Building AI voice agents (without code) using Synthflow - Blog

Do you want to create an AI voice agent—without needing to write a line of code? Tom Osman’s out to show how easy it can be with Synthflow AI.

In a recent live workshop, Tom shared how using Synthflow, businesses can spin up voice agents capable of handling a whole host of interactions, from lead qualification to restaurant bookings.

What’s an AI voice agent? In a nutshell, it’s a phone answering bot powered by AI software. Voice assistants are designed to handle interaction via spoken word. So instead of a human answering a call, such as a customer enquiry, the AI will take it, navigating all the nuances found in regular human interaction.

If you’re looking to implement an AI-powered voice agent to reduce support costs or improve response times, this session is essential viewing.

Here’s a sneak peek into what Tom demoed:

Why voice AI is making a comeback

Tom kicked off by discussing why voice AI is finally reaching its potential—after years of user disappointment with voice assistants like Siri.

As Tom explained,

“The transcription accuracy is way better. The latency is way down. We have human sounding voices with ElevenLabs. [...] You can now create really accurate real-time speech voice agents powered by AI.”

Tom added that Synthflow aims to “make it as simple as possible for people to create AI voice agents” that can handle tasks for their companies. A couple of basic examples Tom gave were:

Having an AI receptionist answer a phone call and take a message.
Get the AI receptionist to forward that call to a human if it meets certain criteria.

Synthflow has powerful customization features, like the ability to modify voice parameters to sound natural and engaging. Using ElevenLabs, users can pick specific accents, inflections, alter the speed of the voice, change the latency, and more, making the AI as human-like as possible. It can even offer multi-language assistance, potentially switching languages as needed mid-conversation.

Crafting a voice AI agent, step-by-step

Tom then took us through a tour of Synthflow’s interface, showing us just how easy it is to set up an AI assistant tailored for specific tasks.

He walked us through creating a real estate assistant using a pre-built template that helps agents qualify leads, collect preferences, and even handle objections.

For each task, Synthflow enables users to set up an agent profile with background info, objectives, and a call script, which “gives the agency a guide on what it is to go through the call.”

After going through this basic setup, Tom showed us how to “make things a bit more interesting.” He listed various tasks the assistant can perform, like:

Sending a simple SMS after the call, with information about an appointment—or just thanking them for taking the call.
Sending a real-time booking link to schedule a live appointment during the call.
Transferring the call to a human, if the script determines that a particular condition has been met (e.g. the caller sounds annoyed, or specifically asks to speak to a human) and during working hours.
Extracting information from the conversation that’s then pushed into a CRM.

AI can go beyond what human agents can do, using real-time data

While using an AI voice agent might seem less personal than having calls answered by a staff member, Tom pointed out that AI agents can tackle tasks faster, doing things that humans can’t easily do.

A great example Tom shared was the integration of real-time weather data to enhance customer service. He demonstrated how a restaurant assistant could offer outdoor seating but, based on an API that checks local weather, suggest indoor seating if it’s raining.

“This happens in a split second and it's something that a human likely wouldn't do. So that's pretty cool.”

He went on to explain that you can connect any API to Synthflow, pulling information into the AI (e.g. the status of an order) and using that within the conversation.

Outbound calls and website chat using Synthflow

Synthflow’s functionality doesn’t stop at inbound calls. It supports outbound “batch campaigns,” allowing voice agents to automatically contact lists of prospects—an efficient tool for sales and follow-up calls. You can set the time period to use for the calls, e.g. 6.45 pm–9.15 pm.

The voice assistant can bring information back into Synthflow or push it into the CRM for later follow-up.

You can also implement Synthflow as an embeddable assistant via a widget on your website or app, available 24/7 for users. Tom suggested, “If you wanted to convert people or engage people on your site, or even have them make reservations or book appointments, you can do that with our widget.”

This widget-based assistant can be configured in the same way as your general voice assistant, offering a consistent user experience.

Wrapping it up

Whether you want a simple AI agent to answer calls and log details, or you’d like to automate your whole booking and appointment reminder system, Tom’s workshop is one not to miss. You can watch the full session (available to Pro Ben’s Bites members) here:

Liked this post? You may be interested in these: