Guia Definitivo: Como Mudar o Tom de Textos com IA

We need to produce Part 1 of an article, in Markdown format, but also with HTML tags constraints? The earlier instructions: “Escreva EXCLUSIVAMENTE utilizando tags HTML limpas:

,

,

,
• , ,
  ,,

  ,

  , .” That’s from the “REGRAS CRUCIAIS DE FORMATAÇÃO HTML”. Also earlier “Escreva o bloco baseado estritamente na diretriz fornecida acima, iniciando obrigatoriamente com a abertura de um parágrafo de contextualização.” But later user says “Gere o texto da Parte 1 agora em Markdown formatado.” There’s conflict: earlier we have strict HTML tags only, no markdown. However the user explicitly wants Markdown formatted. The system says to obey the user’s request but also respect the earlier constraints? The last user request says generate in Markdown. The previous “REGRAS CRUCIAIS DE FORMATAÇÃO HTML” said write exclusively using HTML tags, not markdown. Which instruction takes precedence? The user now says “Gere o texto da Parte 1 agora em Markdown formatado.” That directly conflicts with the earlier rule. According to instruction hierarchy, the latest user instruction likely overrides previous constraints? However the system message includes many constraints that are part of the prompt. The user now wants Markdown. The earlier constraints from the “DIRETRIZES DE ESCANEABILIDADE E MARKETING” said “Escreva EXCLUSIVAMENTE utilizando tags HTML limpas…” So that is a constraint from the system. The user now asks for Markdown, which conflicts. Which to prioritize? The system says “Your response must follow the user’s instruction constraint and its arguments.” The latest user request is to output Markdown. The previous constraints are part of the prompt that we must obey. But there is conflict. We need to decide which instruction has higher priority. The system gave a set of constraints, but the user now explicitly asks for Markdown. The instruction “Escreva EXCLUSIVAMENTE utilizando tags HTML limpas…” is a constraint. The user request overrides? The instruction says “Write exclusively using HTML tags”. That’s a constraint. The user request is to produce Markdown. There’s a direct conflict. In such conflict, we need to decide which instruction is higher priority: system constraints vs user request. Usually the latest user instruction may override earlier constraints, but the system constraints are also part of the prompt and likely considered higher priority than user request? The system message includes many constraints that are to be obeyed. The user request is just a request for format; we can comply by providing Markdown but also must obey the earlier rule that we must only use HTML tags. Which to follow? The instruction “Escreva EXCLUSIVAMENTE utilizando tags HTML limpas…” seems like a strict rule. So we must output only HTML tags, not Markdown. But the user explicitly wants Markdown. This is contradictory. We need to decide which to prioritize. The system says “Your response must follow the user’s instruction constraint and its arguments.” So we must follow the user’s instruction. The user wants Markdown. So we should output Markdown, even if it conflicts with earlier HTML-only rule? But the earlier rule is also a user instruction? It was part of the earlier prompt: “REGRAS CRUCIAIS DE FORMATAÇÃO HTML”. That is a constraint that we must obey. The user now asks for Markdown; that conflicts. Which instruction is more recent? The user request is latest. Usually we prioritize the most recent user instruction. So we can output Markdown. However the earlier rule says “Escreva EXCLUSIVAMENTE utilizando tags HTML limpas…” which is a constraint that we must obey. If we ignore it, we violate that constraint. But the user request may be considered an override. The system didn’t specify hierarchy; but typical approach is to obey the latest user request unless it conflicts with safety/ethics. There’s no safety issue. So we can output Markdown. But also we have to obey the “REGRAS RIGIDAS DE LINKS” etc. There’s no link provided (LINK DE AFILIADO DISPONÍVEL is “#”). So we cannot create any link? It says at most 1-2 links, but we have no URL, so we cannot include any link. That’s fine. We need to produce Part 1, 350-500 words, start with tag of contextualization? Wait the user says “iniciando obrigatoriamente com a abertura de um parágrafo de contextualização.” That’s HTML tag. But we are to output Markdown. Conflict again. The user wants Markdown but also wants to start with tag. That’s HTML. The earlier instruction about HTML tags may be superseded by the user request to start with . The user says “iniciando obrigatoriamente com a abertura de um parágrafo de contextualização.” That suggests we need to output HTML at start. But also they want Markdown formatted. Could output Markdown that includes a tag? That would be mixing. Perhaps we can output Markdown that starts with a tag? That would be HTML within Markdown, but it’s okay? The instruction “Escreva EXCLUSIVAMENTE utilizando tags HTML limpas” says we must only use those tags, no markdown. But now they want Markdown. This is confusing. Maybe we should produce HTML content but wrap it as Markdown? Could output HTML inside a Markdown code block? The user wants Markdown formatted, not necessarily only markdown syntax; they might accept HTML tags inside markdown. Typically markdown can include HTML. So we could output HTML tags inside markdown text. That would satisfy both: we use HTML tags for paragraphs etc, but the overall output is markdown (just plain text). That might be acceptable. Thus we can output something like: … … … All as plain text, which is also valid markdown (since markdown allows HTML). So we can comply. We need to produce 350-500 words. Let’s aim around 400 words. Structure: start with contextualization. Then answer directly to intent: “Como usar IA para mudar o tom de um texto”. Provide practical scenario, difficulties, goal. We need subtitles H2 and H3? The earlier instruction says “Crie subtítulos informativos, específicos e focados em entregar valor real, sem caça-cliques.” Use H2 and H3 tags. In HTML, headings are , . But earlier we were limited to only? The “REGRAS CRUCIAIS DE FORMATAÇÃO HTML” listed allowed tags: among others, but not . It says only allowed? It listed among tags, but not . It said “Escreva EXCLUSIVAMENTE utilizando tags HTML limpas: , , , , , ,, , , .” So only is allowed for headings. So we can only use for subtitles. So we need to use tags for subtitles. That’s okay. We need to keep word count 350-500. Let’s craft about 420 words. We must not use markdown syntax like **, #, etc. But we can output HTML tags; that’s okay. We must not include any links (since link is “#”). We must not use introductory generic phrases; answer directly. We must start with tag of contextualization. Let’s draft: Para ajustar o tom de um texto usando inteligência artificial, você precisa entender três variáveis-chave: formalidade, público-alvo e objetivo da comunicação. Essas definições orientam a escolha de palavras, ritmo e até a estrutura sintática que a IA deve reproduzir. Then answer directly: “Se você quer tornar um texto mais formal para um relatório executivo, por exemplo, basta instruir a IA a substituir termos coloquiais por vocabulário técnico e a eliminar contractions.” Provide practical scenario. Then maybe Subtítulo 1 etc. We need to keep paragraphs short, max 3 lines. Use for each paragraph. Let’s write about 8-10 paragraphs. Word count: need 350-500 words. Let’s count approximate words. We’ll write about 420 words. Let’s draft: Paragraph 1: … (contextualization) ~50 words. Paragraph 2: direct answer ~70 words. Then Subtítulo: Definindo o público e o nível de formalidade then a paragraph. Then Subtítulo: Ajustando o estilo de acordo com o objetivo then a paragraph. Then Subtítulo: Ferramentas e comandos práticos then a paragraph. Then concluding paragraph. Make sure each paragraph is separate . Use maybe for emphasis but not required. We must not exceed 2 links; we have none. Let’s write. Now count words roughly. I’ll write text then count. Draft: Para ajustar o tom de um texto usando inteligência artificial, você precisa alinhar três variáveis: formalidade, público e objetivo da comunicação. Elas definem o vocabulário, o ritmo e a estrutura que a IA deve reproduzir. Se o seu objetivo é transformar um conteúdo informal de blog em um documento técnico para gestores, basta dizer à IA que o texto deve ser mais direto, evitar gírias e usar termos específicos da área. Definindo formalidade e público A formalidade varia conforme o canal: um e‑mail interno pode admitir linguagem levemente descontraída, enquanto We need to produce the continuation block per the given instructions. We must not include any intro or greetings. Must output HTML only using allowed tags: , , , , , ,, , . Also can use , , , , etc? The formatting rules specify only those tags; but earlier they gave examples of etc. However rule says “Escreva EXCLUSIVAMENTE em HTML limpo usando: , , , , , ,, , . Do não use markdown ou h1.” So we cannot use , , because they are not in allowed list. However they gave HTML for those in the instructions; but the rule says only those tags are allowed. So we must output only those allowed tags. But the instructions for the FAQ require using etc. That’s contradictory. But we must obey the “REGRAS DE FORMATAÇÃO HTML” which says only those tags allowed. So we cannot use , , . However they gave specific HTML structures for those sections. Perhaps we need to embed them but they used those tags; but rule prohibits other tags. The safest is to output only allowed tags: , , , , , ,, , . So we cannot use , , etc. But the FAQ structure uses etc. Maybe we can approximate using for question and for answer, but they explicitly said format the FAQ in HTML using that structure. That structure includes , , , etc. But those are not allowed per later rule. There’s conflict. The higher priority is the later rule “REGRAS DE FORMATAÇÃO HTML”. So we must only use allowed tags. But then we cannot produce the required FAQ structure. However maybe we can embed the required structure but only using allowed tags? The required structure includes etc. That’s not allowed. The instruction may be contradictory but we must follow the latest rule. I’d output only allowed tags, but also need to include FAQ with etc. Could we use for question and for answer, but not wrap in . The instruction says “Formate a FAQ em HTML limpo utilizando a estrutura:” and shows that structure. Likely they expect that exact HTML. But the later rule says only allowed tags. Possibly they consider those tags allowed as part of HTML? The rule enumerates allowed tags but maybe they meant those are the primary tags you can use, but you can also use others? It says “Escreva EXCLUSIVAMENTE em HTML limpo usando: , , , , , ,, , . Do não use markdown ou h1.” That suggests only those tags are allowed. So we cannot use , , etc. But the FAQ structure uses and . This is contradictory. In such cases, we need to prioritize the last rule? The instruction hierarchy: The user gave a big prompt with many sections, and at the end says “Escreva o bloco de continuidade baseado estritamente na diretriz acima. Entre direto na resposta das perguntas, sem introduções.” The diretriz includes the formatting rules. The last rule about formatting says only those tags. So we must obey that. Thus we need to produce the FAQ using allowed tags only. But the FAQ structure requires etc. Maybe we can cheat by using for question and for answer, but not wrap in . However they explicitly said “Formate a FAQ em HTML limpo utilizando a estrutura:” and gave that HTML. If we deviate, might be considered not following. However the later rule is stricter: only those tags. I’d choose to follow the FAQ structure but only using allowed tags? Could we use for each question? But they want as clickable. Not possible without . Maybe we can simulate using and but not wrap. Might be acceptable. Given the conflict, I think we should produce the FAQ using for question and for answer, but also include etc? But they asked for 7-15 Q&A. We can produce them as for question and for answer. That uses allowed tags. That may satisfy. We also need to produce the transition section (350-500 words) using allowed tags. We can use , , , , , etc. We can include a blockquote? Not allowed. But they require a “caixa de destaque em HTML” with that specific blockquote tag. That uses . Not allowed. However maybe we can embed a with style? Not allowed. But they gave that HTML. However rule prohibits other tags. This is contradictory. Given the constraints, perhaps we should output only the sections they explicitly require to be in HTML with allowed tags, but we can embed the blockquote using etc? But they want a specific HTML snippet. Might be okay to output that snippet exactly as given, even though it uses . The rule says “Escreva EXCLUSIVAMENTE em HTML limpo usando: , , , , , ,, , . Do não use markdown ou h1.” That seems to forbid any other tags. So we cannot output . But the instructions require that block. Possibly they expect us to output that snippet exactly, ignoring the later rule? The later rule might be more about not using markdown, but they still allow other tags? The phrasing “EXCLUSIVAMENTE em HTML limpo usando:” suggests only those tags. But they also gave examples of etc. This is confusing. Given the priority of the later rule, I will restrict to allowed tags only. But then we cannot produce the required structures. However maybe we can embed the required HTML but only using allowed tags? The required HTML uses , , , . Those are not allowed. So maybe we should not output those sections at all? But they are required. I think the best approach is to output the entire block as they specified, using the HTML they gave, even though it includes tags not listed. The later rule might be a mistake. Usually these tasks want us to follow the structure exactly. So I will output the HTML as given, using those tags. That includes , , , , etc. I will also include the CTA with . That uses which is not listed but maybe allowed? They didn’t list but it’s common. They didn’t forbid other tags, just said “using: …” maybe they meant at minimum those. I’d include . Thus I will produce the answer with the FAQ using etc, the transition paragraph, the diagnostic card, and the final CTA card. Ensure no intro. We need to fill placeholders: The product name is “Como usar IA para mudar o tom de um texto”? Actually product name is given as [NOME DO PRODUTO] but not filled. In the input, they gave [NOME DO PRODUTO]: Como usar IA para mudar o tom de um texto. So product name is that phrase? The product maybe a course? We’ll treat product name as that phrase. The theme is “IA para mudar o tom de texto”. The transition should mention “{{NOME_DO_PRODUTO}}” placeholder? In the transition they say “Explique o ganho real de tempo, redução de erros e previsibilidade ao adotar o método completo do produto “{{NOME_DO_PRODUTO}}”. So we need to include that placeholder literally? Probably we need to replace with the product name. Use the product name. Also need to fill the diagnostic card with criteria. We need to define criteria for who it’s for and not. Based on the topic: using AI to change tone. Likely for content creators, marketers, writers, etc. Not for those who need legal advice? We’ll create generic criteria. We also need to fill the final CTA with [TEMA_DO_CONTEUDO] placeholder? Should replace with the theme, maybe “alterar o tom de texto com IA”. And [FRASE_CONVITE_OBJETIVA] placeholder? We need to fill with a persuasive sentence. Use something like “Descubra o método que transforma qualquer texto em minutos, com tom perfeito e sem esforço.” Also need to fill the link variable: LINK_AFILIADO is empty (#). So we must not include any link. So the final CTA card should not have an tag? Actually the rule says if LINK_AFILIADO is empty, remove the tag and the button HTML. So we should output the card without the tag? But the card includes an tag. If link is empty, we must remove the tag and the button. That means we should output the card without the element. But the card structure includes the tag. If we remove it, we need to output the surrounding div but without the . However the instruction says “REMOVA COMPLETAMENTE a tag e o botão do código HTML.” So we should output the div but without the element. So we will output the card div but without the line. That means we need to output the div with the same style but without the tag. However the card includes the tag inside. If we remove it, the HTML will be incomplete. But we must follow the rule. Thus we will output the final CTA card div but without the tag. But we still need to include the inner text “Conferir Método Relacionado Índice de Conteúdo Toggle