Hey everyone! I signed up here about a year ago but never really used the forum much, I've been working on a new venture since a month or two and it seems promising, so I want to keep an execution thread here to stay motivated and share what I learn along the way. Will try to post on a weekly basis
The idea
I setup Google ads campaigns, landing pages and a voice AI to handle incoming phone calls
Customers get booked appointments without lifting a finger
I charge ad costs + monthly service fee and keep full ownership of all the above
With minor tweaks, the same setup can be rented to several companies of the same niche in different areas, so it can scale
Targets are appliance repair businesses. Why?
1. There's generally enough search volume on Google for their services
2. In populated areas there's also enough appliance repair businesses
3. Incoming phone calls are pretty simple to handle for AI, callers usually just want someone to come as soon as possible and there's little to discuss aside from time, address, and the travel fee of the technician
How I plan to get customers:
getting a list of leads for an area is easy trough google search/maps
Phase A: free appointments
- start running a campaign for the area and wait for a booking
- forward it to my leads(first one who answers gets it)
- wait a day, then send a second customer to the same business
Phase B: free trial
- wait another day, then briefly pitch them my services
- if they're interested, offer a "free trial": they can invest any sum they're comfortable with, it goes all into Google ads and they get the appointments, I don't charge anything
After the free trial: 300€/month + ad costs + telephony/voice AI costs(and eventually a margin on those as well)
Current results
I currently have one guy in Lisbon in the free trial with 100€. Turned off the ads this week because he's off on Christmas, so here's the preliminary results(includes the 2 free appointments + ~30€ of his free trial):
Total spent: 48.20€
Impressions: 283
Clicks: 50
CTR: 17.7%
Conversions: 17 (clicks on landing page tap-to-call button)
Conversion rate: 34%
Calls: 17
Cost per call: 2.84€
Booked appointments: 8
AI failures: 4+ (this could have been booked appointments)
Call-to-booking rate: 47%
Cost per appointment: 6.03€
Looks good to me!
Customer charges 25€ travel fee + repair costs(quote on site), he should be very profitable
I asked how everything went with the first appointments. All good. He is not very talkative, I will have to squeeze the feedback out of him
The voice agent is clearly the weakest part of the chain, so that's the only thing I'll be working on for now
Target call-to-booking rate is 70+%
What's wrong with it:
1. Many speech recognition failures, particularly with location names and addresses. This is a huge mess, I'm currently manually reviewing call recordings and sending the customer corrections when the AI get's it wrong, not good
2. interruption handling: the voice agent abruptly stops speaking on short "filler noises"(like "hmm") and it doesn't handle resuming the conversation from an interruption well. It often generates a completely new sentence from scratch and forgets the point it was making before, skipping to the next stage of the conversation
3. the llm prompt needs work to root out some bad behavior, like hanging up without saying bye... pretty rude
4. high latency on some of the tools
Points 1 and 2 are what cost me most appointments, if fixed, 70% is realistic. But unfortunately the platform I am using(Elevenlabs) doesn't allow me to fix them. It has only 2 choices for the speech recognition model and the relevant settings for turn taking/interruptions handling are not accessible, so I need to start from scratch on a new platform...
At the moment the best candidate is python framework called Pipecat. It's a lot more complicated to use than no-code platforms like Elvenlabs/Vapi but allows for full control of all the underlying models
Plan for the next 2 weeks:
- improve the voice agent
- finish the first free trial and convert him to a paying customer. The google campaign has been running on 5€/day now and it is clearly limited by budget, next week I'll raise it to 10-15 so I can burn the remaining 70€ in under a week
- duplicate my working campaign and landing pages for Porto(next city), get a second business from there in the free trial
Highlight of the week:
A very smooth conversation between my voice agent and a real human, where they agree to book an appointment for the same day between 8 and 10 in the morning. The phone call was recorded at 11 AM...
The idea
I setup Google ads campaigns, landing pages and a voice AI to handle incoming phone calls
Customers get booked appointments without lifting a finger
I charge ad costs + monthly service fee and keep full ownership of all the above
With minor tweaks, the same setup can be rented to several companies of the same niche in different areas, so it can scale
Targets are appliance repair businesses. Why?
1. There's generally enough search volume on Google for their services
2. In populated areas there's also enough appliance repair businesses
3. Incoming phone calls are pretty simple to handle for AI, callers usually just want someone to come as soon as possible and there's little to discuss aside from time, address, and the travel fee of the technician
How I plan to get customers:
getting a list of leads for an area is easy trough google search/maps
Phase A: free appointments
- start running a campaign for the area and wait for a booking
- forward it to my leads(first one who answers gets it)
- wait a day, then send a second customer to the same business
Phase B: free trial
- wait another day, then briefly pitch them my services
- if they're interested, offer a "free trial": they can invest any sum they're comfortable with, it goes all into Google ads and they get the appointments, I don't charge anything
After the free trial: 300€/month + ad costs + telephony/voice AI costs(and eventually a margin on those as well)
Current results
I currently have one guy in Lisbon in the free trial with 100€. Turned off the ads this week because he's off on Christmas, so here's the preliminary results(includes the 2 free appointments + ~30€ of his free trial):
Total spent: 48.20€
Impressions: 283
Clicks: 50
CTR: 17.7%
Conversions: 17 (clicks on landing page tap-to-call button)
Conversion rate: 34%
Calls: 17
Cost per call: 2.84€
Booked appointments: 8
AI failures: 4+ (this could have been booked appointments)
Call-to-booking rate: 47%
Cost per appointment: 6.03€
Looks good to me!
Customer charges 25€ travel fee + repair costs(quote on site), he should be very profitable
I asked how everything went with the first appointments. All good. He is not very talkative, I will have to squeeze the feedback out of him
The voice agent is clearly the weakest part of the chain, so that's the only thing I'll be working on for now
Target call-to-booking rate is 70+%
What's wrong with it:
1. Many speech recognition failures, particularly with location names and addresses. This is a huge mess, I'm currently manually reviewing call recordings and sending the customer corrections when the AI get's it wrong, not good
2. interruption handling: the voice agent abruptly stops speaking on short "filler noises"(like "hmm") and it doesn't handle resuming the conversation from an interruption well. It often generates a completely new sentence from scratch and forgets the point it was making before, skipping to the next stage of the conversation
3. the llm prompt needs work to root out some bad behavior, like hanging up without saying bye... pretty rude
4. high latency on some of the tools
Points 1 and 2 are what cost me most appointments, if fixed, 70% is realistic. But unfortunately the platform I am using(Elevenlabs) doesn't allow me to fix them. It has only 2 choices for the speech recognition model and the relevant settings for turn taking/interruptions handling are not accessible, so I need to start from scratch on a new platform...
At the moment the best candidate is python framework called Pipecat. It's a lot more complicated to use than no-code platforms like Elvenlabs/Vapi but allows for full control of all the underlying models
Plan for the next 2 weeks:
- improve the voice agent
- finish the first free trial and convert him to a paying customer. The google campaign has been running on 5€/day now and it is clearly limited by budget, next week I'll raise it to 10-15 so I can burn the remaining 70€ in under a week
- duplicate my working campaign and landing pages for Porto(next city), get a second business from there in the free trial
Highlight of the week:
A very smooth conversation between my voice agent and a real human, where they agree to book an appointment for the same day between 8 and 10 in the morning. The phone call was recorded at 11 AM...
Dislike ads? Become a Fastlane member:
Subscribe today and surround yourself with winners and millionaire mentors, not those broke friends who only want to drink beer and play video games. :-)
Access Restricted: Unlock 1,000,000+ Posts
Stop Peeking Through the Keyhole.
Open the Door.
You hit a wall because the best advice isn't free—it's earned. Since 2007, MJ DeMarco has been active here daily (99.9% active rate), building a war room for entrepreneurs who refuse to settle for mediocrity.
- Active Daily: Life-changing content posted dozens of times every day. No dead threads.
- MJ's Inner Circle: Direct access to the author of The Millionaire Fastlane.
- Vetted Network: Connect with founders scaling to 7 and 8 figures.
- Proven Roadmaps: Strategy over motivation. Execution over "hustle porn."
"SCALED TO 6-FIGURE MONTHS"
"Tips and advice here saved me from huge mistakes, others allowed me to scale my business to 6-figure months."— User MitchC
"INSANE QUALITY OF PEOPLE"
"The number of high quality people I've met from this forum has been insane. All of it predicated on building real value."— User Richard Peck
"You are the average of the five people you surround yourself with."
Are you surrounding yourself with success?
