Researched: Sourced and summarized, not yet run end-to-end by us Customer support

Classify support emails by sentiment with ChatGPT and open Freshdesk tickets

Last checked: · Sourced and summarized; not yet run end-to-end by us. What this means

At a glance

Setup estimate
1–2 h (ESTIMATE)
Tools
Industries
General (any small business), E-commerce
View source template on make.com (opens in new tab)

Problem

Customer emails pile up in a shared inbox with no ticket and no sense of which ones are upset

Caveats

Built as a learning template for Make's "AI Tools 101" course, so treat it as a starting point. Needs Make, an OpenAI API key (cost UNKNOWN) and Freshdesk. Customer emails are sent to OpenAI. Sentiment labels can be wrong, especially for sarcasm or mixed messages, so don't use them as the only way to spot urgent problems.

Tools

  • Make
  • Gmail
  • OpenAI (ChatGPT)
  • JSON parser
  • Freshdesk

Browse by tool: Make, Gmail, OpenAI, Freshdesk

Steps

  1. Watch Gmail for new emails.
  2. ChatGPT classifies each email as positive, negative or neutral.
  3. Parse the result.
  4. Create a Freshdesk ticket based on the sentiment result.
Status
Researched: Sourced and summarized, not yet run end-to-end by us
Last checked

Setup estimates are rough guesses for someone comfortable with the tools who already has the accounts and API keys. Real time is unknown until tested.

5 AI workflows for small businesses, every Tuesday (launching soon)

One deep-dive plus four quick picks for lead follow-up, CRM, proposals, onboarding and support, labeled Researched or Tested. This form doesn't collect anything yet.

Signups open soon. Privacy policy.

Why trust it: every entry links its source and caveats, and nothing is called Tested until we publish the logs. Our method.

More customer support workflows