Ruby Labs automates 4 million support interactions per month with a 98% resolution rate

We get a lot from the apps on our phones.

They tell us the weather, our horoscope, and which food option fits our fitness goals.

But there’s a massive challenge for the companies that run these apps: customer support.

If you put out a product that a million people use, you’re going to have to up your customer support game.

And if you manage 6 popular apps? You’re probably going to need AI.

The Client

Ruby Labs is a software company specializing in customer-end applications — their suite of apps ranges from health coaching, to astrology, to self-directed tutoring.

Their first major success, the health-coaching app Able, implemented Botpress for customer support after receiving over 500,000 users within the first several months of its launch.

After witnessing the success of AI customer service with Able, the Ruby Labs team decided to expand their Botpress AI agents to their entire suite of app-based businesses.

The Challenge

When they first started out – as a small two-app company – Ruby Labs could manage their customer support with employees alone.

“But after we scaled and brought in more products, it would have been impossible to facilitate subscription management with employees,” explains Alex Bogdan, Head of Support at Ruby Labs. “We would have needed 100 people.”

Instead of hiring 100 people, they streamlined it with a handful of AI agents.

Bogdan and his team were looking to fully automate their tier 1 support functions.“We needed something that could let users cancel subscriptions, get refunds, get technical help, and answer any other questions they might have.”

To provide a satisfactory experience for the customers, Ruby Labs was looking for a solution that could balance cost, value, and customer satisfaction.

It would need to answer questions, manage technical needs, manage subscriptions, and provide refunds after cancellations.

Finding the right AI solution

The team at Ruby Labs experimented with different AI solutions until they found Botpress.

They tried a chatbot built directly on an LLM, but it didn’t offer the blend of company knowledge and natural conversation that are necessary for successful customer service. They tried another AI chatbot platform, but found that it wasn’t extensible enough to fit their needs.

But in Botpress, the team found the best of both worlds: “Botpress provided the best end-user and builder experience.”

The Ruby Labs Support Team especially liked the user support they were able to receive from our Customer Success and Solution Engineering teams. “The Botpress team was so kind and communicative,” Bogdan explains. “They made us feel like part of the team. Our point of contact was happy to jump on calls – he even helped us implement certain features.”

Scaling customer support with AI agents

Because of the sheer scale of their users, the Ruby Labs Support Team wanted to maximize their ability to scale increasing customer requests.

“You no longer need a full team of devs to make an app – you can do it with 1 or 2. You can do a lot more with fewer people and more automation. We wanted to be at the forefront of this movement," explains Bogdan.

The team was able to almost fully automate their customer service — while maintaining a smooth flow for their users.

Accessing services and support

Now, when a user of one of the six Ruby Labs apps needs support, they just have to interact with the chatbot embedded in the company’s help center or contact us page. After opening the widget, the user is presented with four options:

Cancelling an account
Asking billing questions
Troubleshooting technical issues
Asking general questions

The chatbot is attached to the company’s Knowledge Bases, so it easily doubles as an IT support bot and an FAQ bot.

Accelerating subscription management

If a user wants to downgrade, upgrade, or cancel their subscription, all they need to do is input their email address and verify their identity by clicking a link sent to their email. Then they’re eligible to receive a direct refund.

Before cancelling any subscriptions, the chatbots will offer discount deals to entice users to stay. These discounts bring Ruby Labs an extra $30,000 every month.

And if a user is confused by the refund policy? Then the chatbot can direct them to the company’s official refund policy documents.

Identifying high-risk users

Just like a good employee, Ruby Lab’s AI agents know that not all users are the same.

Every time it receives an unsubscription request, the agent will check its Stripe integration to investigate their payment history.

If the agent observes a pattern of requests that suggests fraudulent behavior, it can opt for a different flow than typical users.

This informed discretion allows the AI agent to handle complex edge cases – like fraud – rather than following cut-and-paste resolutions for every user.

A fully automated customer support system for millions

Ruby Labs succeeded in their goal to scale customer support by centring AI in their operations.

Their suite of AI agents has a 98% resolution rate – meaning only 2% of users reach out to speak to a human after using the chatbot.

With a resolution rate like that, Ruby Labs was able to reorganize their customer support team to other positions and fully automate their support processes.

AI is redefining how companies scale customer support.

Ruby Labs proves that with the right approach, businesses can handle millions of users without ballooning operational costs.

With their AI-powered foundation in place, Ruby Labs is set to scale even further as they look for new apps to add to their roster — without the traditional growing pains of customer support.

Table of Contents