Hi everyone,
I built a workflow that turns a simple contact form into an AI front desk for small service businesses (clinics, salons, trades, consultants). Customers get a personal reply within seconds, also at night and on weekends.
Try it live (sample business, reply comes to your email):
https://barnabas.app.n8n.cloud/form/demo-en?branche=dental%20clinic
You can change the business type at the end of the link, e.g. ?branche=hair%20salon or ?branche=accounting%20firm
How it works (5 nodes):
1. Form Trigger – name, email, topic (appointment / prices / urgent / other), message
2. Set node "Business Settings" – business name, type, address, opening hours, services, free slots, reply language, sign-off. All business details live here, so non-technical owners never touch the prompt
3. Basic LLM Chain + OpenAI – the system prompt is built with expressions from the settings, so one template works for any business
4. Send Email (SMTP) – reply goes to the customer
5. Form completion page – shows the same reply on screen
A few things I learned:
- Urgent requests need an explicit rule. Without it, the model happily offered an "emergency" slot three days later. Telling it to offer today or tomorrow morning and ask the customer to call fixed that.
- Never let it invent prices or give medical/legal advice. I added "it depends on the details, happy to clarify" as the default answer to price questions.
- Keeping all business data in one Set node made it much easier to reuse the workflow for different industries. The demo above uses the same workflow for any business type.
- Many owners want to stay personal. A variant that only writes drafts for the owner to review and send works well for them.
If you want to use it without building it yourself, I put the ready-to-import version with a setup guide here: https://barnabuzz85.gumroad.com/l/ai-enquiry-assistant
Happy to answer questions about the setup or the prompt, and I'd love feedback on how you would improve it.