Before Voiceflow—How Intercom fell short
Before Voiceflow, Roam used Intercom’s AI chat tool. It fell short for three core reasons:
- Lack of customizability: The out-of-box AI tool never took over the conversation properly. Instead, the user had to go through many layers of conversation before they were offered a solution. This didn’t achieve their goal of reducing tickets in their support inbox.
- The pricing model: Intercom charged based on resolutions provided by their AI chat. So even if the team found that the agent gave answers that weren’t the best or correct, Roam was still getting charged for them.
- Workflow mismatch: Nick’s team found that Intercom was hard to test and iterate quickly. “I would set something up and then I'd realize it didn't connect to a piece of conversation,” Nick explains, “Unless everything was connected, you couldn't publish it. So I would make these concessions to get it live, but the agent ended up a bit half-baked.”
Jump in at 5:16 on the full interview to learn more about Nick's transition from Intercom to Voiceflow.
The solutions—A template for an AI agent and a trained knowledge base
Nick and his team used Voiceflow to launch an AI chat agent onto their homepage in a matter of days. Here's how they did it:
- Curated a knowledge base (KB): Using Roam’s collection of help articles and FAQs, Nick uploaded resources to a knowledge base for the AI agent to pull relevant answers to common questions. Using the KB, they were able to completely automate their Level 1 support.
- Built an AI agent using a Voiceflow template: Nick got his AI agent up and running quickly by personalizing existing conversational flows from the Voiceflow template library.
- Added chat bubble to homepage: Using a simple copy-and-paste of Voiceflow’s web chat snippet, Nick added the AI agent’s source code to Roam’s homepage.
- Iterated his assistant with rapid testing and transcript reviews: Nick found success with his agent through prototyping his designs, adding new prompts based on live user transcripts, and tweaking his FAQ knowledge base. “We often have an idea of how we think customers perceive our business and the questions they might ask. But the better way is trial by fire,” says Nick.
- Leveraged the analytics dashboard: The team used the top intents view in the Voiceflow analytics dashboard to inform product decisions and gauge where they needed to double down. Nick found that booking your pickup and cancellation were top intents for his agent so quickly brought back to his team, “Are we doing a good enough job onboarding customers to make them understand what the cancellation or the return requirements are or where they can go to access that information?”
“We don't have a team dedicated to this, but it's already working. The ROI on Voiceflow is ridiculously good. And the best part is, we’re just getting started.”
The impact—Customer support time, saved
Since deploying Voiceflow, Nick’s team has saved invaluable time and effort educating prospects on their services. Since the launch of their digital AI agent, the team at Roam has seen:
- A drastic decrease in inbound calls: They’ve saved over 30 hours of customer support hours per week, freeing up time to address more urgent customer needs without additional CS staff hours. “I can’t put a price tag on the brain power and processing that Voiceflow saves us. We get overwhelmed sometimes but It's nice to know that now there's a ‘person’ on the team who will never run out of steam,” says Nick.
- An increase of sales calls booked asynchronously: Nick’s team has experienced a dramatic increase in async sales meetings booked through their digital agent in the short few months it's been live. Roam’s chat agent provides the transcript of the conversation, which offers helpful context to help them prepare for high-value meetings.
- Accurate and comprehensive answers powered by AI: Whether users are inquiring about leasing a car, signing up for a service, learning more about Roam’s offerings, or moving through the sales funnel, Roam’s digital agent answers user questions and shares useful information. By using a knowledge base, Nick’s team trusts that their digital agent will only provide approved and accurate outputs based on their resources, FAQs, and website.
- Efficient launch and easy maintenance: “Voiceflow has a very simple interface and when you use a template, 80% of the work is done for you,” Nick explains, “This is huge as a small business. I'm not making sacrifices in other areas because I can trust the output of our digital agent and simply make changes or tweak things.”
- Insight into customer needs: Before, Nick’s team struggled to understand why users would call into customer support and abandon their requests. Now, Nick has insight into what users are asking about, when they fall off the sales funnel, and what their common questions are. Using this data, Roam is able to pivot their marketing based on users’ pain points, or onboard customers better to address common questions.
“Voiceflow is elastic to each customer. We started with level 1 customer support inquiries and we have the tools to scale our AI agent to handle more advanced customer support tasks.”
.svg)