Bede | AI Voice Agent | GoHighLevel Integration

Voice AI & Automation Engineer | Building Voice AI agents that answer calls, qualify leads & book appointments 24/7 | Retell • n8n • GHL • Make

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 10 of building Voice AI Agents in public and growing on X

I have completed the workflow that checks availability, reschedules and cancels appointments, enabling my Telnyx voice agent to handle these tasks seamlessly.

I also did a test, showing how my GoHighLevel workflow sends data to n8n and triggers the appointment reminder calls.

The basic structures of the build have been set up. What’s left is to ensure that everything works together properly.

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 9 of building Voice AI Agents in public and growing on X

This post was supposed to go live yesterday, but I’ve had to do a lot of API docs reading, tinkering and figuring stuff out with this build.

So far, I have set up:
- The 2 GoHighLevel automations that send client info to n8n
- The call initiation, end-of-call report and time checking workflows on n8n
- The AI voice assistant and TeXML application on Telnyx

I still have to build workflows for checking availability, rescheduling and cancelling appointments.

Every element of this build is slowly coming together, and in the end, all the individual workflows and systems will be connected to work seamlessly with each other.

P.S. I will make a demo video in my next post.

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 9 of building Voice AI Agents in public and growing on X

This post was supposed to go live yesterday, but I’ve had to do a lot of API docs reading, tinkering and figuring stuff out with this build.

So far, I have set up:
- The 2 GoHighLevel automations that send client info to n8n
- The call initiation, end-of-call report and time checking workflows on n8n
- The AI voice assistant and TeXML application on Telnyx

I still have to build workflows for checking availability, rescheduling and cancelling appointments.

Every element of this build is slowly coming together, and in the end, all the individual workflows and systems will be connected to work seamlessly with each other.

P.S. I will make a demo video in my next post.

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 8 of building Voice AI Agents in public and growing on X

As said in my previous post, I’ve started building on Telnyx.

Tbh, the layout is quite different from what I was used to on Retell, but I’m slowly getting a grasp of things.

My new project is an appointment reminder Voice AI agent for a plumbing company.

The idea is to have a GoHighLevel workflow that checks for appointments that will happen within the next 24 hours.

For every client found, our n8n webhook will be triggered, and the agent will call the client via the Telnyx API to confirm they will be available.

And after the call, the outcome/post-call data will be updated on GoHighLevel.

So, what I did today was to create the prompt for the Telnyx voice agent and test it to ensure it works to my satisfaction (see one of the demo calls below).

I’ll start building the backend automations tomorrow to ensure the agent has everything it needs to do its job properly.

More updates coming tomorrow…

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

@Thomas_Tao_1 Same here. It was a necessary switch in a bid to reduce latency.

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 8 of building Voice AI Agents in public and growing on X

As said in my previous post, I’ve started building on Telnyx.

Tbh, the layout is quite different from what I was used to on Retell, but I’m slowly getting a grasp of things.

My new project is an appointment reminder Voice AI agent for a plumbing company.

The idea is to have a GoHighLevel workflow that checks for appointments that will happen within the next 24 hours.

For every client found, our n8n webhook will be triggered, and the agent will call the client via the Telnyx API to confirm they will be available.

And after the call, the outcome/post-call data will be updated on GoHighLevel.

So, what I did today was to create the prompt for the Telnyx voice agent and test it to ensure it works to my satisfaction (see one of the demo calls below).

I’ll start building the backend automations tomorrow to ensure the agent has everything it needs to do its job properly.

More updates coming tomorrow…

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 7 of building Voice AI Agents in public and growing on X

During my time away, I researched on ways to reduce latency for Voice AI agents (while still using low-code/no-code builders).

One platform that kept popping up was Telnyx.

I used to think they were just a telephony provider, but it appears they’ve made it possible for Voice AI developers to create voice agents on their platform.

And having done various test calls, I must say I am impressed with both latency and the platform’s features (you can check my demo call below).

In the coming days, weeks, and months, I’ll be building voice agents with various use cases to see how good voice agents built on Telnyx are and how they integrate with external platforms.

So, watch this space as I’ll be sharing everything I learn along the way.

Is anyone else building voice agents on Telnyx? Please share your experience.

Also, are there any specific use cases you would like me to build?

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 7 of building Voice AI Agents in public and growing on X

During my time away, I researched on ways to reduce latency for Voice AI agents (while still using low-code/no-code builders).

One platform that kept popping up was Telnyx.

I used to think they were just a telephony provider, but it appears they’ve made it possible for Voice AI developers to create voice agents on their platform.

And having done various test calls, I must say I am impressed with both latency and the platform’s features (you can check my demo call below).

In the coming days, weeks, and months, I’ll be building voice agents with various use cases to see how good voice agents built on Telnyx are and how they integrate with external platforms.

So, watch this space as I’ll be sharing everything I learn along the way.

Is anyone else building voice agents on Telnyx? Please share your experience.

Also, are there any specific use cases you would like me to build?

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

@theIBAA Thank you!😁

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 6 of building Voice AI Agents in public and growing on X

So, I’ve been away, but didn’t stop building.

The project I was working on before going MIA was an HVAC AI receptionist.

I have completed it, and you can find the demo video below.

I’m open to hearing your thoughts on what I can improve

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

@DesignByMaeL Yes, I've seen that latency is a big deal because of user experience. I've been researching ways to reduce it drastically.

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

@0xWuki Thanks man😃

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 5 of building Voice AI Agents in public and growing on X

I’ve not been able to post for two days due to situations beyond my control

I made a quick video to show what I’ve built so far and how the voice agent sounds.

I believe there's still room for improvement here.

What do you think needs improving in the voice agent?

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

@Julious_design You're welcome

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 4 of building Voice AI Agents in public and growing on X (late update)

Yesterday, I didn’t get to do much building.

Spent most of the day cleaning and arranging my space, and spending time with family.

The only work I did with my voice agent was to refine the sample interactions section of my prompt.

I haven’t tested it, but that will be happening today. I will also make a video to show what I have built so far.

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 3 of building Voice AI Agents in public and growing on X

I dedicated today to testing and refining the voice agent, and it’s not going according to plan.

It goes off track in certain scenarios and calls functions before they are needed.

After troubleshooting, I found some contradictions in the prompt, which I have fixed. Also discovered that the sample interactions in the prompt are not solid, so I’m working on refining them.

Regarding latency, there’s a 2-second delay between when I finish talking and when the agent responds.

Does anyone have any idea on how I can reduce it to 1 second?

Well, that’s it for today. We go again tomorrow.

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

@MustafyOf You're welcome

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

@Ikimi_Dann1e You're welcome

[Bede | AI Voice Agent | GoHighLevel Integration @ikwuezebede]

Day 2 of building Voice AI Agents in public and growing on X

I started working on the n8n part of the build.

This is where everything comes together in terms of connecting my Retell AI agent to the GoHighLevel CRM.

Since this is an inbound Voice AI agent, I decided to create 2 routes: one for new existing clients and the other for new callers.

In a real-life setting, there is a difference in how both personas are handled, so I want to effect that in the voice AI agent and see how it pans out.

Milestone in the past 24 hours: I finally crossed the 100 followers mark on X, and you all made it possible.😁😁😁😁

I’ll keep showing up more and sharing my progress