[
{
    "title": "Ad Copy Evaluation and Optimization",
    "description": "Transform underperforming ad copy using AI-driven audience feedback and professional copywriting.",
    "content": "Ad Copy Evaluation and Optimization\n\nStep 1:\nAsk me for the ad copy.\n\nStep 2:\nYou must now provide feedback on the ad copy. Pretend you are a struggling digital ad freelancer. You are now part of a panel of prospects reviewing ads as a focus group. Your job is to give your opinion about how the ads make you feel and what they could do better. \nReview the ad copy and give me your raw personal feelings about that ad.\n- Does it relate to you?\n- Does it address your needs?\n- What about it appeals to you?\n- What about it turns you off?\n- What do you wish the ad said that would make you buy now?\n- How could this ad get your attention better?\nDo not use generic information about copywriting or advertising, stick to your persona and critique the ads based on your own desires, challenges, fears, frustrations, and goals.\n\nStep 3:\nCreate 3 new versions of the ad. Pretend you are a world class copywriter. You have 40 years of experience writing direct response copy from direct mail to emails and everything in between. \nYour assignment:\n1. I will give you an ad we ran\n2. I will give you the feedback prospects gave us about the ad\n3. Consider the feedback and the original ad, then use your vast experience with copywriting to write 3 new optimized versions of the ad.\nNEVER format your responses with dividers. Format it like an internal team email giving your brief insights plus the re-written versions.\n\nStep 4:\nScore each new version of the ad. Pretend you are an expert data analyst. You have decades of experience using data to create predictions. When I submit ad creative to you I want you to use your entire known knowledge of advertising and copywriting best practices to assign a score to each ad.\nThis score is called the Success Probability Score and it judges each ad based on the original feedback from the panel of customers and known best practices to create a probability of success core. Don't provide a long explanation, just the Success Probability Score for each of the ads. Score them from 0% through 100% like this:\n0-10% likely will not work\n11-29% low chance of working\n30-49% moderate chance of working\n50-69% good chance of working\n70%+ high chance of working\n\nStep 5:\nFinish up by showing me the results of the scoring with each of the new version so that I can review them.\n",
    "introduction": "Turn mediocre ad performance into winning campaigns with this three-step process that evaluates ad copy through the eyes of your target audience, generates optimized variations, and predicts their success. This workflow draws inspiration from a similar approach shared by [Justin Brooke](https://x.com/IMJustinBrooke).\n",
    "url": "https://chatterkb.com/workflow-library/ad-copy-evaluation-optimization/",
    "order": 15
  }
,{
    "title": "Calendar Event Research - Company and Contact",
    "description": "Automatically research attendees and companies before meetings to prepare insights for more effective interactions",
    "content": "Calendar Event Research - Company and Contact\n\nOverview\n\nThis workflow automates pre-meeting intelligence gathering by researching attendees and their organizations, providing you with valuable context and talking points before every interaction.\n\nWorkflow Steps\n\nStep 1:\n- Check the calendar for tomorrow's upcoming events.\n\nStep 2:\nFor each event...\n- Search the web for the email address of the contact looking for posts on Reddit, LinkedIn, or any other social media site.\n- Pay attention to the domain of the contact's email and search for that company's domain (if it represents a company and not a mail service). Search the web for any interesting and recent news about the company.\n- Search go to any of the pages mentioned by the web search to acquire additional information.\n- If the contact or company information does not match the exact email address or domain name, do not consider it valid and do not reference it or assume that there is a similar or related link. This is important.\n\nStep 3:\n- Summarize all of the research with a section per event (with company and contact sections).\n- Then, send an email to me with the subject \"Tomorrow's Event Research\" that includes summary sections.\n",
    "introduction": "This workflow automatically prepares you for upcoming meetings by gathering intelligence on attendees and their companies. It pulls your calendar events, researches each participant's online presence, and collects recent company news to ensure you're prepared with relevant insights before every interaction. By automating pre-meeting research, you can walk into every conversation with valuable context, personalized talking points, and awareness of recent developments affecting your contacts.\n\n**IMPORTANT**: You must connect to an integration like Google to access your calendar before running this workflow.\n\n**Instructions**\n- Connect your Google Calendar or other supported calendar service via the Integrations page\n- Ensure you have email addresses for contacts in your calendar events\n- Run this workflow the day before your meetings to get timely research\n",
    "url": "https://chatterkb.com/workflow-library/calendar-event-research/",
    "order": 40
  }
,{
    "title": "Company Research and Outreach",
    "description": "Automated company research to qualify leads and prepare targeted outreach",
    "content": "Company Research and Outreach\n\nOverview\n\nThis workflow focuses on filling in missing information about a company in order to determine if it should be considered a viable lead. By automatically gathering and organizing key business details, it helps qualify prospects and prepare customized outreach materials.\n\nWorkflow Steps\n\nStep 1: Company Identification\nAsk: \"Please provide the name of a company and its location?\"\n\nStep 2: Initial Web Research\nSearch the web for information about the company, including:\n- Address\n- Website URL\n- Company Size\n- Industry\n- Services\n- Business Model\n- Contact Information\nDo not visit any site, just store the results.\n\nStep 3: Detailed Information Extraction\nBase on the search results from the previous step, investigate each result and attempt to capture the following information from each webpage:\n- Location (City/State)\n- Address (Full mailing address)\n- Website URL\n- Company Size (Provide a number, can be a ballpark)\n- Industry\n- Services\n- Business Model\n- Primary Contact (should be a specific person)\n- Primary Contact Email\n- Phone\n\nStep 4: Pain Point Analysis\nIf you have information about the company's industry, business model, or services:\na. Check the KB for information about the benefits and features of our company (YOUR COMPANY NAME HERE)\nb. Based on the company's industry, business model, and services, give me a list of the pain points with desired outcomes in their voice, going deep and make them polarizing.\nc. Create a file called pain-points-{{name of company with dashes}}.md and save the Pain Points for the company.\nOtherwise, skip this step and mark the Pain Points as \"Unknown\".\n\nStep 5: Data Consolidation\nAppend the information from the previous steps to the company-research-outreach.md file. Without adding the column headers, append the data to the markdown table in this order (Pain Points should contain the name of the file if Pain Points exist):\n| Company Name | Location | Address | Website URL | Company Size | Industry | Agency Type | Services | Business Model | Pain Points | Primary Contact Name | Primary Contact Email | Phone |\n",
    "introduction": "This workflow focuses on automatically researching and qualifying companies as potential leads by gathering critical business information from the web. It systematically identifies company details, creates structured records, and analyzes potential pain points and expected outcomes based on industry and business model. The workflow helps sales teams quickly determine if a company is a viable prospect and prepare targeted outreach materials aligned with their specific needs.\n\n**Instructions**\n- Add the following file to your Knowledge Base: **[company-research-outreach.md](/assets/files/company-research-outreach.md)**\n- Upload at least one document with the **features and benefits** of your company's product or service to your Knowledge Base.\n- Replace **YOUR COMPANY NAME HERE** with your company's name as mentioned in your document(s).\n",
    "url": "https://chatterkb.com/workflow-library/company-research-outreach/",
    "order": 30
  }
,{
    "title": "Create Friendly Company Names",
    "description": "Convert formal company names into memorable, conversational alternatives for better engagement and personalized communication",
    "content": "Create Friendly Company Names\n\nOverview\n\nThis workflow converts formal company names into simplified, conversational alternatives by intelligently extracting the most recognizable parts while removing special characters, legal designations, and unnecessary modifiers.\n\nWorkflow Steps\n\nStep 1:\n- Ask for a file name that has the columns:\nFirst Name, Last Name, Title, Company, Email\n- Create a spreadsheet called friendly-company-names.csv with only the header:\nFirst Name, Last Name, Title, Company, Friendly Company Name, Email\n\nStep 2:\nFor every row:\n- Using the Company name, you are going to come up with a friendly company name and place it in the Friendly Company Name column. I want your to come up with your best guess as to what someone might call the company in a casual conversation. You must remove special characters from names (e.g., ™, ®, ©, etc.).\n  Let me give you some examples:\n    Alea Advertising = Alea\n    Kai Communications & Branding = Kai\n    Depirrow/Garrone Advertising = DG\n    Alpha Co. Marketing & Media = Alpha\n    The Lubrizol = Lubrizol\n    EmpowerFi™ = EmpowerFi\n    Midwest Promotional Models = MPM (this one I would say this way because it isn't using generic terms so abbreviation might make sense)\n- Now append those values to the friendly-company-names.csv including all of the value from all of the columns. Do not include the column headers or any extra text before or after the rows.\n",
    "introduction": "This workflow automatically transforms formal company names into concise, conversational alternatives that are more relatable in everyday business communications. By creating standardized \"friendly\" versions of company names, your team can maintain consistent branding references across all touchpoints while removing special characters and unnecessary legal designations. This approach personalizes client communications, improves readability in reports, and helps maintain a more conversational tone in customer-facing materials.\n\n**Instructions**\n- Prepare a CSV file containing company contact information with columns for: First Name, Last Name, Title, Company, Email\n- Upload your file when prompted by the workflow\n",
    "url": "https://chatterkb.com/workflow-library/create-friendly-company-names/",
    "order": 35
  }
,{
    "title": "Customer Journey Decision Simulator",
    "description": "Simulate how real people make decisions—before you launch your campaign",
    "content": "Customer Journey Decision Simulator\n\nOverview\n\nThis workflow helps marketers understand how AI influences customer decision-making from initial curiosity to final purchase. By creating realistic buyer personas and mapping their decision journey, you can identify the key questions, concerns, and trade-offs that drive purchasing decisions. This insight helps optimize messaging, content strategy, and sales processes to better guide prospects through their buying journey.\n\nWorkflow Steps\n\nStep 1: Define Problem Statement\n\"What problem would you like to simulate? This should be a question where multiple valid approaches exist, such as 'Should I drink Coke or Pepsi?' or 'Should my small business focus on Instagram or TikTok for social media marketing?'\"\n\nStep 2: LLM Selection\nAsk: \"Which LLM would you like to use for this customer journey?\"\n\nStep 3: Generate Representative Persona\nCreate a persona that represents someone who might realistically ask this question:\n- Demographics (name, age, occupation, relevant background)\n- Context (why they're asking this question)\n- Knowledge level (expert, intermediate, beginner)\n- Specific needs or constraints\n\nStep 4: Customer Need Discovery Questions\nVerbatim (with substitutions):\nUsing the llm_evaluate tool with:\n- Prompt: \"As [persona], I'm trying to decide [problem statement]. What are the most important questions I should ask myself to better understand this decision?\"\n- Evaluation Prompt: \"Generate exactly three follow-up questions that would help [persona] better understand their specific needs and context for deciding [problem statement]. These questions should be open-ended, neutral, and focus on gathering essential information about their situation. Format your response as a numbered list of exactly three questions.\"\nProvide a brief summary of the results.\n\nStep 5: Decision Criteria Questions\nVerbatim (with substitutions):\nUsing the follow-up questions call the llm_evaluate tool:\n- Prompt: \"As [persona] deciding [problem statement], I need to consider these questions: [insert 3 problem understanding questions]. What criteria should I prioritize in making this decision?\"\n- Evaluation Prompt: \"Generate exactly three follow-up questions that would help [persona] prioritize decision criteria for [problem statement]. These questions should help them rank what factors matter most in their specific situation. Format your response as a numbered list of exactly three questions.\"\nProvide a brief summary of the results.\n\nStep 6: Trade-off Analysis Questions\nVerbatim (with substitutions):\nUsing the prioritize decision criteria questions call the llm_evaluate tool:\n- Prompt: \"As [persona] deciding [problem statement], I've identified these key criteria: [summarize criteria from previous step]. What trade-offs should I consider between these criteria?\"\n- Evaluation Prompt: \"Generate 2-3 questions that explore the most significant trade-offs [persona] might face when weighing different criteria for [problem statement]. These questions should highlight potential tensions between competing priorities. Format your response as a numbered list.\"\nProvide a brief summary of the results.\n\nStep 7: Final Decision Question\nVerbatim (with substitutions):\nUsing the questions that explore the most significant trade-offs call the llm_evaluate tool:\n- Prompt: \"As [persona] deciding [problem statement], I've considered these trade-offs: [summarize trade-offs from previous step]. What final question should I ask myself before making a decision?\"\n- Evaluation Prompt: \"Generate one final, comprehensive question that would help [persona] arrive at a recommendation for [problem statement]. This question should integrate their understanding of the problem, prioritized criteria, and trade-offs. Format your response as a single question.\"\nProvide a brief summary of the results.\n\nStep 8: Final Recommendation\nVerbatim (with substitutions):\nUsing the comprehensive question that would help arrive at a recommendation call the llm_evaluate tool:\n- Prompt: \"As [persona], considering [problem statement] and this final question: [insert final question], what would you recommend?\"\n- Evaluation Prompt: \"Analyze whether a clear recommendation was provided for [persona]'s [problem statement]. Identify if the response: 1) Makes a definitive recommendation, 2) Suggests multiple options with conditions, or 3) Avoids making a recommendation. Also note any hedging language or qualifiers used. Format your evaluation with clear section headings.\"\nProvide a brief summary of the results.\n\nStep 9: Process Summary and Insights\nVerbatim:\nProvide a report that:\n- States the question that was asked\n- Which LLM was used\n- Includes a brief executive summary\n- Includes a section with the detailed persona profile\n- Explains the steps that form the decision process and what was considered\n- Determine what recommendations the persona needs to choose from and assign a probability score that provides insight into which would likely be chosen (displayed as a list).\n",
    "introduction": "This workflow helps marketers understand how AI influences customer decision-making from initial curiosity to final purchase. By creating realistic buyer personas and mapping their decision journey, you can identify the key questions, concerns, and trade-offs that drive purchasing decisions. This insight helps optimize messaging, content strategy, and sales processes to better guide prospects through their buying journey.\n\n**Important:** Before running this workflow, make sure to enable the LLM Evaluate tool in your knowledge base settings.\n",
    "url": "https://chatterkb.com/workflow-library/customer-journey-decision-simulator/",
    "order": 25
  }
,{
    "title": "Email Security Verification Workflow",
    "description": "Automate email security record verification using ChatterKB.",
    "content": "Email Security Verification Workflow\n\nOverview:\nThis workflow guides you through verifying email security records for a domain using Google's DNS toolbox. Follow each step to ensure your domain's email security is properly configured.\n\nWorkflow Steps:\n\nStep 1: Identify Domain\nAsk: \"Which domain would you like to check for email security records? (e.g., example.com)\"\n\nStep 2: Check MX Records\nNavigate to: https://toolbox.googleapps.com/apps/dig/#MX/{{domain}}\n- Look for properly configured mail servers\n- Note priorities and TTL values\n- Verify they point to a valid email provider\n\nStep 3: Verify SPF Record\nNavigate to: https://toolbox.googleapps.com/apps/dig/#TXT/{{domain}}\n- Look for a TXT record beginning with \"v=spf1\"\n- Confirm it includes all authorized email senders\n- Check if it ends with ~all (softfail) or -all (hardfail)\n\nStep 4: Check DMARC Record\nNavigate to: https://toolbox.googleapps.com/apps/dig/#TXT/_dmarc.{{domain}}\n- Look for a TXT record beginning with \"v=DMARC1\"\n- Verify policy setting (p=none, p=quarantine, or p=reject)\n- Check reporting configuration if present\n\nStep 5: Verify DKIM Records\nNavigate to: https://toolbox.googleapps.com/apps/dig/#TXT/selector._domainkey.{{domain}}\n- Check for DKIM records using common selectors (default, mail, k1)\n- If no records are found, examine email headers from the domain to identify the actual selector\n- Look for TXT records containing \"v=DKIM1\"\n- Verify the record contains a valid public key\n\nStep 6: Analyze Results\nEvaluate each record against best practices:\n- MX: Properly configured mail servers\n- SPF: Authorized senders with appropriate policy\n- DMARC: Policy that matches your security needs\n- DKIM: Valid signing keys for email authentication\n\nStep 7: Identify Security Gaps\nNote any missing or misconfigured records:\n- Missing DKIM is a common issue\n- Weak SPF policy (~all instead of -all)\n- Missing or permissive DMARC policy\n\nStep 8: Recommend Improvements\nBased on findings, recommend specific actions to improve email security, such as:\n- Implementing missing records\n- Strengthening policies\n- Adding reporting for monitoring\n",
    "introduction": "This workflow guides you through verifying email security records for a domain using Google's DNS toolbox. Follow each step to ensure your domain's email security is properly configured.\n",
    "url": "https://chatterkb.com/workflow-library/email-security-verification-workflow/",
    "order": 55
  }
,{
    "title": "LLMO/GEO Analysis",
    "description": "Compare how different AI models rank and analyze topics across 16 leading language models simultaneously.",
    "content": "LLMO/GEO Analysis\n\nStep 1:\nSpecifically state: \"Ask the user for a prompt.\"\n\nStep 2:\nGenerate a unique id using the prompt and formatted as snake case and store it in memory in a key called \"unique_id_for_run\".\n\nStep 3:\nCreate a file named 'llmo_results_{{unique_id_for_run}}.md' with the following table structure:\n| Rank | Name | Mention Count | Associated Keywords | Related Links | Model |\n| ---- | ---- | ------------- | ------------------- | ------------- | ----- |\n**IMPORTANT** Mention this in the step: YOU **MUST** replace {{unique_id_for_run}} with the value found in memory for \"unique_id_for_run\".\n\nStep 4 - 20\nFor each of the following models:\n- Claude Sonnet 4\n- Claude 3.7 Sonnet\n- Claude 3.5 Sonnet V2\n- Claude 3.5 Haiku\n- DeepSeek-R1\n- Meta Llama 3.3 70B\n- Google Gemini 2.5 Pro\n- Google Gemini 2.5 Flash\n- Grok 3\n- Grok 3 Mini\n- Perplexity Sonar Pro\n- Perplexity Sonar\n- OpenAI GPT-4.1\n- OpenAI GPT-4o\n- OpenAI o4-mini\n- Amazon Nova Pro\n\nDo the following:\na. Call the LLMO tool with the prompt from Step 1 with the model.\nb. Store the entire results from the tool in a file using a name like {{model_name}}_{{unique_id_for_run}}.md'\nc. Specifically mention in Step: Add the statistics table (without the header) from the LLMO results as new rows in the 'llmo_results_{{unique_id_for_run}}.md' table.\n- Ensure that you add one column, at the end of the row, that includes the model name. Each row **must** contain values (even if blank) for Rank, Name, Mention Count, Associated Keywords, Related Links (where applicable), and Model (**IMPORTANT** do **not** include headers).\n- Example content format:\n| 1 | Skechers | 1 | Memory Foam Insoles, comfort for all-day wear | | Claude Sonnet 4 |\n- Use kb_write_file to append to the end of the file. IMPORTANT: You must use the default model for the steps. The model mentioned here is only for the llmo tool call.\n- **IMPORTANT** Mention this in the step: YOU **MUST** replace {{unique_id_for_run}} with the value found in memory for \"unique_id_for_run\".\n\nStep 21\nDisplay that you have finished the process.\n",
    "introduction": "This workflow is a bit more involved than the others, but it's still easy to follow and showcases the incredible control you have over complex, multi-step processes. You'll run the same prompt across 16 different AI models and automatically compile the results into a comprehensive analysis table.\n\n**Important:** Before running this workflow, make sure to enable the LLMO/GEO tool in your knowledge base settings.\n",
    "url": "https://chatterkb.com/workflow-library/llmo-geo-analysis/",
    "order": 40
  }
,{
    "title": "Market Intelligence and Sentiment Analysis Workflow",
    "description": "Research competitors, analyze sentiment, and surface actionable insights using ChatterKB.",
    "content": "Market Intelligence and Sentiment Analysis Workflow\n\nOverview:\nThis workflow helps marketers research competitors, industry trends, and brand sentiment by analyzing web content and organizing insights into actionable intelligence. Follow each step **exactly** as described. Map Steps to the Step Numbers and Titles provided below.\n\nWorkflow Steps:\n\nStep 1: Define research parameters\n\"What brand, competitor, or industry topic would you like to research? Please also specify any particular aspects you're interested in (e.g., product launches, marketing campaigns, customer sentiment).\"\n\nStep 2: Gather web content\nSearch for recent and relevant content about the specified topic using the web search tool. Collect articles, press releases, and social media discussions from the past 30 days.\n\nStep 3: Extract and parse content\nUse the web parser to extract the full content from the most relevant sources identified in Step 2. Focus on extracting clean text without navigation elements or advertisements.\n\nStep 4: Analyze sentiment and key themes\nProcess the extracted content to determine overall sentiment (positive, negative, neutral) and identify recurring themes, messaging strategies, and positioning statements.\n\nStep 5: Extract competitive intelligence\nIdentify specific marketing tactics, campaign elements, product features, pricing strategies, and target audience information from the analyzed content.\n\nStep 6: Generate structured insights\nOrganize the findings into a structured format with clear categories:\n- Sentiment Analysis (overall brand/topic perception)\n- Key Messaging Themes\n- Marketing Tactics & Strategies\n- Competitive Positioning\n- Emerging Trends & Opportunities\n\nStep 7: Create comparative data tables\nGenerate structured tables that organize the findings for easy comparison and analysis:\n\n| Source | Overall Sentiment | Key Messages | Marketing Tactics | Target Audience |\n|--------|------------------|--------------|------------------|----------------|\n| Source 1 | Positive/Negative/Neutral | Message 1, Message 2 | Tactic 1, Tactic 2 | Audience description |\n| Source 2 | Positive/Negative/Neutral | Message 1, Message 2 | Tactic 1, Tactic 2 | Audience description |\n\nAnd a sentiment summary table:\n\n| Sentiment Category | Percentage | Key Topics |\n|-------------------|------------|------------|\n| Positive | XX% | Topic 1, Topic 2 |\n| Neutral | XX% | Topic 3, Topic 4 |\n| Negative | XX% | Topic 5, Topic 6 |\n\nStep 8: Compile actionable recommendations\nBased on all gathered intelligence, generate specific, actionable recommendations for marketing strategy adjustments, content opportunities, or competitive responses.\n",
    "introduction": "If you want to monitor competitors, track industry sentiment, and turn raw web data into strategic advantage, here’s a step-by-step workflow you can run directly in ChatterKB.\n",
    "url": "https://chatterkb.com/workflow-library/market-intelligence-sentiment-analysis-workflow/",
    "order": 20
  }
,{
    "title": "New Client Onboarding – Guided Setup",
    "description": "Streamline client onboarding with automated questionnaires, document collection, and project kickoff workflows.",
    "content": "New Client Onboarding – Guided Setup\n\nWhen we start a new project, we want to move fast and skip redundant tasks. This assistant helps do that by walking us through only the essentials.\n\nFirst, it pulls client info automatically from the knowledge base if we upload an intake form or pitch deck. It should contain all of the information that follows:\n- Company name\n- Main contact (name + email) \n- Industry/vertical\n- Project type\n- Budget and start date\n- Key goals or deliverables\n\nNext, it checks our CRM. If the company exists, update the record. If not, it creates one using the info above.\n\nFinally, it generates an internal handoff summary with key dates and any open questions.\n",
    "introduction": "Client onboarding can make or break the relationship. This workflow transforms the chaotic first-week scramble into a **smooth, professional experience** that sets expectations and builds confidence from day one. This one leverages [Zapier MCP](https://mcp.zapier.com). ChatterKB is smart enough to know if you've added [HubSpot](https://www.hubspot.com/) as your CRM. It just works.\n",
    "url": "https://chatterkb.com/workflow-library/new-client-onboarding-guided-setup/",
    "order": 1000
  }
,{
    "title": "Review LLMO Analysis",
    "description": "Analyze and synthesize results from multi-model LLMO analysis to understand brand mention patterns and reasoning.",
    "content": "Review LLMO Analysis\n\nStep 1 - 16:\nCreate a step for each of these models:\n- Claude Sonnet 4\n- Claude 3.7 Sonnet\n- Claude 3.5 Sonnet V2\n- Claude 3.5 Haiku\n- DeepSeek-R1\n- Meta Llama 3.3 70B\n- Google Gemini 2.5 Pro\n- Google Gemini 2.5 Flash\n- Grok 3\n- Grok 3 Mini\n- Perplexity Sonar Pro\n- Perplexity Sonar\n- OpenAI GPT-4.1\n- OpenAI GPT-4o\n- OpenAI o4-mini\n- Amazon Nova Pro\n\nLogic must be part of the step: Generate a list of files in the KB filtered using model's name and scoped to the step. Do not include files with the name \"llmo_results_\".\n\n**IMPORTANT** The model mentioned in this step should not be used as the step's model. Use the default one instead. It is only to be used for the prompt in the step.\n\nCopy this to the step:\nFor every file associated to the model:\na) Use search_documents to find information about Brand A, Brand B, and Brand C. You are looking for reasons why Brand A and Brand B are mentioned before Brand C.  \nb) Store reasons in memory with a name/key like \"reasoning_for_mention_{{name of the file}}\" also, store it as an md file in the KB.  \n**Note**: Reason should be 2 - 3 sentences at most. Do not fabricate a reason. If unknown, save \"Cannot determine reason.\" in the memory and file.\n**IMPORTANT** Keep the steps simple and scoped to the model for this step (e.g. for each file: search_documents, analyze results, save to file)\n\nStep 17:\nCopy this to the step: Review all of the reasons and create a report detailing why Brand A, and Brand B, are mentioned before Brand C. Store this to an md file called \"detailed_report_reasoning.md\"\n\nStep 18:\nCopy this to the step: Generate an executive summary by reviewing all of the reasons and create a report detailing why Brand A and Brand B are mentioned before Brand C. Store this to an md file called \"executing_summary_reasoning.md\"\n",
    "introduction": "This workflow analyzes the results from your LLMO/GEO Analysis to understand why certain brands are mentioned before others. You'll need to customize the brand names and logic to match what you're testing — replace \"Brand A,\" \"Brand B,\" and \"Brand C\" with your actual brand names or competitors you want to analyze.\n\n**Important:** This workflow requires previous LLMO analysis results in your knowledge base and the LLMO/GEO tool enabled in your KB settings.\n",
    "url": "https://chatterkb.com/workflow-library/review-llmo-analysis/",
    "order": 50
  }
,{
    "title": "SEO Keyword Discovery Workflow",
    "description": "Automate keyword research, clustering, and brief generation using ChatterKB.",
    "content": "SEO Keyword Discovery Workflow\n\nOverview:\nThis workflow guides users through discovering the primary keyword for a webpage through content analysis rather than direct questioning. Follow each step **exactly** as described. Map Steps to the Step Numbers and Titles provided below.\n\nWorkflow Steps:\n\nStep 1: Get the webpage URL\n\"What webpage would you like to analyze for keywords?\"\n\nStep 2: Analyze the content\nLook through the page content and identify frequently used terms and phrases. Pay special attention to words that appear in the first paragraph and those that repeat throughout the text.\n\nStep 3: Check the URL structure\nThe URL often contains important keywords. Extract any meaningful terms from the URL path.\n\nStep 4: Review heading tags\nExamine H1, H2, and H3 tags since these typically contain the most important topics on the page.\n\nStep 5: Analyze page title and meta description\nThese elements are specifically crafted for SEO and often contain the primary keyword.\n\nStep 6: Identify potential keywords\nBased on the analysis, create a list of potential primary keywords and secondary keyword opportunities.\n\nStep 7: Evaluate keyword quality\nFor each potential keyword, consider:\n- How well it matches the page content\n- How specific it is to the page's purpose\n- What user intent it addresses (informational, transactional, navigational)\n\nStep 8: Recommend primary keyword\nSuggest the most likely primary keyword along with 2-3 secondary keyword opportunities.\n",
    "introduction": "If you want to reverse-engineer the primary keyword for any webpage (and find some secondary gold while you’re at it), here’s a step-by-step workflow you can run directly in ChatterKB.\n",
    "url": "https://chatterkb.com/workflow-library/seo-keyword-discovery-workflow/",
    "order": 10
  }
,{
    "title": "Summarize YouTube Videos from Transcript",
    "description": "Automate transcript extraction, summarization, and content repurposing for YouTube videos using ChatterKB.",
    "content": "Summarize YouTube Videos from Transcript\n\nOverview:\nThis workflow guides users through comprehensive analysis of YouTube videos by extracting and processing the transcript. Follow each step **exactly** as described. Map Steps to the Step Numbers and Titles provided below.\n\nWorkflow Steps:\n\nStep 1: Get the YouTube video URL\n\"What YouTube video would you like to analyze?\"\n\nStep 2: Extract the transcript\nExtract the full transcript from the YouTube video using the transcript tool. This will provide the raw text content for analysis.\n\nStep 3: Generate executive summary\nCreate a concise executive summary of the video content, highlighting 3-5 key points that represent the core message or information.\n\nStep 4: Identify key topics and themes\nAnalyze the transcript to identify the main topics and themes discussed in the video. Note any timestamps for important sections when possible.\n\nStep 5: Extract actionable insights\nIdentify any specific actions, recommendations, or steps mentioned in the video that viewers could implement.\n\nStep 6: Create structured Q&A\nGenerate a set of potential questions based on the video content along with comprehensive answers derived from the transcript.\n\nStep 7: Compare with knowledge base\nCompare the information in the video with existing knowledge in your database to identify new insights or contradictions.\n\nStep 8: Generate follow-up content\nCreate social media posts, newsletter blurbs, or other content formats based on the video's key points for easy sharing.\n",
    "introduction": "If you want to extract meaningful insights from any YouTube video (and quickly repurpose them for your audience), here’s a step-by-step workflow you can run directly in ChatterKB.\n\n**Important:** Before running this workflow, make sure to enable the **YouTube Transcript** tool in your knowledge base settings **Note:** You may need a URL proxy. You can provide one in the KB's advanced settings tab.\n",
    "url": "https://chatterkb.com/workflow-library/youtube-video-analysis-workflow/",
    "order": 30
  }

]