Lead follow-up & inside sales

After a showing, send a Telegram voice note and let n8n update the buyer in GoHighLevel, email three similar listings, and create timed follow-up tasks

An n8n template by Rahul Joshi listens for Telegram voice notes from agents after a site visit (and prompts for a voice note if the agent sends text). OpenAI transcribes the note, then GPT-4o-mini extracts the buyer's contact details, preferences, budget, and interest level and builds a CRM note plus a timed follow-up plan. If a phone number or email is present, the workflow upserts the buyer in GoHighLevel and adds the visit note to the contact record. It then reads available listings from Google Sheets, scores them against the buyer's criteria, picks the top three while excluding the property just visited, and emails those matches to the buyer through Gmail. Finally it creates GoHighLevel follow-up tasks for each scheduled step and sends the agent a Telegram confirmation summarizing the CRM update, the emailed matches, and the task due dates.

At a glance

Tools used
n8nTelegramOpenAIGoHighLevelGoogle SheetsGmail
Category
Lead follow-up & inside sales
Audience
agent, team lead
Difficulty
Builder
Source
Follow up real estate site visits from Telegram with OpenAI, GoHighLevel and Gmail
Rahul Joshi (n8n creator) · n8n template · Sep 24, 2026
First step
Create a Telegram bot, connect Telegram, OpenAI, GoHighLevel, Google Sheets, and Gmail credentials in the n8n template, build a listings sheet with listing_id, title, area, price, bedrooms, property_type, url, and status columns, and send one test voice note that names the buyer plus a phone number or email. Tip: put the Telegram message link in the GoHighLevel visit note so the original audio is one tap away when a transcript gets a detail wrong. Tip (our recommendation, not a step in Rahul Joshi's n8n template): have the extraction step save preferred channel, next follow-up day, and the buyer's time zone as GoHighLevel custom fields, so a later workflow can trigger off those fields and a buyer who says "text me Thursday" in another state gets the text Thursday morning in the buyer's time zone. Store the buyer's time zone as a named zone (for example America/Chicago), not a fixed UTC offset like UTC-6, so follow-up times survive daylight saving changes. Tip (our recommendation, prompted by a reader question on X, not a step in Rahul Joshi's n8n template): when a buyer replies twice in a row on a different channel than the stored preferred channel, have the workflow flag the channel change for the agent to confirm, and update the GoHighLevel preferred-channel field only after the agent confirms, so the GoHighLevel preferred-channel field does not go stale. Tip (our recommendation, prompted by a reader question on X, not a step in Rahul Joshi's n8n template): next to the preferred channel, add a GoHighLevel custom field for the date the channel was switched, stored as channel switched on [DATE], and after the transaction closes, have a step reset the preferred channel to the buyer's original choice once the agent confirms. On the closing date stored in GoHighLevel, have the workflow create a task for the agent to confirm the preferred-channel reset, instead of waiting for the deal stage to change to closed.

Compliance note: Get the buyer's consent before emailing listings or adding the buyer to automated follow-up, follow CAN-SPAM and your state's rules for marketing email, keep fair housing in mind when the model summarizes preferences, have the agent approve the same-day thank-you text before the thank-you text sends, and review the first several automated emails before letting the workflow send unattended.

Suggested prompt

Written by this directory, inspired by the source. Replace the [bracketed] parts.

From this post-showing voice note transcript, extract: buyer name, phone, email, property visited, budget range, must-haves, deal-breakers, preferred areas, preferred follow-up channel (text, call, or email), any next-step timing the buyer stated, the buyer's time zone if the buyer stated one, and interest level (hot, warm, or cold) with the words that justify it. Write a three-line CRM note in plain language. Then propose a follow-up plan with timed steps (for example a same-day thank-you, a two-day check-in, and a one-week listing update) and a title for each task, and optionally draft a two-sentence same-day thank-you text that names one feature the buyers said they liked in the voice note (for example the backyard) for the agent to approve before the thank-you text sends. If a field was not said in the note, leave it blank instead of guessing.

Share on X Read the source

← All use cases