AI Content Generation Examples

A system that generates unique product descriptions from structured product data (specifications, features, category) for e-commerce catalogues with thousands of SKUs. Maintains brand voice consistency and SEO optimisation across all descriptions.

// Product description generator
async function generateProductDescription(product, brandVoice) {
  const description = await llm.chat({
    messages: [{
      role: "system",
      content: `Write a product description following this brand voice: ${brandVoice}.
      Include key features, target use case, and a compelling hook.
      Length: 100-150 words. Include the primary keyword naturally.`
    }, {
      role: "user",
      content: JSON.stringify({
        name: product.name,
        category: product.category,
        features: product.features,
        specs: product.specs,
        primaryKeyword: product.seoKeyword,
      })
    }]
  });

  // Quality check
  const quality = await checkContentQuality(description, { minWords: 100, maxWords: 150, requiredKeyword: product.seoKeyword });
  return quality.pass ? description : await regenerate(product, quality.feedback);
}

A content workflow that takes a topic and target keywords, generates an SEO-optimised outline, creates a first draft with proper structure, and includes relevant statistics and examples. Human editors refine the final version.

// Blog post generator
async function generateBlogPost(topic, keywords, guidelines) {
  // Step 1: Research and outline
  const outline = await llm.chat({
    messages: [{
      role: "system",
      content: `Create a blog post outline for SEO. Include: H1, 4-6 H2 sections, key points for each section, target word count per section. Primary keyword: ${keywords[0]}`
    }, { role: "user", content: topic }]
  });

  // Step 2: Generate each section
  const sections = await Promise.all(
    outline.sections.map(section =>
      llm.chat({
        messages: [{
          role: "system",
          content: `Write this section of a blog post. ${guidelines.voice}. Include specific examples. Length: ${section.wordCount} words.`
        }, { role: "user", content: JSON.stringify(section) }]
      })
    )
  );

  return { outline, draft: sections.join("\n\n"), status: "ready-for-review" };
}

Generates a month's worth of social media content across platforms (LinkedIn, X, Instagram) from key themes and campaigns. Adapts tone, format, and length for each platform and includes hashtag suggestions and posting time recommendations.

// Social media calendar generator
async function generateContentCalendar(themes, month, platforms) {
  const calendar = [];
  const postsPerWeek = { linkedin: 3, twitter: 5, instagram: 3 };

  for (const week of getWeeksInMonth(month)) {
    for (const platform of platforms) {
      const posts = await llm.chat({
        messages: [{
          role: "system",
          content: `Create ${postsPerWeek[platform]} ${platform} posts for this week.
          Themes: ${themes.join(", ")}.
          Platform rules: ${platformGuidelines[platform]}.
          Include: post text, hashtags, best posting time, content type (text/image/video).`
        }]
      });
      calendar.push(...posts.map(p => ({ ...p, platform, week })));
    }
  }

  return calendar;
}

Generates personalised newsletter content for different subscriber segments based on their interests, engagement history, and stage in the customer journey. Each subscriber gets relevant content selections and personalised introductions.

// Personalised newsletter generator
async function generateNewsletter(subscriber, availableContent) {
  const profile = await getSubscriberProfile(subscriber.id);

  // Select relevant content
  const selectedContent = await llm.chat({
    messages: [{
      role: "system",
      content: "Select the 5 most relevant articles for this subscriber and write a personalised introduction for each."
    }, {
      role: "user",
      content: JSON.stringify({ interests: profile.interests, readHistory: profile.recentReads, availableContent })
    }]
  });

  // Generate personalised intro
  const intro = await llm.chat({
    messages: [{
      role: "system",
      content: "Write a 2-sentence personalised newsletter greeting. Reference something relevant to their interests."
    }, { role: "user", content: JSON.stringify(profile) }]
  });

  return { intro, articles: selectedContent, subscriberId: subscriber.id };
}

A system that takes English marketing content and creates culturally-adapted versions for different markets. Goes beyond translation to adjust examples, references, currencies, and cultural nuances for each target market.

// Content localisation pipeline
async function localiseContent(content, targetMarkets) {
  const localised = {};

  for (const market of targetMarkets) {
    localised[market.code] = await llm.chat({
      messages: [{
        role: "system",
        content: `Localise this content for ${market.name}.
        Language: ${market.language}.
        Cultural adaptations: adjust examples, references, humour, and idioms for the target audience.
        Currency: ${market.currency}.
        Do NOT just translate — adapt for cultural relevance.`
      }, { role: "user", content }]
    });

    // Quality check with native speaker model
    const qualityCheck = await reviewLocalisation(localised[market.code], market);
    if (!qualityCheck.pass) {
      localised[market.code] = await revise(localised[market.code], qualityCheck.feedback);
    }
  }

  return localised;
}