the demo i want a plumber to hear is his own shop picking up the phone. so i built that its one agent in Retell. the prompt has a business_name and a trade variable, and my demo page fills them in when it starts the web call each prospect gets their own page, /demo/their-name. theres a small allowlist on my server that maps the slug to the real company name, otherwise anyone could type whatever in the url and make the agent say it did a test call for one shop and the first thing the agent said was their company name. same agent, same prompt, but honestly it stops feeling like a template once you hear your own business answer keep the demo agent fully separate from your prod agent btw. one bad edit and a real clients line answers with the wrong name also check the phone number before you build anything. the one in the contractor licence registry wasnt even the one the shop has on its website anyone actually measured if personalized demos close better than generic ones?

I spent some time on a plumbing forum to see how owners actually talk about missed calls. One thread stuck with me. A guy running a 3-person shop posted 47 missed calls in a month, only 5 voicemails. He'd already tried his wife on the phone and a part-time receptionist at $1,500 a month. The plumbers who answered weren't impressed. One misses 5 to 10 calls a day and says most are price shoppers. Another said to tag every missed call for 2 or 3 weeks before buying anything, and when his shop did that, most of the calls were junk. From now on, when I talk to an owner, my first question will be how many of his missed calls were real jobs, before I mention the agent at all. If he doesn't know, that's the first thing we find out together. How do you handle owners who say "most of those calls are garbage anyway"?

Quick one for anyone whose agent texts the owner or the caller after a call. In my Twilio logs I found 23 texts that never went out between February and June, while my n8n workflow showed them as sent. All of them failed with error 30044. My account was still a trial, and trial accounts refuse texts that are too long. 10 of them also had a long dash, and one character like that makes each part of the text hold 67 characters instead of 153. For an agent whose whole job is to text the owner about an urgent call, a text that fails without anyone noticing is about the worst bug it can have. So my templates use plain characters only now, and I check the delivery status in Twilio rather than trusting the workflow. Has anyone else found failures like this long after they happened?

Hey everyone, Haris here, from Canada (originally Cameroon). Some of you might remember Sentinelle, the property-maintenance agent that took first place in the Retell hackathon here. I've been building quietly since then. Today I posted my backstory on LinkedIn for the first time: I got my American high school diploma at 12, studying online from Cameroon after my regular school day, on a connection that kept dropping. The diploma is in the slides. Now I run DICSIO, where I build AI agents for small businesses. Right now I'm focused on phone agents for plumbing and HVAC businesses. I'm going to keep posting what I build there, including what breaks. Here it is: https://lnkd.in/p/g4ptBXrB It's my very first post there, so if it resonates, a like or a comment would honestly help a lot. And a real question for the builders here: what made you start posting about your work, and did it actually bring you clients?

I build phone agents for plumbing and HVAC shops, and last month I found a hole in my own demo prompt. The flow was intake first: name, address, callback number, what's going on. Then fixed keywords like "gas smell" or "sewage backing up" would trigger an urgent text to the owner. Then I was reading through an open-source intake template for contractors and saw they put one step before everything else. If the caller mentions a gas smell or a CO alarm, or anything electrical that's sparking or wet, the agent drops the intake and tells them to get out and call 911. Before it even asks for a name. Mine didn't have that. Someone smelling gas would have been asked to spell their street name. I added it as step 0 and checked the live prompt through the Retell API after the update to make sure it was really there. What safety rules do you put in your agents before the intake starts?