Strong front end; the open question is what happens after the conversation ends.
• Relevance AI — custom AI agents and multi-agent workflows positioned around assembling an "AI workforce"; Series B, $38M, 2026. Intuitive framing, harder to debug without technical staff.
• Gumloop — agent and workflow builder for non-technical users with model choice, data connectors and scheduled automations; Series B, $50M, Benchmark, March 2026. Strong on internal process, less on what a customer sees.

An operator who wants a customer-facing intake experience with an internal workflow behind it is looking at MindStudio or Voiceflow plus Gumloop, plus whatever CRM holds the result. Three vendors to assemble one flow. Until an AI builder can carry a small business from idea to deployed application in one place, the plug-and-play comparison stays aspirational.

What does the demand signal look like from the seller's side?

Unusually clean, and it is not a marketing problem. Kerum is not a prospect resisting a pitch; she is a motivated buyer with a defined application who cannot complete a purchase. When a motivated buyer with money cannot complete a purchase, that is a product gap rather than a market gap.

The line worth pinning above a roadmap is her description of how AI capability is distributed:

"I feel like there's only a small portion of people that really understand how to use AI. And then I feel like there's everyone else that is just like: I'm hungry, what do I eat?"

That is a capability cliff, not a knowledge gap, and education does not close it. It is an interface problem, and the companies in this category are the ones nominally solving it. Operators in this position are not concentrated in yacht charter. They are in every industry Silicon Valley has decided is too small, too physical or too far away to build for.

AI tools named in this report

None. Jennifer Kerum named no no-code product. She described charter as an industry that "almost get[s] forgotten in the AI world" and described her own build as five hundred loops. MindStudio, Voiceflow, Relevance AI and Gumloop were identified by Artificial Reality as apparently relevant to her stated requirements. Inclusion is reporting, not endorsement.

So the roadmap question is not how to reach these buyers. It is this: can a non-technical owner get from idea to a deployed, customer-facing flow inside your product alone, without a second vendor and without a training project? Where the answer is no, that distance is the product, and the operator has already told you she will pay for it.

Where this comes from

S1E2: How a luxury yacht company uses AI without losing the personal touch — the full interview with Jennifer Kerum. Listen or watch: YouTube, Spotify or Apple Podcasts.