Workflow Templates

8 nodes

Recover overdue invoices with GPT-4o-mini, Gmail, Google Sheets and Slack

by Rahul Joshi

Quick overview This workflow runs every 24 hours to read overdue invoices from Google Sheets, uses OpenAI (GPT-4o-mini) to draft personalized recovery emails and determine escalation, sends emails via Gmail, logs outcomes back to Google Sheets, and posts status and error alerts to Slack. How it works Runs every 24 hours on a schedule trigger. Reads invoice rows from a Google Sheets “invoices” tab and checks whether there are any valid invoice entries to process. If no invoices are found, posts a “no overdue invoices” message to a Slack channel. For each invoice found, uses OpenAI (GPT-4o-mini) to assign an escalation level, draft a personalized HTML email (including payment plans when applicable), and recommend an escalation action. Parses the AI output into structured fields and sends the recovery email to the client via Gmail. If the AI recommends escalation, sends an internal escalation alert email to the legal team via Gmail. Updates the invoice row in Google Sheets with the new escalation level and AI summary, then appends a full audit log row to a “Recovery sheet” tab. If AI processing fails, posts an error alert to Slack and continues with the next invoice. Setup Add Google Sheets OAuth2 credentials and set the Spreadsheet ID plus the sheet names (“invoices” and “Recovery sheet”). Add an OpenAI API credential and ensure the GPT-4o-mini model is available for the AI agent. Add Gmail OAuth2 credentials and set the reply-to address and recipients (client email field and the internal legal email address). Add Slack OAuth2 credentials and set the target channel ID for both the “no invoices” and error alerts. Ensure your invoices sheet includes columns used by the workflow (Invoice, Client, Email, Amount, Due Date, Days Overdue, Reminders Sent, and row_number for updates).

6 nodes19 views

Schedule and manage appointments with Groq, Google Calendar, Gmail and Sheets

by BytezTech

AI appointment bot with Google Calendar, Gmail and Sheets 📌 Overview This workflow automates end-to-end appointment scheduling for your business using an AI-powered chatbot. Clients can book, reschedule, or cancel meetings through a simple chat interface — no manual coordination needed. Two independent flows run in parallel. The AI Chat Flow handles real-time client conversations using Groq LLaMA 4 Scout, managing calendar availability, event creation, updates, and Gmail confirmations automatically. The Daily Sync Flow runs every morning to refresh your Google Sheet with the day's appointments and email an admin summary report. All times are handled in IST (GMT+05:30) with full ISO 8601 formatting. ⚙️ How it works AI Chat Flow Client sends a message to the bot AI Agent collects name, email, purpose, date, time, and duration Checks Google Calendar for availability Creates, updates, or deletes calendar events based on request Sends Gmail confirmation or cancellation email to the attendee Daily Sync Flow Triggers every morning at 9:30 AM IST Clears previous rows from Google Sheets (keeps header) Fetches all of today's Google Calendar events Formats and inserts each event into the sheet Emails the admin a summary report with total meeting count and sheet link 🛠️ Setup steps Import this workflow into n8n Add a Google Calendar OAuth2 credential and set your calendar email Add a Gmail OAuth2 credential for the agent and admin report Add Google Sheets OAuth2 (for Clear node) and Service Account (for Insert node) Add your Groq API key for the LLaMA 4 Scout model Update your admin email, calendar email, and Google Sheet ID in the workflow Activate the workflow — both flows run independently 🚀 Features AI-powered scheduling Books, reschedules, and cancels appointments via chat Checks real-time calendar availability before confirming Suggests 3 alternative slots if the requested time is unavailable Sends automatic Gmail confirmations to attendees Maintains conversation memory across the session (last 50 messages) Daily reporting Auto-clears and refreshes the Google Sheet every morning Syncs all calendar events with full details (name, email, time, duration) Emails admin a daily count summary with a direct sheet link 📋 Requirements n8n (cloud or self-hosted) Google Calendar, Gmail, and Google Sheets access Groq API key (free tier supported) 🎯 Benefits Zero manual appointment coordination Clients self-serve via chat 24/7 Admin always has a fresh daily schedule in Google Sheets Automatic email confirmations build client trust IST timezone enforced consistently across all events 👨‍💻 Author BytezTech Pvt Ltd

Track AI search visibility across ChatGPT, Claude, DeepSeek and Perplexity

by Incrementors

Description This workflow automates AI Search Engine Optimization (ASEO) tracking for digital marketing agencies. It tests your client's website visibility across four major AI platforms—ChatGPT, Claude, DeepSeek, and Perplexity—using brand-neutral prompts, analyzes ranking position and presence strength on each platform, identifies top competitors, and returns a structured 27-field scorecard with actionable recommendations. Designed as a sub-workflow, it integrates directly into your existing client audit or reporting pipeline. Key Features Brand-neutral prompt generation (no client name used—tests true organic AI discoverability) Simultaneous visibility testing across ChatGPT, Claude, DeepSeek, and Perplexity Presence strength scoring (0–100%) per platform Competitor identification across all four AI platforms Strongest and weakest platform detection AI-generated actionable recommendations for improvement Structured 27-field output ready for Google Sheets or database insertion Error handling on all agent nodes (partial results if any platform fails) Sub-workflow design—integrates cleanly into larger audit pipelines What This Workflow Does Input This workflow is triggered by a parent workflow and receives two parameters: Website**: The client's website URL (e.g., https://example.com) Website Summary**: A plain-text description of what the business does and its core services Processing Stage 1 — Brand-Neutral Prompt Generation GPT-4.1-mini generates a realistic search prompt that potential customers would type into an AI chatbot to find a company like the client. Critically, the prompt does not include the client's brand name—it focuses on their services and industry instead. For example, for a Los Angeles product photography studio, the prompt would be something like "best product photography studio for Amazon listings in Los Angeles" rather than the studio's name. This tests true organic discoverability, not brand recall. Stage 2 — Four-Platform Sequential Testing The same generated prompt is submitted sequentially to four AI platforms: ChatGPT via GPT-4o-mini Claude via Claude Sonnet 3.7 DeepSeek Perplexity Each platform agent runs independently with error handling enabled. If one platform API is down or throws an error, the workflow continues and returns partial results—it does not fail entirely. Stage 3 — Cross-Platform Analysis DeepSeek analyzes all four platform outputs together and produces a structured JSON report covering each platform's ranking (Yes/No), position (1–10 or null), presence strength percentage, key mentions, and top competitors. It also generates an overall summary comparing all platforms. Stage 4 — Data Flattening The nested JSON is flattened into 27 individual fields that can be directly inserted into a Google Sheet row, database, or passed back to the parent workflow for reporting. Output The workflow returns 27 structured data fields: Search prompt used (1 field) Per-platform metrics for ChatGPT, Claude, DeepSeek, Perplexity: Ranking (Yes/No), Position, Presence Strength %, Key Mentions, Top Competitors (5 fields × 4 platforms = 20 fields) Overall summary: Total platforms ranking, Average presence strength, Strongest platform, Weakest platform, Main competitors across all platforms, Recommendations (6 fields) Setup Instructions Prerequisites Active n8n instance (self-hosted or n8n Cloud) Parent workflow with Execute Workflow node (this workflow does not run standalone) OpenAI API key (used for prompt generation and ChatGPT testing) Anthropic API key (used for Claude testing) DeepSeek API key (used for DeepSeek testing and final analysis) Perplexity API key (used for Perplexity testing) Estimated setup time: 20–25 minutes Step 1: Understand how this workflow is triggered This is a sub-workflow. It does not have its own schedule trigger. It runs when a parent workflow calls it using n8n's Execute Workflow node. Setting up the parent workflow: Open or create your parent workflow (e.g., a client audit scheduler, a Google Sheets loop, or a manual trigger) Add an Execute Workflow node to your parent workflow Inside the Execute Workflow node: Source: Select "Database" Workflow: Search for and select this AI Search Ranking Analyzer workflow Mode: Choose "Run once for all items" or "Run once for each item" depending on your setup Under Fields, add two parameters to pass: Name: Website | Value: your client's website URL expression (e.g., ={{ $json['Website URL'] }}) Name: Website Summary | Value: your client's business description (e.g., ={{ $json['Business Description'] }}) Example parent workflow structure: Schedule Trigger (Weekly / Monthly) → Read Client List from Google Sheets → Loop Over Each Client → Execute Workflow (this AI Search Ranking Analyzer) Pass: Website = {{ $json['Website URL'] }} Pass: Website Summary = {{ $json['Summary'] }} → Append 27 Fields to Reporting Sheet → Send Report Email or Slack Notification Testing the trigger connection: Open this workflow and click on the Receive Website and Summary from Parent node You will see "Waiting for input from parent workflow..." Go to your parent workflow and click Execute node on the Execute Workflow node The data will flow into this workflow for testing Both workflows must be set to Active for production use Step 2: Connect OpenAI credentials This workflow uses two OpenAI models: GPT-4.1-mini** — used by Generate Brand-Neutral Search Prompts, Parse Prompt as JSON, and GPT Model for Parser Support GPT-4o-mini** — used by Test Visibility on ChatGPT To connect: In n8n go to Credentials → Add credential → OpenAI API Paste your API key from https://platform.openai.com/api-keys Name it clearly (e.g., "OpenAI Main") Open each of these nodes and select your credential: GPT Model for Prompt Generation → select your OpenAI credential, set model to gpt-4.1-mini GPT Model for Parser Support → select your OpenAI credential, set model to gpt-4.1-mini GPT-4o-mini for ChatGPT Test → select your OpenAI credential, set model to gpt-4o-mini Step 3: Connect Anthropic credentials Used by the Test Visibility on Claude agent via Claude Sonnet 3.7 Model. To connect: Go to Credentials → Add credential → Anthropic API Get API key from https://console.anthropic.com/ Open the Claude Sonnet 3.7 Model node and select your credential Verify the model is set to claude-3-7-sonnet-20250219 Step 4: Connect DeepSeek credentials Used by two nodes: DeepSeek Model for Testing (platform test) and DeepSeek Model for Analysis (final summarizer). To connect: Go to Credentials → Add credential → DeepSeek API Get API key from https://platform.deepseek.com/ Open DeepSeek Model for Testing node → select your credential Open DeepSeek Model for Analysis node → select your credential Step 5: Connect Perplexity credentials Used by the Test Visibility on Perplexity node (Perplexity native node, not an AI agent). To connect: Go to Credentials → Add credential → Perplexity API Get API key from https://www.perplexity.ai/settings/api Open the Test Visibility on Perplexity node and select your credential Step 6: Test the complete workflow Temporarily add a Manual Trigger node at the start and connect it to Generate Brand-Neutral Search Prompts (bypass the executeWorkflowTrigger for isolated testing) Set the Manual Trigger to pass test data: { "Website": "https://your-test-site.com", "Website Summary": "A company that provides [your service] in [your city]" } Execute and verify: Generate Brand-Neutral Search Prompts produces a sensible search query Each platform node returns output (or gracefully continues on error) Analyze All Platform Results produces structured JSON Flatten JSON to 27 Data Fields produces all 27 fields correctly Remove the test Manual Trigger once testing is complete Activate this workflow and your parent workflow Workflow Node Breakdown Receive Website and Summary from Parent The entry point of this sub-workflow. Listens for execution from a parent workflow via n8n's Execute Workflow node. Receives two inputs: Website (client URL) and Website Summary (business description text). These values are referenced by subsequent nodes throughout the workflow. Generate Brand-Neutral Search Prompts An AI agent powered by GPT-4.1-mini that creates a realistic search query a potential customer might type into an AI chatbot to find a business like the client—without using the client's brand name. This tests organic discoverability based on services and industry positioning rather than brand recognition. The output is a single focused search prompt. Parse Prompt as JSON A Structured Output Parser that enforces JSON schema {"Prompts": "..."} on the generated prompt. Uses GPT Model for Parser Support as its language model and has autoFix enabled, so malformed outputs are automatically retried and corrected. Test Visibility on ChatGPT An AI agent that submits the generated search prompt to ChatGPT (GPT-4o-mini) and records the response. This captures what ChatGPT currently recommends when users search for services like the client's. Test Visibility on Claude An AI agent powered by Claude Sonnet 3.7 (Anthropic) that receives the same prompt and records Claude's recommendations. Has onError: continueRegularOutput so the workflow continues if Claude's API is unavailable. Test Visibility on DeepSeek An AI agent powered by DeepSeek that tests the same prompt on DeepSeek's platform. Also has onError: continueRegularOutput for resilience. Test Visibility on Perplexity Uses n8n's native Perplexity node (not an AI agent) to submit the prompt to Perplexity's search-augmented AI. Perplexity is particularly important because it uses real-time web search, making its recommendations highly relevant for current visibility. Has onError: continueRegularOutput. Analyze All Platform Results A DeepSeek-powered AI agent that receives all four platform outputs simultaneously along with the client website URL and the original search prompt. It analyzes each platform independently—determining whether the client appears (Yes/No), at what position (1–10), how strongly (0–100%), how they are mentioned, and which competitors appear. It also generates an overall summary comparing all platforms and provides specific improvement recommendations. Uses Parse Analysis as Structured JSON as its output parser. Flatten JSON to 27 Data Fields A Set node that extracts values from the nested JSON output of the analyzer into 27 flat fields. This makes the data ready for direct insertion into a Google Sheets row, Airtable record, or database table—or for return to the parent workflow. Output Data Complete A No Operation node marking the successful completion of the workflow. The parent workflow receives all 27 fields as the execution output. Usage Guide Adding clients for analysis In your parent workflow, maintain a Google Sheet with columns: | Client Name | Website URL | Business Description | Last Checked | |---|---|---|---| | Example Corp | https://example.com | A SaaS company that provides... | 2025-01-15 | Your parent workflow reads each row, passes the Website URL and Business Description to this sub-workflow, and writes the 27 returned fields back into the sheet for tracking. Understanding the output After execution, check the Flatten JSON to 27 Data Fields node output. For each platform you get: Ranking:** Yes (client appears) or No (client not mentioned) Position:** Numeric position in the AI's recommendations (1 being top) Presence Strength:** 0–100% measuring how positively and prominently the client is featured Key Mentions:** How the AI described or mentioned the client Ranking Competitors:** Which competitors the AI recommended instead The Overall Summary tells you: How many of 4 platforms are currently ranking your client The average presence strength across all platforms Which platform is your strongest opportunity Which platform needs the most improvement The 3 main competitors appearing consistently Specific recommendations for improving AI discoverability Tracking over time Run this workflow monthly per client. Append results to a Google Sheet with a date column. Track whether presence strength is improving, whether the client appears on more platforms over time, and whether competitors are losing or gaining ground. Customization Options Change the number of platforms: Remove any platform agent node and update the Analyze All Platform Results prompt to exclude that platform's output reference. Add more platforms: Add new AI agent nodes (e.g., Grok, Gemini) between Test Visibility on Perplexity and Analyze All Platform Results. Update the analyzer prompt to include the new platform's output. Generate multiple prompts: Modify Generate Brand-Neutral Search Prompts to produce 3–5 different prompts. Loop through each and aggregate results for more comprehensive testing. Write results directly to Google Sheets: After Flatten JSON to 27 Data Fields, add a Google Sheets Append node in your parent workflow to log each audit automatically. Add email or Slack notifications: After the workflow completes in the parent, add a Send Email or Slack node that formats the key metrics (Overall Ranking, Average Presence Strength, Recommendations) into a readable client report. Adjust presence strength scoring: Modify the Analyze All Platform Results prompt to change how the AI scores presence strength—for example, weighting first-position mentions more heavily. Troubleshooting Parent workflow not triggering this workflow Verify both workflows are toggled to Active In the Execute Workflow node, confirm the correct workflow is selected Check that the Mode is set (not left blank) Test by clicking Execute node directly on the Execute Workflow node in the parent Website and Website Summary parameters not passing In the Execute Workflow node, confirm the field names are exactly Website and Website Summary (case-sensitive, space in second parameter) Check the parent workflow is actually passing values, not empty expressions Use the Receive Website and Summary from Parent node's input panel to verify received data One platform returns empty output The workflow continues even if one platform fails (onError: continueRegularOutput is set) Check the specific platform node for the error message Verify API credentials are valid and have available credits Perplexity free tier has strict rate limits—upgrade plan if hitting limits Structured output parser fails Parse Prompt as JSON has autoFix enabled—it will retry malformed outputs automatically If Parse Analysis as Structured JSON fails, simplify the prompt in Analyze All Platform Results or increase max tokens Check that DeepSeek credentials are active (DeepSeek handles the analysis output parsing) Generated prompt includes client brand name The Generate Brand-Neutral Search Prompts agent prompt instructs GPT to avoid brand names If brand names slip through, add to the system prompt: "Never mention any specific company name, brand, or trademark in the generated prompt" All 27 fields not appearing in output Run the workflow with test data and inspect Analyze All Platform Results node output If a platform returned empty output due to an error, its fields will be null Check that Flatten JSON to 27 Data Fields expressions reference the correct node names Use Cases Digital marketing agencies offering ASEO services: Run monthly AI visibility audits for 20–50 clients from one parent workflow. Generate client reports showing AI platform rankings, presence strength trends, and competitor comparisons. Position ASEO as a premium new service. SEO teams expanding beyond Google: Use this alongside traditional Google ranking reports. Show clients their full search visibility picture—covering both Google and the AI chatbots that are increasingly influencing purchase decisions. Competitive intelligence: Run this workflow for your own site and 3–5 competitors simultaneously. Identify which competitors dominate AI recommendations and reverse-engineer their content strategy. Brand monitoring: Track how AI chatbots describe your brand over time. Detect if competitors are gaining ground or if negative associations are appearing in AI responses. New market entry research: Before entering a new market or launching a new service line, test whether your website would appear in AI searches for that service category. Use results to guide content strategy before launch. Expected Results Time savings: 45–60 minutes of manual AI testing per client, eliminated per audit cycle Coverage: 4 major AI platforms tested in a single automated run Output quality: Structured, consistent 27-field data format—ready for Google Sheets, dashboards, or PDF reports Scalability: Process 50+ clients per parent workflow run with no additional manual effort Competitive advantage: One of the first systematic approaches to measuring AI Search Engine Optimization (ASEO)—a space with no established tooling yet For any questions, custom development, or workflow integration support: 📧 Email: info@incrementors.com 🌐 Website: https://www.incrementors.com/

6 nodes

Detect human vs AI text using stylometric metrics and multi‑agent LLM debate

by Mychel Garzon

Stop guessing if text came from ChatGPT. Let three AI agents argue about it using forensic data. Paste any text and get a verdict on whether it was written by a human, AI, or a hybrid mix. Instead of trusting one black-box score, this workflow runs your text through statistical analysis and a three-agent debate where each agent challenges the others using hard numbers. This is not another "detect AI with AI" template. The workflow measures six forensic markers first, then makes three separate agents argue about what those numbers mean. You see the raw data, the debate, and the final verdict with confidence scores. How it works The workflow runs in five stages: Extract forensic metrics: A code node measures burstiness (sentence length variation), type-token ratio (vocabulary diversity), hapax rate (words appearing once), repetition score (repeated phrases), transition density (filler words like "furthermore"), and AI fingerprints (100+ known LLM phrases stored in a data table). Short texts under 150 words get recalibrated because metrics are less reliable. Agent 1 - The Scanner: Reads the text cold with zero metrics. Gives a gut impression (human/AI/hybrid) based purely on instinct. Acts like an editor who has read thousands of manuscripts. Agent 2 - Forensic Analyst: Gets the text, all metrics, and Agent 1's verdict. Writes a data-driven report that must cite specific numbers. Either agrees or disagrees with Agent 1 and explains why using the forensic evidence. Agent 3 - Devil's Advocate: Gets everything above and argues the opposite of whatever Agent 2 concluded. If Agent 2 said AI, Agent 3 must argue human. Finds holes in the logic and metrics that got ignored. Weighted verdict: A code node scores all three agents (35% Analyst, 15% Scanner, 15% Devil's Advocate, 35% raw metrics) and classifies as human (score under 0.35), AI (score over 0.60), or AI-augmented (in between). Confidence is calculated separately so you get verdicts like "AI with 67% confidence." Chat output format The chat response shows: Verdict badge:** 🙎🏻 Human-Written, 🤖 AI-Generated, or 🦾 AI-Augmented Confidence bar:** Visual bar (██████████ 85%) showing how certain the verdict is Metrics table:** All six forensic markers with 🟥 AI or 🟩 Human flags Agent debate:** Three verdicts with reasoning. Agent 1's gut check, Agent 2's forensic report, Agent 3's counter-argument. Each shows classification and confidence percentage. Example output for AI text: 🤖 Verdict: AI-Generated Confidence: ████████░░ 87% 📊 Stylometric Metrics: Burstiness: 0.18 🟥 AI Vocabulary Diversity: 0.36 🟥 AI Hapax Rate: 0.32 🟥 AI Repetition: 0.21 🟥 AI Transition Density: 0.024 🟥 AI 🔎 Agent 1 (Gut Check): AI (90%) "Monotonous rhythm, corporate vocabulary, zero personality" 🔬 Agent 2 (Data): AI (95%) "Five of six metrics flag AI. Burstiness of 0.18 well below human threshold..." 😈 Agent 3 (Critic): AI-AUGMENTED (65%) "Could be human technical writing. Transition density alone not conclusive..." Self-updating fingerprint database A separate workflow branch runs weekly to keep the AI phrase list current: Check existing words: Reads all fingerprint phrases from the data table Find new AI tells: Asks an LLM what phrases modern models currently overuse Filter duplicates: Removes words already in the database Add to table: Stores new phrases for future detection Requires: A data table (Google Sheets, Airtable, or n8n Data Table) to store fingerprint words. The workflow includes a starter list of 100+ phrases like "delve into," "it's worth noting," "as of my last update." LLM writing patterns shift fast. What worked for GPT-3 detection does not work for GPT-4. This keeps the detector current without manual updates. Key benefits Three classifications instead of binary.** Human, AI, or AI-augmented. Most real content is hybrid. You see the reasoning.** Full agent debate included. When verdicts are borderline, you can read which argument won. Transparent metrics.** Raw numbers exposed with red/green flags. No hidden scoring. Self-updating detection.** Weekly workflow finds new AI phrase patterns as models evolve. Error resilient.** If one agent fails, the workflow continues and redistributes weights. Who this is for Content teams verifying contractor submissions are not AI-generated Educators checking student essays for AI assistance Publishers screening submissions to maintain editorial standards SEO teams ensuring content meets Google's helpful content guidelines Researchers analyzing hybrid human-AI writing patterns Setup Add API credentials for at least one LLM provider (Groq, OpenAI, Gemini, or Anthropic) Create a data table for AI fingerprint phrases or use n8n's built-in Data Table node Populate the table with the starter list (included in workflow documentation) Activate the workflow and open the chat interface Paste text and wait 30-60 seconds for forensic analysis Required APIs & credentials At least one LLM provider: OpenAI, Anthropic, Google Gemini, Groq, or any other provider with JSON output support. Each agent can use a different provider or all can use the same one. Data storage for fingerprint phrases: n8n Data Table (built-in), Google Sheets, or Airtable. The workflow checks this table to identify known AI phrases during analysis. How to customise it Swap models:** Each agent node has a chat model sub-node. Replace with any provider. Scanner works with smaller models. Analyst needs strong reasoning. Devil's Advocate needs good instruction-following. Tune thresholds:** Open Extract Stylometric Metrics code. Burstiness under 0.3 flags AI. Type-token ratio under 0.4 flags AI. Adjust for stricter or looser detection. Change agent weights:** Open Final Verdict code. Default is 35% Analyst, 15% Scanner, 15% Devil's Advocate, 35% metrics. Increase metric weight to trust data more. Modify agent personas:** Edit system prompts. Make Scanner more skeptical. Make Analyst cite sources. Make Devil's Advocate more aggressive. Add quality gate:** Drop a Filter node after verdict. Only proceed if confidence exceeds 70%. Batch process:** Replace Chat Trigger with Schedule Trigger looping over a file list. Known limitations The workflow works best on long-form content (500+ words). Short texts under 100 words produce less reliable metrics because statistical patterns need more data to emerge. The recalibration helps but is not perfect. AI fingerprint phrases evolve as models improve. GPT-5 might not use "delve into" but will have new tells. The self-updating workflow helps but lags current releases by a few weeks. The three-agent debate architecture assumes disagreement is meaningful. For extremely niche topics where only one agent has relevant training data, the minority opinion might be correct but gets outvoted. Review the individual agent reasoning when dealing with specialized content.

Generate daily stock BUY/HOLD/SELL signals with GPT-4o, Gemini, FMP, Sheets and Telegram

by Mo AlBarrak

Overview This is a production-grade, fully automated stock analysis system built entirely in n8n. It combines institutional-level financial analysis, dual AI model consensus, and a self-improving backtesting loop — all running on autopilot, every single day. Every morning, the engine screens the US stock market, collects deep financial data, reads the latest news, and sends two independent AI analysts (GPT-4o and Gemini 2.5 Pro) to debate each stock. When they disagree, a structured bull-vs-bear tiebreaker is triggered. The result: a daily ranked list of BUY, HOLD, and SELL signals — with price targets, confidence scores, and risk assessments — delivered straight to your Telegram. A companion backtesting workflow runs silently in the background, grading every past signal 7 days after it was issued and sending you a weekly performance report every Monday morning. This is not a toy workflow. This is the kind of system that would cost thousands of dollars to build as a SaaS — running entirely on your own infrastructure. ✨ What Makes This Template Unique 🤖 Dual AI Consensus Engine — GPT-4o and Gemini 2.5 Pro analyze every stock independently. Their outputs are compared, and consensus is only declared when both models agree within a tight price target band ⚖️ Structured Tiebreaker Architecture — When models disagree, a bull analyst (GPT-4o) and a bear analyst (Gemini) re-run with opposing mandates. The final verdict is derived from their averaged price target plus a Piotroski F-Score gate 📊 Institutional-Grade Financial Modeling — Piotroski F-Score (9-point), Graham Number intrinsic value, DCF anchor, TTM revenue & margins, net debt, FCF, revenue growth YoY, and sector-relative P/E valuation — all computed automatically 📰 Live News Sentiment — Latest headlines per stock are fed into the AI prompt, adjusting confidence scores in real time based on positive or negative sentiment signals 🎯 Scenario Price Targets — Every stock gets three targets: pt_bear (downside), pt_base (fair value), pt_bull (upside case), giving you a full risk/reward picture 🔁 Self-Improving Backtester — Every signal is automatically graded 7 days later. Win rate, average return, and best/worst calls are reported every Monday via Telegram 📡 Smart Screener with Sector Diversity — Scores 100 candidates daily using volume health, market cap sweet spot ($5B–$100B), and beta gradient — with a sector diversity cap so you never end up with 15 tech stocks 💾 Full Google Sheets Audit Trail — Every signal, confidence score, rationale, and outcome is logged permanently for your own review and analysis 📋 Workflow Breakdown Workflow 1 — AI Institutional Stock Valuation Engine Phase What Happens Phase 1 — Screening FMP screener fetches 100 US stocks. Score_and_Prefilter scores and selects the top 20 with sector diversity Phase 2A — Financial Data 13 FMP endpoints per stock: income, balance sheet, cash flow, ratios, profile, sector P/E Phase 2B — News Latest headlines fetched and passed into AI context Phase 3 — AI Round 1 GPT-4o and Gemini 2.5 Pro analyze in parallel. Verdicts and price targets compared Phase 3 — Tiebreaker Bull vs Bear re-analysis when models disagree or price target gap > 25% Phase 4 — Strong Buy Alert Stocks with BUY verdict + upside ≥ 20% + confidence ≥ 65 trigger an immediate alert Phase 5 — Storage & Summary All results written to Google Sheets. Daily Telegram summary sent with top picks Workflow 2 — Signal Outcome Checker & Weekly Backtester Trigger What Happens Daily 8AM Finds signals that are 7 days old, fetches current price, grades WIN / LOSS / NEUTRAL, writes outcome back to sheet Monday 9AM Computes weekly win rate, average return on BUY signals, best and worst call — sends full report to Telegram 🛠️ What You Need Requirement Details FMP API Key Financial Modeling Prep — Starter plan or above (~$25/mo). Covers all financial data, screener, news, and historical prices OpenAI API Key GPT-4o access via API or ChatGPT Plus Google Gemini API Key Gemini 2.5 Pro via Google AI Studio (free tier available) Google Sheets One sheet named Stock_Signals with the column headers listed in the setup guide Telegram Bot Create via @BotFather in 2 minutes. Free n8n Self-hosted or n8n Cloud Estimated running cost: $0.43/day in AI tokens for 20 stocks ($10–$13/month). FMP and Telegram are the only other costs. ⚙️ Setup Time ~30–45 minutes for a first-time setup. All credentials, Sheet IDs, and API keys are clearly labeled in each node. No coding required — every parameter is documented. 📈 Example Daily Telegram Output 📊 Daily Valuation Report — 2026-04-02 Stocks Analyzed: 20 🟢 BUY: 7 🟡 HOLD: 10 🔴 SELL: 3 🚨 STRONG BUY ALERTS: • NVDA — Upside 34% | Confidence 81 | F-Score 7/9 • MSFT — Upside 22% | Confidence 74 | F-Score 8/9 Top Picks: NVDA pt_base $172 | pt_bull $198 | pt_bear $124 MSFT pt_base $485 | pt_bull $530 | pt_bear $410 AMGN pt_base $318 | pt_bull $355 | pt_bear $275 📊 Example Weekly Backtest Report 📈 Weekly Signal Performance — Week of Mar 31 Signals Graded: 18 ✅ Win Rate: 72% | BUY Accuracy: 78% 📈 Avg Return on BUY signals: +4.3% 🏆 Best Call: NVDA +11.2% (BUY ✅) 💔 Worst Call: BA -6.8% (BUY ❌) 💡 Who Is This For? Retail investors who want institutional-quality analysis without paying for a Bloomberg terminal Quantitative traders looking for a customizable, data-driven signal generation pipeline n8n builders who want to see a real-world, production-grade multi-node workflow in action AI enthusiasts interested in multi-model consensus systems and structured debate architectures 📬 Questions, Customizations & Feedback Have a question about setup, want to adapt this workflow to your own strategy, or found something to improve? 📧 mambarrak@gmail.com All feedback is welcome. If you build something interesting on top of this, I'd love to hear about it. Built with ❤️ using n8n, Financial Modeling Prep, OpenAI GPT-4o, and Google Gemini 2.5 Pro.

6 nodes

Send weekly narrative client reports with Google Sheets, Claude, and Gmail

by Monfort N. Brian | 宁俊

Quick overview This workflow runs every Friday at 9:00 AM, pulls active client status data from Google Sheets, uses Anthropic Claude to generate a narrative weekly report with structured HTML and plain text output, and sends the finished report to each client via Gmail. How it works Runs every Friday at 9:00 AM on a schedule trigger. Reads all client rows from a specified Google Sheets document. Filters out paused/inactive or incomplete rows, extracts any “Metric:” columns, and builds a reporting week label for each client. Sends each client’s normalized data to Anthropic Claude to write a 3–4 paragraph narrative report and return JSON containing an email subject plus HTML and plain-text bodies. Emails the generated HTML report to each client’s email address using Gmail. Setup Create a Google Sheet with the required columns (at minimum “Client Name” and “Client Email”) and optional “Status” plus any “Metric:” columns you want included. Add Google Sheets OAuth2 credentials in n8n and replace the placeholder Sheet ID in the Google Sheets read step. Add an Anthropic credential for the Claude chat model used to generate the structured report. Add Gmail OAuth2 credentials and confirm the “send to” field maps to your sheet’s “Client Email” values (and optionally adjust the prompt tone/system message).

6 nodes1 views

Track competitor prices and email AI insights with Google Sheets, Groq and Gmail

by Avkash Kakdiya

How it works This workflow automatically monitors competitor product prices stored in Google Sheets. It scrapes product pages, extracts pricing and offer data using AI, and compares it with historical values. Based on changes, it updates records and generates a market intelligence report. The workflow then emails the report and resets data for the next execution cycle. Step-by-step Step 1: Database sync** Schedule Trigger – Runs the workflow at a scheduled time. Get row(s) in sheet – Fetches competitor data and product URLs. Step 2: Scraping** Loop Over Items – Processes each competitor entry. HTTP Request3 – Retrieves raw HTML using ScraperAPI. Clean Content – Cleans and prepares text for AI processing. Step 3: Price extraction** AI Agent1 – Extracts product name, price, and offers. Groq Chat Model1 – Provides AI extraction capability. current Price and offer – Converts AI output into structured data. If2 – Checks if it's the first recorded entry. First time price and offer added – Stores initial values. If1 – Compares current vs previous price and offers. Updated current price and offer in sheet – Updates if changes detected. If No changes then update – Updates sheet even when no change is found. Step 4: Analysis** Get row(s) in sheet1 – Retrieves updated dataset. Data Aggregator – Builds structured market comparison data. AI Agent – Generates strategic insights and recommendations. Groq Chat Model – Powers the analysis output. Update row in sheet – Saves AI-generated summary in sheet. Step 5: Reporting** Edit Fields1 – Formats the report into HTML email layout. Send a message – Sends the final report via Gmail. Step 6: Reset** Get row(s) in sheet2 – Retrieves final processed data. Update row in sheet1 – Moves current data to history and clears fields. Why use this? Ensures all price scenarios (change or no change) are handled properly Keeps your Google Sheets always updated with accurate data Provides AI-powered competitive intelligence automatically Sends clean, formatted reports without manual effort Maintains structured historical tracking for better decision-making

Discover SaaS creators from HubSpot with influencers.club and Gmail/SendGrid

by Influencers Club

How it works: Get multi social platform data for SaaS clients with their email and send personalized comms to onboard them as organic creators, partners and ambassadors. Step by step workflow to enrich customer emails with multi social (Instagram, Tiktok, Youtube, Twitter, Onlyfans, Twitch and more) profiles, analytics and metrics using the influencers.club API and sending tailored outreach to activate them as creators. Set up: Hubspot (can be swapped for any CRM like Salesforce, Attio or DB) Influencers.club Gmail Sendgrid (can be swapped for any programmatic email sender like Mailgun)

7 nodes6 views

Send personalized B2B cold outreach emails with Google Gemini and Gmail

by Nguyen Thieu Toan

Quick overview This workflow runs on a schedule to pull leads from an n8n Data Table, fetches and cleans each lead’s website content, uses Google Gemini to draft a personalized cold outreach email in structured JSON, sends it via Gmail, and updates the lead status back in the Data Table. How it works Runs on a schedule and loads sender details, the n8n Data Table ID, email length limits, and rate-limiting settings. Retrieves all leads from the n8n Data Table and checks each lead’s email address against a strict regex format. Marks leads with invalid email addresses as INVALID_EMAIL in the n8n Data Table with an updated timestamp. Fetches the lead’s website HTML and strips scripts, styles, and tags to produce a short plain-text website summary. Sends the lead details, website summary, and sender context to Google Gemini to generate a personalized outreach email returned as a JSON object (subject, greeting, opening line, main body, ending). Sends the generated email to the lead using Gmail with the AI-generated subject and body. Updates the lead record in the n8n Data Table to SENT (including sent time and email subject) and waits for the configured delay to rate-limit sending. Setup Create or choose an n8n Data Table for your leads, ensure it includes fields like id, email, first_name, last_name, company_name, and website plus status tracking columns, and paste its ID into data_table_id in the Set Context step. Add a Google Gemini (PaLM) API credential for the Google Gemini Flash model used to generate the outreach email. Configure Gmail sending credentials (Google API) for the Gmail node, and verify the sending account is allowed to send outbound email. Update the sender/company values and constraints in Set Context (sender name/email, company name/solution, max words, and rate-limit seconds) before activating the workflow. Requirements n8n Version:* Built and tested on *n8n 2.20.0+*. *(Note: You may encounter errors on older versions. It is highly recommended to update to the latest n8n version to use this workflow effectively). Google Gemini** API key credentials. Gmail OAuth2** credentials. Built-in n8n Data Table feature enabled. Customization Change the Email Provider:* Swap out the *Gmail* node for an *Outlook* or *SMTP** node if you use a different mailing service. Change the AI Model:* Replace the *Google Gemini* chat model with *OpenAI (ChatGPT)* or *Anthropic (Claude)** depending on your preference. Integrate your CRM:* Instead of using n8n Data Table, replace the fetch and update nodes with your CRM of choice, such as *HubSpot* or *Pipedrive**. Additional info About the Author Created by: Nguyễn Thiệu Toàn (Jay Nguyen) Email: me@nguyenthieutoan.com Website: nguyenthieutoan.com Company: GenStaff (genstaff.net) Socials (Facebook / X / LinkedIn): @nguyenthieutoan More templates: n8n.io/creators/nguyenthieutoan

6 nodes

Orchestrate patient admission, discharge and post-care with NVIDIA and Claude

by Cheng Siong Chin

How It Works This workflow automates patient risk assessment and clinical alerting for healthcare providers using NVIDIA AI models. Designed for hospitals, clinics, and healthcare organizations, it addresses the critical challenge of timely identification and response to high-risk patients requiring immediate intervention. The system monitors patient data webhooks, enriches records with external EHR data, and analyzes aggregated information through Claude AI for comprehensive risk stratification. Healthcare operations data is fetched and combined with patient metrics to provide contextual risk assessment. NVIDIA's structured generation capabilities ensure standardized clinical outputs, while parallel execution routes enable simultaneous processing: critical cases trigger immediate alerts via email and escalation flags, whereas routine cases follow standard documentation paths. The workflow maintains an audit trail, merges execution results, and generates detailed reports for compliance and quality improvement initiatives. Setup Steps Configure Patient Event Webhook with your EHR system endpoint URL and authentication headers Add NVIDIA API credentials (API key) in Fetch Patient Data and Structured Generation nodes Connect Claude Model node with Anthropic API key and configure healthcare risk assessment prompt Set up Gmail node with sender credentials and configure recipient email addresses for clinical alerts Enable Google Sheets integration for audit logging and specify spreadsheet ID for execution reports Prerequisites NVIDIA API access, Anthropic Claude API key, Google Workspace account (Gmail, Sheets) Use Cases Emergency department triage automation, post-operative monitoring for deterioration detection Customization Modify risk scoring algorithms, add disease-specific assessment criteria Benefits Reduces clinical response time through automated risk detection

Detect and enforce abuse cases with OpenAI, Slack, Gmail and Sheets

by Cheng Siong Chin

How It Works This workflow automates platform trust and safety operations by deploying a multi-agent AI system that detects abuse signals, investigates behaviour, scores risk, checks policy compliance, and enforces actions automatically. Designed for platform safety teams, content moderation managers, and compliance officers, it eliminates manual triage delays and ensures high-severity violations are actioned immediately. An abuse signal webhook triggers behaviour analysis via OpenAI, classifying signals by severity. A routing node directs cases to a Governance Agent, which orchestrates Investigation, Risk Scoring, and Policy Compliance Checker sub-agents. Enforcement data is prepared, then routed by action type-logging to abuse records, alerting the security team via Slack, sending escalation emails, or triggering auto-enforcement actions based on threshold checks—before all outcomes are logged. Setup Steps Configure Abuse Signal Webhook URL and authenticate incoming POST requests. Add OpenAI API credentials to all OpenAI Model nodes. Connect Google Sheets for abuse records and enforcement action logging. Configure Slack credentials and set security team alert channel. Add Gmail/SMTP credentials to Send Escalation Email node. Prerequisites Slack workspace with bot token Gmail or SMTP credentials Google Sheets for abuse and enforcement logging Use Cases Real-time abuse detection and auto-suspension on social platforms Customization Replace OpenAI with Anthropic Claude or NVIDIA NIM models Benefits Eliminates manual abuse triage with real-time AI signal processing