Overview
This is a case study of one of my customers for whom I built and implemented a conversational AI chatbot. Hopin purchased Drift's Conversational AI as a part of their ~$130K contract to launch a chatbot on their website. Since they were new to AI chatbots and were looking to learn more about the goals of their site visitors they wanted to start by introducing a chatbot on a page that would capture the maximum percentage of site visitors.
GOAL: To build a chatbot that is informative, and helps their first-time buyers get the right resources so they're motivated to to purchase Hopin for their event. Most importantly, a chatbot that qualifies high intent leads and deflects low-intent questions to support.
CLIENT: Hopin, an all-in-one events management platform
MY ROLE: Conversation Designer
The Problem
Hopin had decision-tree (DT) chatbots converting audiences from paid ads, targeting specific campaign pages etc. where they worked effectively. However the DT chatbots performed poorly on their homepage that received the maximum site traffic at varying stages of their buying journey. Like IVR, the DT chatbots did not allow the site visitors to interact on their own terms. The 'one size fits all' approach failed especially when targeting the high traffic pages with a high volume of questions. Hopin's customers and buyers could only get qualified if they followed the scripted path by clicking on specific buttons. There were too many drops offs. Hopin was losing quality leads with strong potential for conversion, because not all site visitors necessarily wanted to wait to be connected to a human agent to get their questions answered. It was either that or too many unqualified leads were getting through to sales.
How Conversational AI Solves the Problem?
AI powered chatbots's biggest advantage is that it allows for visitors to communicate using free text. The bot could be trained to exact intent from open-text conversations to route only the high intent leads to assigned sales reps. immediately, while disqualifying or deflecting low intent leads to support. Here's what we knew based on implementing AI chatbots for other customers:
Discovery Research
My discovery research for Hopin began by identifying the gaps in ongoing conversations that Hopin's site visitors were having with the human agents by clicking through the button options in the DT chatbots.
The Conversational AI Framework
Timeline
The Build - Default Paths & Custom Experiences
Testing, Training & Iterating!
After conducting 2-3 rounds of testing, both internally as well as with the customer team, I launched Hopin's chatbot live on their website on Feb. 15 2022. I am continuing to train the chatbots on new topics while adding more utterances to existing topics to strengthen them. I do this by routinely reviewing their conversation logs, monitoring drop offs, engagement rates, email captures, meetings booked and topic usages on a continued basis.
Results
The CAI chatbot has influenced a pipeline of $2M in 3 months since launch. Hopin just renewed their contract with Drift for another year. Email captures have accelerated by 5.3% and meetings booked increased by 1.9%.
