AI Agent workflow'ları
518 sonuç — tümü kaynağa bağlı n8n referansı
Monitor procurement spend and risks with Google Sheets, Groq Llama and Gmail
Quick Overview This workflow runs every 15 minutes to pull procurement transactions and department budgets from Google Sheets, calculate spend utilization and forecasts, generate executive risk insights with Groq Llama, write dashboard records back to Google Sheets, and email critical-risk alerts via Gmail. How it works Runs every 15 minutes on a schedule. Reads procurement transactions and department budget/threshold data from Google Sheets and combines them by department. Aggregates spend per department, calculates utilization, remaining budget, projected monthly spend, vendor counts, pending transactions, and assigns SAFE/WARNING/CRITICAL risk levels. Uses Groq (Llama 3.1) to generate an executive-friendly risk analysis and recommendations for each department based on the calculated metrics. Builds organization-wide dashboard totals (spend, projections, alert counts, highest-spend and top-risk departments) and attaches workflow metadata. Formats the results into executive summary, department insight, and risk alert records and appends them into separate Google Sheets tabs. Sends a Gmail escalation email when a CRITICAL risk alert record is detected. Setup Connect Google Sheets O
Run AI employee onboarding with Groq, Google Workspace, and Gemini
Quick overview This workflow ingests onboarding documents from Google Drive into an in-memory vector store with Google Gemini embeddings, then runs a multi-channel onboarding chatbot powered by Groq, sends role-based welcome emails via Gmail when a new hire appears in Google Sheets, and schedules day 1/7/30 check-in emails. How it works Manually runs an ingestion flow that lists files in a specified Google Drive folder, downloads them, extracts text, and indexes them into an in-memory vector store using Google Gemini embeddings. Receives employee questions via n8n Chat (and optionally Slack or Telegram), normalizes the message into a single format, and rejects empty or overly long inputs. Uses a Groq LLM agent with per-user session memory to retrieve relevant excerpts from the in-memory company_docs knowledge base and generate a grounded answer, or directs the user to HR when no relevant content is found. Routes the assistant’s response back to the originating channel (n8n Chat, and optionally Slack or Telegram). Triggers every minute on updates to a Google Sheets “Hires” sheet, validates rows marked as Hired, and skips hires already present in the “Tracking” sheet. Builds a depart
Send a daily AI-ranked job match digest with Anthropic and Gmail from 6+ boards
Quick Overview This workflow runs every morning to pull jobs from multiple free job boards (and optional Adzuna), filters and de-duplicates results, scores each role against your resume and preferences using Anthropic, then emails a ranked digest via Gmail. How it works Runs every morning at 08:00 using a schedule trigger. Builds a list of job board API URLs (RemoteOK, Remotive, Arbeitnow, Jobicy, Himalayas, TheMuse, and optional Adzuna) based on your search settings. Fetches job listings from each board over HTTP. Normalizes all responses into a single job format, applies your filters (title keywords, tags, locations/remote, salary, and excluded terms), and de-duplicates roles across boards. Sends your resume, preferences, and the short-listed jobs to an Anthropic-powered AI agent to return a structured fit score, fit reason, risk/concerns, and a tailored first-person pitch for each job. Keeps only roles at or above your minimum score, ranks them by score, and generates an HTML email digest with links and source board attribution. Sends the digest to your target address using Gmail. Setup Add an Anthropic API credential, select an Anthropic chat model, and keep the structured outp
Chat with Google Docs and Sheets using a Claude AI agent
Quick Overview This workflow opens an n8n chat interface where an AI agent powered by Anthropic Claude answers questions by reading a connected Google Doc and/or Google Sheet on demand, using short-term conversation memory for follow-up questions. How it works Receives a user question through n8n’s built-in chat trigger. Uses an AI agent powered by Anthropic Claude to decide whether it needs to consult Google Docs, Google Sheets, or both. Reads the full text of the configured Google Doc when the question is about document content. Fetches rows from the configured Google Sheet tab when the question is about spreadsheet data. Answers in the chat based only on the retrieved Doc/Sheet content (or states the information isn’t present), while keeping the last 10 messages in memory for contextual follow-ups. Setup Add credentials for Anthropic (Claude), Google Docs, and Google Sheets. Replace YOUR_GOOGLE_DOC_ID_OR_URL with your target Google Doc URL or ID. Set the Google Sheet document ID and choose the tab (sheet name) to read in the Google Sheets tool. Confirm the connected Google account has access to both the Doc and the Sheet, then open the workflow chat and start asking questions.
WMonitor competitor pricing page changes with OpenAI, Google Sheets and Slack
Quick overview This workflow checks competitor pricing pages each morning, compares the latest visible text to the previous snapshot stored in Google Sheets, and uses OpenAI to summarize meaningful changes before posting an alert to a Slack channel. How it works Runs every day at 08:00 on a schedule. Reads your configured list of competitor page URLs and processes each URL individually. Fetches each page’s HTML, extracts the visible body text, and trims it to a snapshot for comparison. Looks up the previous snapshot for the same URL in Google Sheets. If a previous snapshot exists and the content differs, OpenAI generates a 1–3 bullet summary of meaningful changes based on your watch criteria. Posts the change summary to Slack and then appends or updates the URL’s latest snapshot and “Last Checked” timestamp in Google Sheets. If no previous snapshot exists or no change is detected, it skips the Slack alert and only saves the current snapshot to Google Sheets. Setup Add credentials for Google Sheets, Slack, and OpenAI (Chat) and select the target Slack channel. Create or choose a Google Sheet with columns named URL, Content, and Last Checked, and set the Spreadsheet ID and sheet name
Triage Gmail and draft replies with GPT-5 Mini and Google Sheets
Quick overview This workflow watches your Gmail inbox for new unread emails, uses OpenAI (GPT-5 Mini) to classify each message by priority and category, logs the triage results to Google Sheets, and creates a Gmail draft reply only when the email needs a personal response. How it works Triggers every minute when a new unread email arrives in Gmail. Applies your configured preferences (name, role, categories, and drafting style) to guide the triage. Sends the email content to OpenAI (GPT-5 Mini) to return a structured priority, category, one-line summary, whether a reply is needed, and an optional draft reply. Appends or updates a row in Google Sheets with the email details and triage results. If the email needs a reply, creates a draft reply in Gmail in the same thread addressed to the original sender. Setup Connect your Gmail account (Gmail Trigger and Gmail Draft) and ensure the trigger is set to monitor unread mail as desired. Add an OpenAI credential and select the GPT-5 Mini chat model in the OpenAI node. Connect Google Sheets OAuth credentials, then set the target spreadsheet document ID and sheet name for the triage dashboard. Create a Google Sheet with columns for Date, Fro
Run a Telegram voice assistant with OpenAI GPT-5 Mini and Google Calendar
Quick overview This workflow turns a Telegram bot into a personal assistant that accepts text or voice notes, transcribes audio with OpenAI, uses a memory-enabled AI agent for replies and calculations, and creates Google Calendar reminder events from natural-language requests. How it works Triggers whenever a new Telegram message is received by your bot. Routes the message based on type, sending text directly to the assistant, voice notes to transcription, and other message types to a fallback reply. For voice notes, downloads the audio from Telegram and transcribes it to text using OpenAI. Sends the user’s text (typed or transcribed) to an OpenAI chat model powered agent that remembers the last 15 messages per chat and can use a calculator tool. When the user asks for a reminder or event, the agent creates a Google Calendar event with AI-generated title and start/end times. Sends the agent’s final response back to the originating Telegram chat. Setup Create a Telegram bot with @BotFather, add a Telegram credential in n8n, and enable updates for the Telegram Trigger. Add an OpenAI credential with access to the chat model and audio transcription used in the workflow. Add a Google Ca
Handle WhatsApp support text and voice messages with OpenAI GPT-5 Mini
Quick overview This workflow responds to incoming WhatsApp Business messages, handling both text and voice notes by transcribing audio with OpenAI and replying with an OpenAI chat agent that uses per-contact conversation memory and answers only from the business information you provide. How it works Triggers when a new WhatsApp message is received. Routes the incoming message by type, sending text messages directly to the support flow, audio messages to a voice-transcription flow, and other message types to a fallback reply. For voice notes, retrieves the WhatsApp media URL, downloads the audio file, and transcribes it to text with OpenAI. Sends the customer’s text (original or transcribed) to an OpenAI chat agent configured to answer only using your provided business information. Stores and reuses a per-contact conversation history keyed by the sender’s WhatsApp number to maintain context across messages. Sends the agent’s response back to the customer as a WhatsApp text message. Setup Connect WhatsApp Business Cloud credentials for both receiving messages (trigger) and sending replies, and ensure your WhatsApp webhook is configured to point to the trigger URL. Add OpenAI credenti
Answer customer questions on WhatsApp with OpenAI, Gmail, and Google Sheets
Quick overview This workflow answers incoming WhatsApp customer questions using an OpenAI-powered agent that reads your live website via HTTP requests, escalates unclear or sensitive cases to your support team via Gmail, and logs each conversation to Google Sheets. How it works Triggers when a new WhatsApp message is received. Loads configuration values (company name, website root URL, AI model, language rules, escalation email, and WhatsApp phone number ID) and extracts the customer phone number and message text. Ignores non-text WhatsApp updates (such as images, audio, and status messages). Uses OpenAI with conversation memory to navigate your website by listing links and fetching relevant pages, then generates a structured result containing an answer, escalation flag, sentiment, topic, and source URLs. If the agent marks the request as needing a human, sends an escalation email via Gmail to your support address with the customer details and the drafted reply. Sends the final plain-text reply back to the customer on WhatsApp. Appends the conversation details to a Google Sheets log. Setup Connect credentials for WhatsApp Trigger/WhatsApp Cloud API, OpenAI, Gmail, and Google Sheets
Run a personal assistant for calendar, Gmail and utilities with OpenAI and MCP
Quick overview This workflow runs a chat-based OpenAI assistant in n8n that calls three MCP servers for Google Calendar management, Gmail search and draft replies, and utility tools for weather, currency conversion, and calculations. How it works Receives a chat message in n8n to start or continue a conversation. Loads the assistant configuration (model, memory window, and system prompt) and sends the user message to an OpenAI chat model with short-term conversation memory. When the assistant needs calendar actions, it calls an MCP server that searches, creates, updates, or deletes Google Calendar events. When the assistant needs email actions, it calls an MCP server that searches Gmail messages and creates draft replies for user review. When the assistant needs quick utilities, it calls an MCP server that fetches weather from wttr.in, converts currency using frankfurter.app, or runs calculations. Returns the assistant’s response back to the chat, including any results from the MCP tool calls. Setup Add credentials for OpenAI, Google Calendar (OAuth2), and Gmail (OAuth2) in n8n. Activate the workflow, then copy each MCP Server Trigger Production URL for the Calendar, Gmail, and Uti
Send daily AI news digests from RSS feeds with OpenAI, Telegram and Gmail
Quick overview This workflow runs every morning at 8 AM, pulls articles from multiple RSS feeds, deduplicates and trims the list, uses OpenAI (GPT-5 Mini) to write a short plain-text news digest, and delivers the same digest to a Telegram chat and a Gmail inbox. How it works Runs daily at 8:00 AM on a schedule. Splits your configured RSS feed URL list into individual feeds and fetches the latest articles from each one, continuing even if a feed fails. Removes duplicate articles by link, keeps only the newest items up to your maxArticles limit, and trims each item to title, link, publish date, and a short snippet. Combines all cleaned articles into a single JSON list and sends it to OpenAI (GPT-5 Mini) to select 5–8 top stories and generate a single plain-text digest under 3,500 characters. Sends the generated digest to your specified Telegram chat. Emails the same digest as a plain-text message through Gmail. Setup Add an OpenAI credential and ensure the GPT-5 Mini model is available in your account. Create a Telegram bot with @BotFather, add the Telegram credential in n8n, and set your target telegramChatId (for example by using @get_id_bot). Connect a Gmail account in n8n and set
Run a multimodal WhatsApp AI assistant with OpenAI for text, voice and images
Quick overview This workflow turns WhatsApp into a multimodal assistant that answers text, voice notes, images, and PDF documents using OpenAI, and can generate images with DALL·E on demand, replying back in WhatsApp with either text, audio, or an image. How it works Triggers when a new WhatsApp message is received. Detects whether the incoming message is text, a voice note, an image, a PDF document, or an unsupported type. For text messages, generates an image with OpenAI DALL·E when the message starts with the configured /image command, and sends the generated image back to the sender. For voice notes, fetches the WhatsApp media URL, downloads the audio, transcribes it with OpenAI, and forwards the transcript to the assistant. For images, fetches the WhatsApp media URL, downloads the image, describes it with OpenAI vision using the configured analysis prompt, and forwards the description (plus any caption) to the assistant. For PDF documents, downloads the file from WhatsApp, extracts the PDF text, and forwards the extracted content (plus any caption) to the assistant, while non-PDF files receive an error message. Uses an OpenAI chat model with per-contact short-term memory to ge
Screen LinkedIn jobs and send tailored applications with Apify and OpenAI
Quick overview This workflow collects your job preferences via an n8n form, uses Apify to scrape LinkedIn job listings, and has OpenAI screen and rank them. After you approve a shortlist by Gmail, it generates tailored resume and cover-letter PDFs with DocRaptor, emails them to you, and logs matches to Sheets How it works Receives your job preferences and resume text through an n8n form submission. Sends the preferences to Apify to scrape recent LinkedIn jobs, then removes duplicates and obvious mismatches (seniority, years of experience, and location). Uses OpenAI to score each job’s relevance and eligibility and returns a JSON verdict with a fit score and short reason. Keeps the top matches above the configured threshold and emails you a shortlist in Gmail with approve/decline buttons. If you approve, uses OpenAI to draft a job-specific resume and cover letter from only the provided profile text and generates HTML interview prep questions. Converts the resume and cover letter HTML into PDFs via DocRaptor, emails the tailored application package to you in Gmail, and appends the match details to a Google Sheets tracker. Setup Add credentials for OpenAI, Gmail OAuth2, Google Sheets
Draft personalized cold emails from Google Sheets using Gemini and HubSpot
Quick overview This workflow batch-processes leads from Google Sheets, researches each person and company using Tavily plus Google Gemini, drafts a personalized cold email, then creates/updates the contact in HubSpot with a research note and a saved email draft before marking the lead as Done in the sheet. How it works Runs manually and sets a batch size that controls how many leads to process per run. Reads leads from Google Sheets, skips rows already marked as Done, and normalizes the selected batch into clean lead fields. Processes the batch one lead at a time and uses Tavily search (last 60 days) with Google Gemini to extract the most recent public activity, pain points, and a personalization hook. Cleans and parses the research JSON so it can be reused reliably downstream. Uses Google Gemini to draft a 90–120 word cold outreach email (subject and body) that incorporates the research and offers a free “Automated LinkedIn Post Creator” n8n workflow. Creates or reuses the HubSpot contact, logs a HubSpot note with the research and drafted email, and saves the email as a HubSpot engagement in DRAFT status. Updates the matching row in Google Sheets (by Email) to set Status to “Done
Send and track Slack invoice requests with Gemini, Google Sheets and Gmail
Quick overview Youtube Video: https://youtu.be/42-oHh9w9Eo This workflow listens for Slack @mentions, uses Google Gemini to classify the request and extract invoice details, then reads and updates Google Sheets, generates invoice PDFs with PdfBro, sends invoices or reminders via Gmail, and posts confirmations or errors back to Slack. How it works Triggers when the workflow’s Slack app is mentioned in a specified Slack channel. Uses Google Gemini to classify the message intent (send invoice, remind invoice, or check status) and extract the recipient email and invoice due date. Routes the request based on the extracted intent and stops if the intent cannot be determined. For SEND_INVOICE, looks up the client in a Google Sheets “Client Bill” spreadsheet and posts a Slack error if no matching email is found. If a client is found, generates an invoice PDF with PdfBro, emails it to the client via Gmail, appends/updates the invoice record in a Google Sheets “Client Invoices” spreadsheet, and posts a Slack confirmation. For REMIND_INVOICE, finds an UNPAID invoice for the email in the “Client Invoices” Google Sheet, regenerates the invoice PDF with PdfBro, emails a reminder via Gmail, and p
CWReport workflow errors to Asana tasks with Google Chat and Grok 4.5 analysis
Quick overview When a monitored workflow fails, this reads its Description field to find the owning client and assignee, looks them up in Asana, analyses the error with Grok 4.5, and creates an assigned task with a severity-based due date. Repeat failures comment on the existing task. How it works Any workflow that names this one as its error workflow triggers it on failure. There is no production trigger by design. The error payload only carries a workflow ID, so an HTTP Request calls the n8n API to read that workflow's own Description field, which holds a small JSON object naming the client and the assignee. Asana projects and users are fetched live and matched by name, so no GID is ever hardcoded. A failed match raises a warning that travels downstream instead of failing silently. Asana is searched for an open task covering the same workflow and node. If one exists, a comment carrying the new execution ID is added rather than a duplicate task. For a new error, Grok 4.5 via OpenRouter returns a root cause, a solution, concrete steps, a severity, and a category, held to a fixed shape by a structured output parser. Severity sets the due date: Critical today, High tomorrow, Medium i
Handle WhatsApp customer service chats with Google Gemini and Google Sheets
Quick Overview This workflow handles WhatsApp customer service via OpenWA, using Google Gemini and a Google Sheets knowledge base to answer personal chats, while routing conversations to a human queue in Google Sheets when escalation is needed. How it works Triggers on OpenWA WhatsApp events and ignores message reactions except for specific ✅/☑ reactions used to re-enable AI handling. Normalizes incoming event fields and processes only personal chats (IDs ending with @c.us) that are not sent by your own WhatsApp account. Checks a Google Sheets “Session Chat” list to decide whether the conversation is currently handled by a human, and only continues to AI when no active human session is found. Uses a Google Gemini chat model with short-term memory and a Google Sheets “Database Company” tool to draft a friendly Indonesian response based only on the sheet data. If the AI response does not contain the #admin escalation token, it sends the reply back to the customer via OpenWA. If the AI response contains #admin, it appends the chat to the Google Sheets “Session Chat” sheet to flag it for human follow-up. When a ✅/☑ reaction is received for a chat, it finds and deletes the corresponding
Answer Slack knowledge base questions with Notion, Pinecone, and OpenAI
Quick overview This workflow syncs an internal Notion knowledge base to Pinecone on a daily schedule, then runs a Slack bot that answers questions in a specific channel using Pinecone retrieval and an OpenAI chat model, replying in a thread with citations back to the source Notion pages. How it works Runs on a daily schedule to sync your Notion knowledge base into a Pinecone vector index. Fetches all pages from a specified Notion database and then pulls each page’s full block content (including nested blocks). Combines each page’s blocks into a single document and attaches metadata such as the Notion title, URL, and last edited time. Splits each document into overlapping text chunks, generates OpenAI embeddings, and inserts them into Pinecone while clearing and rebuilding the target namespace. Triggers when a new message is posted in the configured Slack channel and ignores bot/system messages to prevent reply loops. Uses an OpenAI-powered agent that always queries Pinecone for relevant chunks (including source metadata) and generates an answer in the user’s language. Posts the response back to Slack as a threaded reply, including cited Notion sources. Setup Create a Notion interna
Assess visa eligibility and send PDF checklists with Claude and SendGrid
Quick overview This workflow receives visa eligibility requests via webhook, uses Anthropic Claude to extract passport details and assess visa requirements, generates a purpose-specific document checklist, converts it to a PDF via PDFShift, logs each request to Google Sheets, and emails the PDF to the applicant using SendGrid. How it works Receives a webhook POST with applicant details, travel plan, and raw passport OCR text. Uses Anthropic Claude to extract and validate passport data (name, nationality, passport number, expiry) and scores whether the passport text is readable. If passport data is insufficient, immediately returns a webhook response asking the applicant to resubmit a clearer passport scan. If passport data is sufficient, waits briefly and uses Anthropic Claude again to determine whether a visa is required, the visa category, a recommended visa type, processing time, risks, and confidence. Builds a document checklist from a standard base list plus purpose-specific items and any AI-flagged eligibility risks, then renders it into HTML. Sends the HTML to PDFShift to generate a PDF, formats a summary row, and appends the application record to a Google Sheets tab. Emails
Generate sales research briefs from Google Sheets with Claude and Slack
Quick overview This workflow polls a Google Sheets prospect queue, researches each new company via SerpAPI and enrichment/contact APIs, then uses Anthropic Claude to generate a structured sales brief and outreach draft, logs the result back to Google Sheets, marks the prospect as researched, and notifies a sales rep in Slack. How it works Runs every 15 minutes and reads rows from the Prospects tab in Google Sheets. Filters for prospects with Status set to "New" and processes each prospect as an individual item. Researches each prospect in parallel by pulling recent news via SerpAPI, company funding/technology data via a company enrichment API, and decision-maker contacts via a people data API. Combines the research signals into one profile per prospect and calculates a priority score and a research completeness score. Skips prospects with insufficient data and sends the rest to Anthropic Claude to generate a JSON sales brief and first-touch outreach message. Appends the brief to the SalesBriefs tab in Google Sheets, updates the original prospect’s Status to "Researched", and posts a summary plus the draft message to Slack via an incoming webhook. Setup Create a Google Sheets OAuth2
Route procurement helpdesk tickets from Gmail with Groq and Slack
Quick Overview This workflow monitors a Gmail inbox for procurement requests, uses Groq (Llama 3.1) to classify and route each ticket with SLA logic, escalates high-risk items to Slack, creates a ticket via HTTP, sends an auto-reply from Gmail, and logs the full record to Google Sheets. How it works Triggers every minute when a new email arrives in a Gmail inbox. Extracts the sender, subject, body, timestamps, and Gmail thread/message IDs, and generates a procurement ticket ID. Looks up the email thread ID in Google Sheets to identify whether the request already exists. Uses Groq (Llama 3.1) to classify the request (category, urgency, sentiment, risk, team) and applies routing rules to set SLA deadlines, queues, and escalation flags. If escalation is required, posts a detailed alert to a Slack channel and creates an escalated ticket via an HTTP request; otherwise, creates a standard ticket via an HTTP request. Generates a short agent-facing status summary with Groq, emails the customer an acknowledgement and suggested reply via Gmail, and appends the full ticket and AI outputs to Google Sheets. Setup Connect your Gmail OAuth2 credential for both the inbox trigger and the customer a
Detect early payment invoice opportunities with Groq and Google Sheets
Quick Overview This workflow runs daily to fetch invoice data from an HTTP endpoint, calculate early-payment savings and annualized ROI, and use Groq (Llama 3.3) to prioritize eligible discounts. It logs eligible and expired invoices to Google Sheets and emails a digest of HIGH-priority recommendations via Gmail. How it works Runs every day at 9:00 to start an invoice scan. Fetches invoice records from an external HTTP API endpoint. Calculates potential savings, days saved by paying early, annualized ROI, and whether each invoice is still eligible before the discount deadline. Routes eligible invoices to Groq (Llama 3.3) to return a JSON decision with a YES/NO recommendation, HIGH/MEDIUM/LOW priority, and a short reason, then normalizes the AI response. Appends eligible invoice evaluations (including ROI, savings, priority, recommendation, and reason) to a Google Sheets log. Filters for invoices marked YES and HIGH priority, aggregates them into a daily batch, and emails a plain-text digest to the finance team via Gmail. Marks ineligible invoices as expired with a skip reason and appends them to a separate Google Sheets audit log. Setup Add a Groq API credential and ensure the sele
Triage supplier capacity risk with Groq, Gmail, Slack, and Google Sheets
Quick overview This workflow runs daily to pull supplier performance data from an HTTP endpoint, calculate fulfillment and delay risk metrics, use Groq (Llama 3.3) to classify supplier risk and generate mitigation actions, then notify via Gmail and Slack and log outcomes to Google Sheets. How it works Runs every day at the scheduled time. Fetches supplier performance data from an HTTP API and stops if the supplier has no orders. Normalizes the supplier fields and calculates fulfillment rate, delay, rule-based flags, and a risk score. Sends the metrics to Groq (Llama 3.3) to classify risk (Low/Medium/High) with a reason and confidence, then merges the AI result with the computed metrics (falling back to the rule-based classification if parsing fails). If the AI risk is Medium, Groq generates three mitigation actions and the workflow fetches backup supplier options from an HTTP API, then combines both into a single alert payload. Sends the mitigation actions by Gmail email and Slack message, then appends a “high risk” audit row to Google Sheets. If the AI risk is not Medium, appends a “no alert sent” audit row to Google Sheets with the assessed risk details. Setup Add Groq API creden
Handle WhatsApp sales inquiries with OpenAI, Pinecone, Google Calendar, and Slack
Quick overview This workflow handles inbound WhatsApp Business messages with an OpenAI-powered sales agent that uses Pinecone RAG for product answers and Google Calendar tools for scheduling, then either replies on WhatsApp or hands the conversation to Slack, with Slack-based error alerts. How it works Triggers on every inbound WhatsApp Business message and checks whether the payload contains a text body. Drops non-text messages (such as images or voice notes) and continues only with text conversations. Uses an OpenAI chat model with per-customer session memory to generate a concise sales reply and decide whether a human handoff is needed. Lets the agent query a Pinecone-backed product catalog via OpenAI embeddings to answer pricing, specs, and availability questions. Lets the agent check availability and book appointments in Google Calendar when the customer requests scheduling. Parses the agent’s HANDOFF flag and either posts the customer message plus draft reply to a Slack channel or sends the reply back to the customer on WhatsApp. Sends a Slack alert to an ops channel if any step in the workflow fails. Setup Connect WhatsApp Business Cloud credentials and set the WhatsApp Phon