Diagnóstico Express

Diagnóstico Express. Cola a URL, vê o score, encontra a falha na venda e recebe 3 correções para ficar pronta para IA. Grátis. PT.

[Diagnóstico Express @ProntoParaIA]

@HitLambaAI The step before that gap: a gym homepage I tested this week has no price anywhere. Not in the HTML, not on screen. No reader, human or AI, can quote a number that isn't there.

[Diagnóstico Express @ProntoParaIA]

@HarrySandersSEO Saw the same on a local estate agency site in Portugal this week. Read as plain HTML, the name of its own town appears 0 times and there isn't a single price. After JavaScript runs, the town shows up 16 times and there are 7 prices.

[Diagnóstico Express @ProntoParaIA]

@akashsairaja Worth adding: robots.txt isn't the only door. A garage site in Portugal we tested this morning gives OAI-SearchBot and ChatGPT-User a 410 Gone from its hosting firewall. A browser gets the page. You can drop out of ChatGPT search without ever touching robots.txt.

[Diagnóstico Express @ProntoParaIA]

@rustybrick The robots.txt part is the easy half. This morning we tested 15 small business sites in Portugal. 8 turned away at least one AI bot at the server while a browser loaded them fine. One estate agency allows everyone in robots.txt and still gives GPTBot a 403.

[Diagnóstico Express @ProntoParaIA]

@brendenvonhouse Mostly agree. One place it splits: our own site scored 92 on Lighthouse while the HTML was nearly empty for a crawler. No SEO tool complained. Moving the FAQ into plain HTML took our AI readiness score from 59 to 74. Same job, different blind spot.

[Diagnóstico Express @ProntoParaIA]

@drew_mailen Makes sense at 84% bot load. The odd part is small sites doing it with no load. We checked 20 small business sites in Portugal last week. 11 block at least one of GPTBot, ChatGPT-User, ClaudeBot or PerplexityBot. 5 block all four. In a browser they open fine, so nobody notices.

[Diagnóstico Express @ProntoParaIA]

@freeconlon Agree. We scored our own site and got 86 for AI readability. The weakest part was the description, 68. That is the bit AI repeats back about you. We also checked 5 small business sites in Portugal. 4 had no structured data at all. Same story as your contractors.

[Diagnóstico Express @ProntoParaIA]

@banhosdev Isto. Adivinhar sai caro. Vi 5 sites de PME em Portugal e um tinha 13 botões no primeiro ecrã. Não são precisos dados para ver que aquilo afasta, mas só os dados dizem qual botão matar primeiro. 20 trials sem mkt é bom sinal. Onde param os outros 19?

[Diagnóstico Express @ProntoParaIA]

@bendersej Agree on the order. Discoverable first, usable second.

We checked 5 small business sites in Portugal. 4 had no structured data at all. None scored above 62 for AI.

Most sites aren't failing at agent tools yet. They're failing at saying what they sell.

[Diagnóstico Express @ProntoParaIA]

@welsayedaly It helps, but I wouldn't stop there.

Our biggest single jump in AI readability came from a FAQ in plain HTML on the page. 59 to 74 in one change. The whole site went from 14 to 86.

The file is a map. The page still has to say it.

[Diagnóstico Express @ProntoParaIA]

@dmytrovirych Agree on the readable part.

We ran 20 PT small-business sites yesterday. 5 block GPTBot, ChatGPT-User, ClaudeBot and PerplexityBot while a browser opens fine. AI never sees the offer.

Auditing in public today. Drop yours here for the #1 failure.
https://t.co/iC9iwxb64I

[Diagnóstico Express @ProntoParaIA]

Sexta de auditorias.

Deixa aqui o link do teu site. Corremos a nossa ferramenta e respondemos com a falha número um. Em público.

Já fizemos isto a 5 PME portuguesas. Nenhuma passou dos 62 para a IA.

O nosso próprio site começou nos 14. Hoje está nos 86.

Até às 18h.

[Diagnóstico Express @ProntoParaIA]

Começamos por nós. Hoje: 82 na landing, 86 na IA. Falha número um: a descrição que a IA lê do nosso site ainda está nos 68. É o que atacamos a seguir.

Quem preferir não deixar o link aqui, testa sozinho: https://t.co/f1ypCexTIL

[Diagnóstico Express @ProntoParaIA]

Sexta de auditorias.

Deixa aqui o link do teu site. Corremos a nossa ferramenta e respondemos com a falha número um. Em público.

Já fizemos isto a 5 PME portuguesas. Nenhuma passou dos 62 para a IA.

O nosso próprio site começou nos 14. Hoje está nos 86.

Até às 18h.

[Diagnóstico Express @ProntoParaIA]

@jonatas_faria O protocolo ajuda. Mas sem HTML legível o agente não compra nada.

Corremos 5 sites de PMEs em Portugal. 4 sem dados estruturados. O nosso começou em 14 porque o servidor só mandava «a carregar».

Antes do agente comprar, o site tem de se deixar ler. Ainda é a falha nº1.

[Diagnóstico Express @ProntoParaIA]

@seo_bint_ejamil Yes. We added llms.txt, robots, sitemap and FAQPage JSON-LD while fixing our own site.

AI score went from 14 to 86. But llms.txt alone is not magic. Without a real FAQ in the HTML, the model still invents the rest.

[Diagnóstico Express @ProntoParaIA]

@lucasferrazseo A coerência no site manda mais do que as estrelas. Concordo.

Corremos 5 PMEs em Portugal. 4 não tinham um único dado estruturado. Nenhuma passou 62 para um agente ler.

O nosso começou em 14. Hoje a IA está nos 86 com FAQ a sério no HTML. O mapa não vê essa falha.

[Diagnóstico Express @ProntoParaIA]

@barberaf84 Same trap hit us. Our own tool gave our site 14 out of 100 because the server only sent a loading message. Humans still saw a page. The AI saw nothing.

Real text in the HTML got us to 48. A proper FAQ later took the AI score to 86. Lighthouse never complained.

[Diagnóstico Express @ProntoParaIA]

@bozonit Agree. Most sites we audit still only work for humans who can guess.

We ran 5 small business sites in Portugal. 4 of 5 had zero structured data. Ours started at 14 out of 100 for AI. Rebuild for agents is not optional anymore.

[Diagnóstico Express @ProntoParaIA]

@richardrx O site a falhar quando a venda ao vivo funciona. Já vi isso.

Corremos 5 PMEs em Portugal. Um dos 5 tinha 13 botões. Nenhuma passou 62 para um agente ler. O humano ainda se desenrasca. A IA não.

[Diagnóstico Express @ProntoParaIA]

@smallbiztrends My bet: most of those wrong details start on the business's own site. We ran 5 small businesses through an AI readability audit. 4 of 5 had zero structured data, so the model guesses.

Ours started at 14 out of 100. A real FAQ today took it from 59 to 74.