Build a pay-first service flow with human approval
Use an agent to quote a service, collect Stripe payment, then hand over the deliverable, with Telegram approval for any refund or dispute.
- What they did
- A custom agent was arranged around a set order of steps: it examined the request, suggested the offer, handled payment, delivered the service, and then finished the job. It produced Stripe checkout links for the service and confirmed the money had landed before releasing the deliverable. The system was run on a local machine with no separate interface or hosting, and customers could reach it through messaging tools or a browser-based component. When a refund or dispute appeared, the owner got a Telegram message and the agent waited for that reply before taking action.
- How it went
- The project was presented as a practical way for non-technical business owners to sell services without building a full app. No performance data, sales totals, or customer outcomes were reported. Because refunds and disputes depend on a Telegram response, someone has to be available to approve them.
- Tips
- Make sure the approval step is clear for anything that could move money backward or end an order. Test the checkout and delivery path in Stripe test mode before turning on real payments.
Try it yourself with n8n
Set up a workflow for selling [service]. When a customer asks for [service], have the agent draft the offer and create a Stripe checkout link for [price]. Wait until payment is confirmed before sending [deliverable]. If a refund or dispute comes up, stop and message me on Telegram with the order details, then wait for my approval before doing anything else. Test the full process in Stripe test mode first, and do not enable live payments until I say so.
Fill in the [brackets]. Agents can make mistakes, so review anything before it is sent, paid or deleted.
Get 3 agent uses like this every morning
With the day's top AI news. Free, and one click to unsubscribe.
Related agent uses
Set up an outreach agent to call and email leads
An agent got its own contact details, reached out to matching leads, and kept following up until it could report interest.
Use a guarded loop to publish pages and a tool
An agent ran 365 short rounds, checked a revenue dashboard each time, and ended up with 394 linked pages plus a webhook checker site.
Build a website and WhatsApp support agent from a URL
A founder showed a setup that turns a business URL into a support agent that answers questions and books Calendar appointments.
Turn a client site into a ready ad campaign
An agent read a client website and drafted a multilingual ad plan, including keywords, copy, tracking, and a conversion issue to fix.
Automate renter replies and tour booking
An embedded leasing agent handled questions, tour scheduling, and follow-ups for multifamily listings.
Set up a voice agent for calls and simple support
A team used a voice agent for after-hours support, FAQs, and order-status checks, then pushed structured call details into their internal tools.