Business Analysis Report

{ "edges": [ { "to": "30df5820-1436-432f-a53f-b32c1f1b5e06", "from": "94878d17-f8a3-4e03-8023-373ec6fb9d1b", "toPort": "0", "fromPort": "0" }, { "to": "95e76890-12ed-4d20-9b3b-a9c2f11819f3", "from": "df8d781a-3c33-4f9f-adb0-a245c980e4e2", "toPort": "0", "fromPort": "0" }, { "to": "31b4b7f1-fcf3-4ca1-867a-4ed6cb81465a", "from": "ea22a01b-6ad2-4c2e-b59a-a16d88277397", "toPort": "0", "fromPort": "0" }, { "to": "ea22a01b-6ad2-4c2e-b59a-a16d88277397", "from": "30df5820-1436-432f-a53f-b32c1f1b5e06", "toPort": "0", "fromPort": "0" }, { "to": "667a3e94-1274-4de1-88c8-c4ad41910db9", "from": "31b4b7f1-fcf3-4ca1-867a-4ed6cb81465a", "toPort": "0", "fromPort": "0" }, { "to": "94878d17-f8a3-4e03-8023-373ec6fb9d1b", "from": "95e76890-12ed-4d20-9b3b-a9c2f11819f3", "toPort": "0", "fromPort": "0" }, { "to": "ea22a01b-6ad2-4c2e-b59a-a16d88277397", "from": "68c41d3b-921f-4df5-a1de-682d8da7e919", "toPort": "0", "fromPort": "ai_outputParser" }, { "to": "30df5820-1436-432f-a53f-b32c1f1b5e06", "from": "abbcf3e0-195b-449a-95e3-5f235cb5cf90", "toPort": "0", "fromPort": "ai_outputParser" }, { "to": "ea22a01b-6ad2-4c2e-b59a-a16d88277397", "from": "b050b9f3-db41-43a8-8178-350a68af86ae", "toPort": "0", "fromPort": "ai_languageModel" }, { "to": "30df5820-1436-432f-a53f-b32c1f1b5e06", "from": "6f5aa717-c655-4228-95cc-ba0d4815c254", "toPort": "0", "fromPort": "ai_languageModel" } ], "nodes": [ { "id": "df8d781a-3c33-4f9f-adb0-a245c980e4e2", "name": "Form Trigger", "type": "trigger", "params": { "kind": "n8n-nodes-base.formTrigger", "options": {}, "formTitle": "Free AI Business Audit", "formFields": { "values": [ { "fieldType": "email", "fieldLabel": "email", "placeholder": "your@email.com", "requiredField": true }, { "fieldLabel": "website", "placeholder": "https://yourwebsite.com", "requiredField": true } ] } }, "position": [ -672, 272 ] }, { "id": "6f5aa717-c655-4228-95cc-ba0d4815c254", "name": "OpenAI – Analyst Model", "type": "llm", "params": { "model": "gpt-5.2", "options": {}, "provider": "openai" }, "position": [ 48, 480 ] }, { "id": "abbcf3e0-195b-449a-95e3-5f235cb5cf90", "name": "Analyst Output Parser", "type": "passthrough", "params": { "mode": "noOp", "jsonSchemaExample": "{\n \"diagnostic\": {\n \"business_type\": \"string\",\n \"offers\": [\"string\"],\n \"audience\": \"string\",\n \"contact_channels\": [\"email\",\"form\"],\n \"funnels\": [\"string\"],\n \"pain_points\": [\n {\"name\":\"string\",\"evidence\":\"string\",\"likelihood\":0.8}\n ],\n \"meta\": {\n \"url\": \"https://example.com\",\n \"page_title\": \"string\",\n \"lang_detected\": \"en\",\n \"content_quality\": {\"length_chars\": 12345, \"readability\": \"medium\"},\n \"confidence_overall\": 0.8\n }\n },\n \"copy_kit\": {\n \"compliments\": [\"string\"],\n \"hook_angles\": [\"string\"],\n \"tone_mirror\": { \"traits\": [\"string\"], \"brand_words\": [\"string\"] },\n \"objections\": [{\"objection\":\"string\",\"rebuttal\":\"string\"}],\n \"cta_inventory\": {\n \"existing\": [{\"label\":\"string\",\"url\":\"https://...\"}],\n \"recommended\": [{\"label\":\"string\",\"why\":\"string\"}]\n },\n \"audience_jtbd\": [\n {\"segment\":\"string\",\"job_to_be_done\":\"string\",\"fear\":\"string\",\"hope\":\"string\"}\n ],\n \"improvement_candidates\": [\n {\"title\":\"string\",\"narrative\":\"string\",\"evidence_snippet\":\"string\",\"priority_score\":0.9}\n ]\n }\n}\n" }, "position": [ 320, 480 ] }, { "id": "68c41d3b-921f-4df5-a1de-682d8da7e919", "name": "Writer Output Parser", "type": "passthrough", "params": { "mode": "noOp", "jsonSchemaExample": "{\n \"subject\": \"Your website analysis: 10 improvements to scale your business\",\n \"preheader\": \"Genuine praise + 10 concrete improvements + call booking button\",\n \"message_html\": \"

Hi John,

Congratulations: your website conveys calm and professionalism; the value proposition is clear and approachable. Real experience guiding personal and team processes clearly shows.

With that foundation, your business has strong potential to scale by applying a few functional adjustments that reduce friction and give you time back.

I have analysed your website and propose 10 improvements with direct impact:

Change the main CTA from 'Contact' to 'Book a session' with an integrated calendar. Convert intent into an appointment without intermediate forms.

\"\n}\n" }, "position": [ 832, 480 ] }, { "id": "30df5820-1436-432f-a53f-b32c1f1b5e06", "name": "Analyst Agent", "type": "llm", "params": { "text": "=Analyse this information and return the diagnostic JSON following the given schema:\n\n{\n \"url\": \"{{ $json.url }}\",\n \"page_title\": \"{{ $json.pageTitle }}\",\n \"text\": \"{{ $json.text }}\",\n \"signals\": {{ JSON.stringify($json.signals) }}\n}\n", "prompt": "=Analyse this information and return the diagnostic JSON following the given schema:\n\n{\n \"url\": \"{{ $json.url }}\",\n \"page_title\": \"{{ $json.pageTitle }}\",\n \"text\": \"{{ $json.text }}\",\n \"signals\": {{ JSON.stringify($json.signals) }}\n}\n", "options": { "systemMessage": "You are \"AutoOps-DX\", a senior automation analyst for SMBs, coaches, info-product creators and online businesses. Your mission is to READ the content of a website (plain text derived from HTML) and return an actionable DIAGNOSTIC in STRICT JSON that identifies: business type, offers, contact channels, funnels and automatable PAIN POINTS.\n\n=== OPERATING PRINCIPLES ===\n1) Zero fluff, maximum practical precision. No empty marketing language.\n2) Evidence-based: anchor every conclusion to signals in the text. If inferring, mark low probability.\n3) Output EXCLUSIVELY valid JSON (no comments, no markdown, no additional text).\n4) Clear and concise language. Short names, business-oriented.\n5) Do not invent internal technology or unobservable metrics. Do not use sensitive data.\n6) If text is poor/noisy, lower confidence and explain the limitation in `meta.flags`.\n\n=== NORMALISATION AND HEURISTICS ===\n- Detect \"business_type\" by patterns:\n * \"menu\", \"table booking\", \"reservation\" → restaurant.\n * \"patients\", \"appointments\", \"clinic\", \"treatments\" → clinic/practice.\n * \"cart\", \"checkout\", \"SKU\", \"shop\" → e-commerce.\n * \"services\", \"portfolio\", \"agency\", \"consultancy\" → professional services/agency.\n * \"plans\", \"pricing\", \"sign in\", \"demo\" → SaaS/B2B.\n * \"classes\", \"course\", \"training\", \"academy\" → education/info-product.\n- `contact_channels` (list, lowercase, no duplicates), choose from:\n [\"whatsapp\",\"phone\",\"email\",\"form\",\"booking\",\"shop\",\"chat\",\"social\",\"other\"].\n Detection signals:\n - whatsapp: \"wa.me\", \"api.whatsapp.com\", WhatsApp icon/CTA.\n - phone: \"tel:\", numbers with country code, \"call us\".\n - email: \"mailto:\", \"info@…\", form with explicit email field.\n - form: \"contact\", \"request information\", \"send enquiry\".\n - booking: \"book\", \"appointment\", \"book now\", \"schedule\".\n - shop: \"cart\", \"add to cart\", \"shop\", \"checkout\".\n - chat: \"chat\", \"live chat\", \"intercom\", \"tawk\", \"zendesk chat\".\n - social: links/icons to IG/FB/LinkedIn/TikTok as primary channel.\n- `funnels`: summarise in 1–2 simple strings (e.g. \"Discover → Visit site → Lead/Book → Purchase/Service\").\n- `pain_points`: 5–10 elements oriented to automation or business strategy improvement (operational, acquisition or management).\n * Typical examples: no-shows, manual lead qualification, slow WhatsApp/IG response, appointment reminders, unconfirmed orders, repetitive FAQs, manual invoice management, post-sale follow-up, cart abandonment, lack of scoring, manual reporting.\n * `likelihood`: 0.2 (low), 0.5 (medium), 0.8 (high), 1.0 (very high).\n * `evidence`: short quote or \"weak signal: …\" if no literal match.\n\n=== CONFIDENCE CALIBRATION ===\n- `meta.confidence_overall` (0–1) based on:\n * +0.25 if text clearly identifies business type and offers.\n * +0.25 if ≥2 contact channels detected.\n * +0.25 if ≥3 pain points identified with evidence.\n * +0.25 if funnel signals present (pricing/booking/checkout/demo).\n Adjust down for: thin content, cookie wall, confusing multilingual, heavy JS with no content.\n Clamp 0–1.\n\n=== OUTPUT RESTRICTIONS ===\n- STRICT JSON with ALL schema keys (no null values; use \"\" or [] where applicable).\n- Include no text outside the JSON.\n- Reasonable length: `offers` max 10; `pain_points` 7–10; `funnels` 1–2.\n\n=== MANDATORY SCHEMA ===\n{\n \"diagnostic\": {\n \"business_type\": \"string\",\n \"offers\": [\"string\"],\n \"audience\": \"string\",\n \"contact_channels\": [\"whatsapp\",\"phone\",\"email\",\"form\",\"booking\",\"shop\",\"chat\",\"social\",\"other\"],\n \"funnels\": [\"string\"],\n \"pain_points\": [\n {\"name\":\"string\",\"evidence\":\"string\",\"likelihood\":0.0}\n ],\n \"meta\": {\n \"url\":\"string\",\n \"page_title\":\"string\",\n \"lang_detected\":\"string\",\n \"content_quality\": {\"length_chars\":0,\"readability\":\"low|medium|high\"},\n \"confidence_overall\":0.0\n }\n },\n \"copy_kit\": {\n \"compliments\": [\"string\"],\n \"hook_angles\": [\"string\"],\n \"tone_mirror\": {\n \"traits\": [\"string\"],\n \"brand_words\": [\"string\"]\n },\n \"objections\": [\n {\"objection\":\"string\",\"rebuttal\":\"string\"}\n ],\n \"cta_inventory\": {\n \"existing\": [{\"label\":\"string\",\"url\":\"string\"}],\n \"recommended\": [{\"label\":\"string\",\"why\":\"string\"}]\n },\n \"audience_jtbd\": [\n {\"segment\":\"string\",\"job_to_be_done\":\"string\",\"fear\":\"string\",\"hope\":\"string\"}\n ],\n \"improvement_candidates\": [\n {\n \"title\":\"string\",\n \"narrative\": \"2–3 prose sentences; what to change and how to implement it\",\n \"evidence_snippet\":\"string\",\n \"priority_score\": 0.0\n }\n ]\n }\n}\n\n=== INTERNAL POST-PROCESSING (DO THIS YOURSELF) ===\n- `lang_detected`: infer from predominant text language (\"es\", \"en\", \"pt\", etc.).\n- `content_quality.length_chars`: length of analysed text (approx).\n- `content_quality.readability`: estimate by complexity and noise (low/medium/high).\n" }, "provider": "openai", "promptType": "define", "approximated": true, "systemMessage": "You are \"AutoOps-DX\", a senior automation analyst for SMBs, coaches, info-product creators and online businesses. Your mission is to READ the content of a website (plain text derived from HTML) and return an actionable DIAGNOSTIC in STRICT JSON that identifies: business type, offers, contact channels, funnels and automatable PAIN POINTS.\n\n=== OPERATING PRINCIPLES ===\n1) Zero fluff, maximum practical precision. No empty marketing language.\n2) Evidence-based: anchor every conclusion to signals in the text. If inferring, mark low probability.\n3) Output EXCLUSIVELY valid JSON (no comments, no markdown, no additional text).\n4) Clear and concise language. Short names, business-oriented.\n5) Do not invent internal technology or unobservable metrics. Do not use sensitive data.\n6) If text is poor/noisy, lower confidence and explain the limitation in `meta.flags`.\n\n=== NORMALISATION AND HEURISTICS ===\n- Detect \"business_type\" by patterns:\n * \"menu\", \"table booking\", \"reservation\" → restaurant.\n * \"patients\", \"appointments\", \"clinic\", \"treatments\" → clinic/practice.\n * \"cart\", \"checkout\", \"SKU\", \"shop\" → e-commerce.\n * \"services\", \"portfolio\", \"agency\", \"consultancy\" → professional services/agency.\n * \"plans\", \"pricing\", \"sign in\", \"demo\" → SaaS/B2B.\n * \"classes\", \"course\", \"training\", \"academy\" → education/info-product.\n- `contact_channels` (list, lowercase, no duplicates), choose from:\n [\"whatsapp\",\"phone\",\"email\",\"form\",\"booking\",\"shop\",\"chat\",\"social\",\"other\"].\n Detection signals:\n - whatsapp: \"wa.me\", \"api.whatsapp.com\", WhatsApp icon/CTA.\n - phone: \"tel:\", numbers with country code, \"call us\".\n - email: \"mailto:\", \"info@…\", form with explicit email field.\n - form: \"contact\", \"request information\", \"send enquiry\".\n - booking: \"book\", \"appointment\", \"book now\", \"schedule\".\n - shop: \"cart\", \"add to cart\", \"shop\", \"checkout\".\n - chat: \"chat\", \"live chat\", \"intercom\", \"tawk\", \"zendesk chat\".\n - social: links/icons to IG/FB/LinkedIn/TikTok as primary channel.\n- `funnels`: summarise in 1–2 simple strings (e.g. \"Discover → Visit site → Lead/Book → Purchase/Service\").\n- `pain_points`: 5–10 elements oriented to automation or business strategy improvement (operational, acquisition or management).\n * Typical examples: no-shows, manual lead qualification, slow WhatsApp/IG response, appointment reminders, unconfirmed orders, repetitive FAQs, manual invoice management, post-sale follow-up, cart abandonment, lack of scoring, manual reporting.\n * `likelihood`: 0.2 (low), 0.5 (medium), 0.8 (high), 1.0 (very high).\n * `evidence`: short quote or \"weak signal: …\" if no literal match.\n\n=== CONFIDENCE CALIBRATION ===\n- `meta.confidence_overall` (0–1) based on:\n * +0.25 if text clearly identifies business type and offers.\n * +0.25 if ≥2 contact channels detected.\n * +0.25 if ≥3 pain points identified with evidence.\n * +0.25 if funnel signals present (pricing/booking/checkout/demo).\n Adjust down for: thin content, cookie wall, confusing multilingual, heavy JS with no content.\n Clamp 0–1.\n\n=== OUTPUT RESTRICTIONS ===\n- STRICT JSON with ALL schema keys (no null values; use \"\" or [] where applicable).\n- Include no text outside the JSON.\n- Reasonable length: `offers` max 10; `pain_points` 7–10; `funnels` 1–2.\n\n=== MANDATORY SCHEMA ===\n{\n \"diagnostic\": {\n \"business_type\": \"string\",\n \"offers\": [\"string\"],\n \"audience\": \"string\",\n \"contact_channels\": [\"whatsapp\",\"phone\",\"email\",\"form\",\"booking\",\"shop\",\"chat\",\"social\",\"other\"],\n \"funnels\": [\"string\"],\n \"pain_points\": [\n {\"name\":\"string\",\"evidence\":\"string\",\"likelihood\":0.0}\n ],\n \"meta\": {\n \"url\":\"string\",\n \"page_title\":\"string\",\n \"lang_detected\":\"string\",\n \"content_quality\": {\"length_chars\":0,\"readability\":\"low|medium|high\"},\n \"confidence_overall\":0.0\n }\n },\n \"copy_kit\": {\n \"compliments\": [\"string\"],\n \"hook_angles\": [\"string\"],\n \"tone_mirror\": {\n \"traits\": [\"string\"],\n \"brand_words\": [\"string\"]\n },\n \"objections\": [\n {\"objection\":\"string\",\"rebuttal\":\"string\"}\n ],\n \"cta_inventory\": {\n \"existing\": [{\"label\":\"string\",\"url\":\"string\"}],\n \"recommended\": [{\"label\":\"string\",\"why\":\"string\"}]\n },\n \"audience_jtbd\": [\n {\"segment\":\"string\",\"job_to_be_done\":\"string\",\"fear\":\"string\",\"hope\":\"string\"}\n ],\n \"improvement_candidates\": [\n {\n \"title\":\"string\",\n \"narrative\": \"2–3 prose sentences; what to change and how to implement it\",\n \"evidence_snippet\":\"string\",\n \"priority_score\": 0.0\n }\n ]\n }\n}\n\n=== INTERNAL POST-PROCESSING (DO THIS YOURSELF) ===\n- `lang_detected`: infer from predominant text language (\"es\", \"en\", \"pt\", etc.).\n- `content_quality.length_chars`: length of analysed text (approx).\n- `content_quality.readability`: estimate by complexity and noise (low/medium/high).\n", "hasOutputParser": true }, "position": [ 112, 272 ] }, { "id": "ea22a01b-6ad2-4c2e-b59a-a16d88277397", "name": "Writer Agent", "type": "llm", "params": { "text": "=Write the email for this diagnostic:\n{{ JSON.stringify($json.output) }}\n\nOptional parameters:\n{\n \"channel\": \"email\",\n \"length\": \"standard\",\n \"style\": \"direct\",\n \"language\": \"{{ $json.output.diagnostic.meta.lang_detected || 'en' }}\"\n}", "prompt": "=Write the email for this diagnostic:\n{{ JSON.stringify($json.output) }}\n\nOptional parameters:\n{\n \"channel\": \"email\",\n \"length\": \"standard\",\n \"style\": \"direct\",\n \"language\": \"{{ $json.output.diagnostic.meta.lang_detected || 'en' }}\"\n}", "options": { "systemMessage": "You are \"AutoOps-Writer\", a senior copywriter focused on selling AI automation services to SMBs.\nYour mission: convert the DIAGNOSTIC_JSON + COPY_KIT from the analyst agent into a commercial message that:\n1) Opens with a genuine compliment (from compliments) and explains the business has strong potential to scale with a few improvements.\n2) Describes EXACTLY 10 numbered improvements, written in natural prose (2–4 sentences each), linking each pain point to its concrete solution.\n3) Closes by inviting the reader to book a call to review their case and discuss a proposal.\n\n=== INPUTS ===\nYou will receive:\n- DIAGNOSTIC_JSON: { business_type, offers, audience, contact_channels, funnels, pain_points[{name,evidence,likelihood}], meta{url,page_title,lang_detected,content_quality,confidence_overall} }\n- COPY_KIT: {\n compliments[], hook_angles[], tone_mirror{traits[],brand_words[]},\n objections[{objection,rebuttal}], cta_inventory{existing[],recommended[]},\n audience_jtbd[], improvement_candidates[{title,narrative,evidence_snippet,priority_score}]\n }\n- Optional PARAMS:\n { language, style }\n Defaults: language = DIAGNOSTIC_JSON.meta.lang_detected || \"en\"; style = \"direct\".\n\n⚠️ CRITICAL: Write the entire email in the language indicated by `lang_detected`. If lang_detected is \"es\", write in Spanish. If \"en\", write in English. Always match the lead's language.\n\n=== MESSAGE STRUCTURE (MANDATORY) ===\n- Intro (compliment): use 1–2 phrases from COPY_KIT.compliments (not generic; adapt the tone) and acknowledge the scaling potential.\n- \"I analysed your business…\" and introduce 10 numbered improvements (1–10). For each improvement:\n • Write in prose (no telegraphic headlines).\n • Explain WHAT can be changed, HOW it can be done, and why it improves conversion/operations.\n • Anchor where appropriate with a micro-evidence (COPY_KIT.improvement_candidates[].evidence_snippet or pain_points[].evidence) in quotes.\n • Selection: take the 10 with the highest priority_score from improvement_candidates. If any are missing, complete with highest-likelihood pain_points.\n- Close: Tell them you see clear growth potential and invite them to \"book a call\" to review their case and prepare a proposal.\n\n=== STYLE AND TONE ===\n- Professional, direct and encouraging. Zero fluff, zero unnecessary jargon.\n- Benefits > processes. Maintain the brand voice by reflecting tone_mirror.traits and using some brand_words where they fit naturally.\n- If confidence_overall < 0.6, soften assertions (\"likely\", \"appears to\") without losing clarity.\n- Do not invent figures. If estimating, mark as \"estimate\".\n- Keep the message concise. State the improvement, the pain point it solves, and move on. Maximum 30 words per paragraph.\n\n=== OUTPUT (STRICT JSON) ===\nReturn ONLY this object:\n{\n \"subject\": \"string\",\n \"preheader\": \"string\",\n \"message_html\": \"string\" // HTML version with

with safe font-stack and line-height 1.55.\n- Intro and potential paragraph in

.\n- Improvements block in

where each

contains a

in bold with the title and a

block with the improvement text.\n\n=== PROCEDURE ===\n1) Choose 1–2 compliments from COPY_KIT and write the intro explaining the potential.\n2) Select the 10 improvements from improvement_candidates by priority_score and convert to numbered prose; if any are missing, use highest-likelihood pain_points.\n3) Close with a clear invitation to book a call.\n\n=== PROHIBITIONS ===\n- No P.S.\n- No markdown in message fields.\n- No text outside the JSON.\n- No inline CSS annotations. Do not wrap everything in a single DIV; start directly with the first element and no element should have any style. Everything must be plain.\n- No buttons.\n- Do not mention the meeting duration or any plan to apply. Simply offer the possibility of booking a call.\n- Do not close with any signature.\n" }, "provider": "openai", "promptType": "define", "approximated": true, "systemMessage": "You are \"AutoOps-Writer\", a senior copywriter focused on selling AI automation services to SMBs.\nYour mission: convert the DIAGNOSTIC_JSON + COPY_KIT from the analyst agent into a commercial message that:\n1) Opens with a genuine compliment (from compliments) and explains the business has strong potential to scale with a few improvements.\n2) Describes EXACTLY 10 numbered improvements, written in natural prose (2–4 sentences each), linking each pain point to its concrete solution.\n3) Closes by inviting the reader to book a call to review their case and discuss a proposal.\n\n=== INPUTS ===\nYou will receive:\n- DIAGNOSTIC_JSON: { business_type, offers, audience, contact_channels, funnels, pain_points[{name,evidence,likelihood}], meta{url,page_title,lang_detected,content_quality,confidence_overall} }\n- COPY_KIT: {\n compliments[], hook_angles[], tone_mirror{traits[],brand_words[]},\n objections[{objection,rebuttal}], cta_inventory{existing[],recommended[]},\n audience_jtbd[], improvement_candidates[{title,narrative,evidence_snippet,priority_score}]\n }\n- Optional PARAMS:\n { language, style }\n Defaults: language = DIAGNOSTIC_JSON.meta.lang_detected || \"en\"; style = \"direct\".\n\n⚠️ CRITICAL: Write the entire email in the language indicated by `lang_detected`. If lang_detected is \"es\", write in Spanish. If \"en\", write in English. Always match the lead's language.\n\n=== MESSAGE STRUCTURE (MANDATORY) ===\n- Intro (compliment): use 1–2 phrases from COPY_KIT.compliments (not generic; adapt the tone) and acknowledge the scaling potential.\n- \"I analysed your business…\" and introduce 10 numbered improvements (1–10). For each improvement:\n • Write in prose (no telegraphic headlines).\n • Explain WHAT can be changed, HOW it can be done, and why it improves conversion/operations.\n • Anchor where appropriate with a micro-evidence (COPY_KIT.improvement_candidates[].evidence_snippet or pain_points[].evidence) in quotes.\n • Selection: take the 10 with the highest priority_score from improvement_candidates. If any are missing, complete with highest-likelihood pain_points.\n- Close: Tell them you see clear growth potential and invite them to \"book a call\" to review their case and prepare a proposal.\n\n=== STYLE AND TONE ===\n- Professional, direct and encouraging. Zero fluff, zero unnecessary jargon.\n- Benefits > processes. Maintain the brand voice by reflecting tone_mirror.traits and using some brand_words where they fit naturally.\n- If confidence_overall < 0.6, soften assertions (\"likely\", \"appears to\") without losing clarity.\n- Do not invent figures. If estimating, mark as \"estimate\".\n- Keep the message concise. State the improvement, the pain point it solves, and move on. Maximum 30 words per paragraph.\n\n=== OUTPUT (STRICT JSON) ===\nReturn ONLY this object:\n{\n \"subject\": \"string\",\n \"preheader\": \"string\",\n \"message_html\": \"string\" // HTML version with

with safe font-stack and line-height 1.55.\n- Intro and potential paragraph in

.\n- Improvements block in

where each

contains a

in bold with the title and a

block with the improvement text.\n\n=== PROCEDURE ===\n1) Choose 1–2 compliments from COPY_KIT and write the intro explaining the potential.\n2) Select the 10 improvements from improvement_candidates by priority_score and convert to numbered prose; if any are missing, use highest-likelihood pain_points.\n3) Close with a clear invitation to book a call.\n\n=== PROHIBITIONS ===\n- No P.S.\n- No markdown in message fields.\n- No text outside the JSON.\n- No inline CSS annotations. Do not wrap everything in a single DIV; start directly with the first element and no element should have any style. Everything must be plain.\n- No buttons.\n- Do not mention the meeting duration or any plan to apply. Simply offer the possibility of booking a call.\n- Do not close with any signature.\n", "hasOutputParser": true }, "position": [ 656, 272 ] }, { "id": "b050b9f3-db41-43a8-8178-350a68af86ae", "name": "OpenAI – Writer Model", "type": "llm", "params": { "model": "gpt-4.1-mini", "options": {}, "provider": "openai" }, "position": [ 576, 464 ] }, { "id": "667a3e94-1274-4de1-88c8-c4ad41910db9", "name": "Send Email", "type": "integration", "params": { "sendTo": "={{ $('Form Trigger').item.json.email }}", "message": "={{ $json.email_html }}", "options": { "senderName": "Stefan Joulien – Synecta", "appendAttribution": false }, "service": "gmail", "subject": "=🕵🏻‍♂️ {{ $('Writer Agent').item.json.output.subject }}", "operation": "read" }, "position": [ 1664, 272 ] }, { "id": "31b4b7f1-fcf3-4ca1-867a-4ed6cb81465a", "name": "Build Email HTML", "type": "code", "params": { "jsCode": "// ============================================================\n// BUILD EMAIL HTML — Synecta-branded template\n// ============================================================\n// ⚙️ CONFIGURE THESE VALUES BEFORE USING:\nconst YOUR_NAME = 'Stefan Joulien – Synecta'; // ⚙️ Change to your name\nconst YOUR_EMAIL = 'user@example.com'; // ⚙️ Change to your email\nconst CAL_URL = 'https://cal.eu/stefan-joulien-synecta/discovery'; // ⚙️ Change to your booking link\nconst INSTAGRAM_URL = 'https://www.instagram.com/stefanjd.fit/'; // ⚙️ Change to your Instagram\nconst LINKEDIN_URL = 'https://www.linkedin.com/in/stefan-joulien-ditchev/'; // ⚙️ Change to your LinkedIn\nconst LOGO_URL = 'https://res.cloudinary.com/dmdmat24m/image/upload/v1773779124/Synecta_Logo_Horizontal_Alargado_v1sw74.png'; // ⚙️ Change to your logo URL\n// ============================================================\n\nconst RAW = ($input.first().json.output && $input.first().json.output.message_html) || '';\n\n// --- Synecta Brand Tokens ---\nconst C_PRIMARY = '#6C63FF'; // Synecta violet\nconst C_DARK_BLUE = '#1C5D99'; // Synecta dark blue\nconst C_LIGHT = '#AEB8FE'; // Synecta light violet\nconst C_BG = '#F5F7FA'; // Off-white background\nconst C_DARK_BG = '#1A1A1A'; // Footer dark\nconst FONT = \"'Segoe UI', Tahoma, Geneva, Verdana, sans-serif\";\n\n// CTA button: gradient from violet to dark blue\nconst CTA_INLINE =\n `display:inline-block;` +\n `background:linear-gradient(90deg, ${C_PRIMARY} 0%, ${C_DARK_BLUE} 100%);` +\n `color:#fff;text-decoration:none;padding:16px 40px;border-radius:14px;` +\n `font-weight:700;font-size:15px;box-shadow:0 10px 30px rgba(108,99,255,0.35);` +\n `text-align:center;min-width:220px;letter-spacing:.5px;text-transform:uppercase;`;\n\n// Avoid reprocessing\nif (//i.test(RAW)) {\n return [{ email_html: RAW }];\n}\n\n// 1) Strip any inline styles on tags, reapply brand colour\nlet src = RAW.replace(/]*?)style=\"[^\"]*?\"([^>]*)>/gi, '');\nsrc = src.replace(/]*)>/gi,\n ``\n);\n\n// 2) Convert

into numbered card layout\nlet idx = 1;\nfunction liToCard(n, innerHtml) {\n const m = innerHtml.match(/]*>\\s*([\\s\\S]*?)<\\/strong>\\s*<\\/p>/i);\n const title = (m ? m[1] : `Improvement ${n}`).trim();\n const body = (m ? innerHtml.replace(m[0], '') : innerHtml).trim();\n return `\n \n \n \n \n
\n \n \n \n \n \n
\n
\n ${n}\n
\n \n
${title}
\n
${body}
\n
\n
`;\n}\n\nsrc = src.replace(//gi, (ol) => {\n const items = [...ol.matchAll(/]*>([\\s\\S]*?)<\\/li>/gi)].map(m => m[1]);\n if (!items.length) return ol;\n return items.map(li => liToCard(idx++, li)).join('');\n});\n\n// 3) Assemble full email-safe HTML\nconst html = `\n\n\n\nBusiness Analysis Report\n\n\n\n\n \n \n \n
\n\n \n \n\n \n \n \n \n\n \n \n \n \n\n \n \n \n \n\n
\n $\"Logo\"$ \n
\n
\n ${src}\n
\n
\n Book a Free Call\n
\n\n \n \n \n \n \n
\n\n \n $\"Synecta\"$ \n\n \n \n \n \n \n \n \n \n \n \n
\n \n $\"Email\"\n$ \n \n \n \n $\"Instagram\"\n$ \n \n \n \n $\"LinkedIn\"\n$ \n \n
\n\n
\n This analysis was prepared specifically for your business.
\n Questions? Reply directly to this email.\n
\n
\n\n
\n`;\n\nreturn [{ email_html: html }];\n" }, "position": [ 1216, 272 ] }, { "id": "94878d17-f8a3-4e03-8023-373ec6fb9d1b", "name": "Clean HTML", "type": "code", "params": { "jsCode": "const html = $input.first().json.data || '';\nconst cleaned = html\n .replace(//gi,' ')\n .replace(//gi,' ')\n .replace(//g,' ')\n .replace(/<[^>]+>/g,' ')\n .replace(/\\s+/g,' ')\n .trim();\n\nconst text = cleaned.slice(0, 28000); // hard limit ~6-8k tokens\nconst signals = {\n url: $json.url || '',\n page_title: $json.pageTitle || '',\n has_wa: /wa\\.me|api\\.whatsapp\\.com/i.test(html),\n has_tel: /tel:/i.test(html),\n has_mailto: /mailto:/i.test(html),\n has_checkout: /checkout|cart|add to cart|woocommerce/i.test(html),\n has_booking: /calendly|book now|appointment|booking|schedule/i.test(html),\n social: [...new Set((html.match(/instagram\\.com|linkedin\\.com|facebook\\.com|youtube\\.com/gi) || []).map(s => s.toLowerCase()))]\n};\n\nreturn [{ url: signals.url, pageTitle: signals.page_title, text, signals }];\n" }, "position": [ -208, 272 ] }, { "id": "95e76890-12ed-4d20-9b3b-a9c2f11819f3", "name": "Fetch Website HTML", "type": "http", "params": { "url": "={{ $('Form Trigger').item.json.website }}", "options": {} }, "position": [ -384, 272 ] } ] }

Send AI website audits with GPT-4.1 and Gmail as a lead magnet