Hey, I have created a prompt to generate some bullet points according to the job description. It's giving me invalid output error. I have tried different models, changed output schema. Double checked input variables still the problem is there. How can I solve this?
Invalid output errors in AI prompts can happen for a few reasons:
Make sure all required input fields have data—use the "Required to Run" toggle in the Prompt Editor to skip empty rows [1].
Double-check your prompt variables and output schema. If you’re referencing dynamic data, ensure you’re using the right column names and syntax [2].
If you’re using JSON, watch for formatting issues like missing quotes, trailing commas, or unescaped characters. Use a JSON validator to catch these [3].
If you’re using an HTTP API, ensure your body is properly formatted with quotes around strings and correct punctuation [4].
Try these steps and let me know if you’re still stuck!
Was that helpful?
If you still need help resolving the invalid output error, I’m here to assist further. Would you like to provide more details about your prompt or the specific error you’re encountering?
The variables are mapped correctly, data is not missing as shared in the image. Toggle required to run is off but still not working
"Invalid input" errors usually mean the prompt or data format isn't right, even if variables are mapped and data looks present. Try these:
Double-check your prompt's output schema—make sure it matches what the model expects (like JSON structure, if used). Watch for missing quotes, commas, or brackets.
If you're using JSON, validate the format with a tool to catch syntax issues.
Click into the cell with the error to see more details about what's wrong.
If you're using conditional runs, make sure the run conditions are set correctly and not blocking execution.
If you can share the exact prompt or error message, I can help spot the issue faster. [1]
Was that helpful?
If you still need help with the invalid output error, I’m here to assist further and apologize for the inconvenience. Would you like to share more details about the prompt or error messages so we can continue troubleshooting?
Inputs: - job_description: - company_domain: IMPORTANT CONSTRAINTS - DO NOT: - Do NOT copy any wording, phrasing, or exact terminology directly from Job Description - Do NOT create a 100% perfect match candidate — aim for approximately 80% match only - Do NOT include soft skills (e.g., "team player," "strong communicator," "problem solver") - Do NOT include a candidate name - Do NOT list all technologies mentioned in the JD — select only 1-2 specific technologies, then use broader skill categories - Do NOT invent skills or technologies not reasonably inferred from the JD - Do NOT fabricate years of experience that exceed what the JD requests — always go slightly under (e.g., if JD asks for 5+ years, suggest 4 years) - Do NOT write fewer than 5 words or more than 10 words per bullet point ONLY: - Generate ONE fictional candidate persona based on Job Description - Include technical skills only - Paraphrase all skills and requirements into your own words - Create a realistic, market-available candidate profile (80% match makes them findable) - Keep each bullet point to 5-10 words --- OBJECTIVE: Analyze Job Description from company_domain and generate ONE fictional ideal candidate persona that represents an 80% match for the role. This candidate should be realistic and representative of what HAILO Consulting can actually source from the market — not a unicorn. The 80% match principle means: - Slightly fewer years of experience than requested - Not all technologies listed, but core relevant ones - Strong fit on 2-3 key requirements, acceptable gaps on others --- GENERATION PROCESS: 1. ANALYZE to identify: - Core role requirements (must-haves) - Technical stack and tools mentioned - Years of experience requested - Industry or domain context 2. SELECT for the candidate profile: - 1-2 specific technologies from the JD (paraphrased if possible) - 1-2 broader skill categories that complement the role (e.g., "cloud infrastructure" instead of just "AWS") - Years of experience slightly below JD requirement 3. GENERATE the candidate persona with exactly: - 3 bullet points for "Why this candidate" - 3 bullet points for "Reasons" - 3 bullet points for "Technical Skills" --- VALIDATION RULES: - All skills must be technical only — zero soft skills - No skill should use exact wording from the JD (paraphrase required) - Years of experience must be equal to or less than JD requirement - Candidate must NOT be a perfect match — intentional gaps are required - At least one skill should be a broader category, not a specific tool - Each bullet point must be 5-10 words (aim for 5-7, extend to 10 for clarity) IF NOT FOUND: If is too vague, lacks technical requirements, or provides insufficient information, set is_valid to false and return empty arrays. --- CRITICAL REMINDERS: - Do NOT copy exact wording from — paraphrase everything - Do NOT create a 100% match — aim for 80% (realistic, sourceable candidate) - Do NOT include any soft skills — technical skills only - Do NOT include a candidate name - Keep all bullets between 5-10 words - Mix 1-2 specific technologies with broader skill categories
"Invalid input" errors usually mean there's a formatting issue in your prompt or output schema, even if variables are mapped and data looks present.
Double-check your prompt's output schema—make sure it matches what the model expects (like JSON structure, if used). Watch for missing quotes, commas, or brackets.
If you're using JSON, validate the format with a tool to catch syntax issues.
Click into the cell with the error to see more details about what's wrong.
If you can share the exact prompt or error message, I can help spot the issue faster.
Is that what you were looking for?
If you still need help resolving the invalid output error, I apologize for the inconvenience and am here to assist further. Would you like to provide more details about your prompt or error messages so we can continue troubleshooting?
talk to human support
I understand your frustration and will connect you with a human agent who can assist you further.
Our support team has got your message and we'll get back to you soon! We're currently outside of our standard office hours (9 AM to 9 PM EST, Monday through Friday), so it may take a little longer for us to respond.
If you’re dealing with a specific table, drop the URL below so we can help you quicker. Otherwise, someone from our team will be in touch soon!

