1What it does
Stellantis &You UK built two-way SMS with ClickSend and Knack in Make, then added OpenAI to classify intent and sentiment. Messages that need no reply get auto-closed. Subtle unhappiness gets flagged for a person.
2Who it is for
- Industry
- Automotive dealership
- Platform
- Make
- Category
- Support triage
- Tools used
- MakeClickSendKnackOpenAI
3How it works
Every step, in order, is in Workflow at a glance, pinned on the right as you scroll.
- 01 Trigger
Inbound customer SMS.
- 02 Action
ClickSend receives the SMS.
- 03 AI / action
OpenAI classifies intent and sentiment.
- 04 Action
No action needed: mark closed.
- 05 Action
Signs of dissatisfaction: flag for human follow-up.
- 06 Action
Booking confirmations and reminders run in the same system.
- Trigger
- Inbound customer SMS.
- AI step
- Unknown. Intent and sentiment classification.
4How it went
Per Make's case study: 47,000+ messages analyzed and 18,000+ auto-closed over 12 months, 151 hours saved (based on 30 seconds per review), and 'Communication' became a top reason for 9 to 10 NPS scores.
Only what the source reports. Numbers above come from the source.
5Worth knowing
- They piloted with no real customer data first, then tested models for accuracy and cost.
- They run 700+ Make scenarios.
6Recipe
Written by Builds That Work The source did not publish this. We wrote it as a starting point.
Starter classifier: "Classify this customer SMS. Return JSON {intent: booking|question|complaint|thanks|other, needs_reply: true|false, unhappy: true|false}.\nSMS: {{text}}"