AI Automation for Voice-based Lead Generation
In April 2024, I worked on a project where I developed an automated voice–based quiz to help quiz takers understand how they could improve their gym consistency. At the end of the quiz:’
● Each candidate received a personalized gym readiness score as well as guidance on what to do next to improve their physical fitness.
● The Quiz Creator (a gym owner) received the quiz takers contact information as well as answers to each of the questions in the quiz and their final score, all in real–time into a dashboard hosted in a spreadsheet.
This personalized assessment was a very powerful lead generation mechanism because most people have a high interest in knowing how good they are and how they can improve at something they care about.
By answering the questions in the quiz, not only do the candidates learn new things about themselves, the quiz creator is also able to access valuable insights as to how to position their products and services and to make personalized recommendations to the new leads thus exponentially improving sales conversion rates.
The click through rate on the test version of this campaign was staggering – a 94.4% CTR is quite unheard of for lead generation campaigns, traditionally!
The intriguing and most satisfying part of this project was that the voice based quiz was deployed natively and intentionally on WhatsApp. Unlike anything I had seen before, I was determined to build this WhatsApp Bot so that it could “speak” and also receive voice based input from the quiz takers.
It was incredibly fulfilling for me as I had to invent a way to push the boundaries of the norm with WhatsApp Bots.
I leveraged the APIs from:
● PlayHT (to convert quiz questions to speech)
● OpenAI’s Whisper Model (to convert the candidates’ voice responses to text)
● WhatsApp Business (to deliver the verbal quiz, drive CTAs and receive inputs in an interactive manner)
● Make.com (to dynamically receive and send inputs from the quiz taker into Open AI’s whisper Model
● Google Drive to store all audio files received from quiz takers, and
● Google sheets to provide the quiz funnel creator with a real time dashboard of all user response audios, transcripts, sentiment analysis and contact information.
I took advantage of ManyChat’s visual conversational logic builder to tie all the above together and deploy the Voice–based quiz.
To this day, this
is the marketing
automation project I am most proud of as
it created a profound impact
on both end users and the business owners while forcing me to
learn, invent and to
push the envelope
of what’s common.