PROJECT: Lindy AI Email Sales Assistant + CRM Automation Objective Set up a lean AI sales assistant using Lindy that automatically handles inbound email enquiries for: • SnapCor software • Thermal inspection services • Technical/system solutions The assistant should qualify leads, send the correct response (trial link, quote, or meeting booking), and log key data into the CRM — while only involving Rachael when human input is required. This is not a traditional sales funnel. It is a conversational AI assistant working inside email. System Overview Function: AI Sales Assistant Tool: Lindy Function: Email Inbox Tool: Gmail Function: CRM Tool: HubSpot (Free) Function: Automation Connector (if needed) Tool: Zapier Function: Meeting Scheduling Tool: HubSpot Meeting Links PART 1 — Gmail + Lindy Setup • Connect Lindy securely to the sales inbox • Lindy monitors new inbound enquiries only • Avoid reprocessing entire threads to control credit usage PART 2 — Lead Type Identification The AI must classify enquiries into: Software (SnapCor) Inspection Services System / Technical Enquiries PART 3 — Conversation Logic Software Enquiries • Do not send pricing • Send the free trial link • Offer a meeting only if they request a demo, have a large team, or ask about advanced features Inspection Enquiries AI asks: “Roughly how many items or assets need to be inspected?” Items: 1–50 Action: Send 1 Day Inspection quote – 1,900 AED Items: 51+ Action: Send 2 Day Inspection quote – 3,800 AED System / Technical Enquiries • Do not ask questions • Do not send pricing • Immediately send fact-finding meeting link so Rachael can scope requirements PART 4 — Quote Handling • Quotes must use predefined templates • No AI-generated pricing • Use rule-based selection only • Include pricing and what’s included • Notify Rachael when a quote is sent PART 5 — Follow-Up Logic If no reply: • Follow up after 24 hours • Follow up again after 72 hours • Stop if the lead replies • Do not over-message PART 6 — CRM Sync When key information is gathered, the system should: • Create or update contact in HubSpot • Create a deal • Populate fields: Deal Type – Software – Inspection – System Qualification Fields – Number of Users – Use Case – Timeline – Number of Items – Location PART 7 — Notifications Notify Rachael only when: • A meeting is booked • A quote is sent • Lead requests a human • Lead asks complex technical questions No notifications for standard AI conversations. PART 8 — Credit Efficiency Rules To prevent Lindy credit overuse: • Only send the latest email message to AI • Do not summarize long threads • Keep prompts short and structured • Avoid multi-step AI reasoning where simple rules can be used Goal: maximum one AI generation per reply. Timeline Target completion: 1–2 weeks End Result An AI assistant that: • Replies to leads automatically • Sends trial links for software • Quotes inspections using simple rules • Books meetings for system enquiries • Follows up • Updates CRM • Only involves when necessary