Ada’s Only Getting Smarter: 4 New Machine Learning Features to Enhance Your CX

With Conversational AI, businesses are talking to their customers more than ever before. In fact, Ada automated over 1.5 billion customer interactions in 2020 alone. And while we’re enabling businesses to scale their customer conversations to record breaking heights, we’ve always maintained a priority on making sure businesses weren’t just having more conversations, but more valuable conversations.

AI is unlocking value at every moment across the customer journey. At Ada, improving our industry-leading AI isn’t limited to only improving your customer’s experience, but everyone who works in our platform as well. That’s why we’re launching 4 new machine learning features to improve the experience of everyone that touches Ada’s product—sales and support agents, customers, and the bot builders behind the scenes.

Keep agent handle time low with Transcript Summaries

Let’s put on our customer hat. Think of a time you reached out to a company for support. You communicated for some time with a bot and maybe realized your question could be best answered by a human. When you were handed off, your agent had to get caught up to speed. Maybe they spent even more time reading through your conversation with the bot for context, leaving you waiting even longer. Or perhaps they simply started from scratch - leaving you to repeat what you’ve already told the bot.

It’s a frustrating experience all around. For customers, who want faster answers to their questions. For agents, who can’t do their job as efficiently as they like. And for businesses, who care deeply about customer and employee satisfaction.

We’re launching Transcript Summaries to provide agents with short, impactful AI-powered summaries of customer conversations at handoff to reduce customer and agent effort, decrease average handle time, and improve customer satisfaction.

Transcript Summaries 2

What can you expect? Ada’s Conversation Summarization Model works to summarize the entire conversation, pulling out up to 3 of the most relevant customer messages to get to the heart of the customer’s intent. At handoff, live agents have the option to view the Summary, and if need be, the full transcript is still available to dive into more detail.

Save bot builder time with Conversation Summaries

We wanted to expand on giving Ada clients the ability to easily review the conversations their customer base is having with their bot. Conversation Summaries will now provide bot builders with quick AI-powered summaries of every conversation that takes place in a businesses’ bot.

Much like Transcript Summaries, this feature leverages Ada’s Conversation Summarization Model to surface up to 3 messages that capture the contextual intent in the conversation.

Convo Summaries ML Blog

What does this mean for builders? They now get an at-a-glance view of customer conversations, making interactions easy to review and action on evolving customer needs. We’re giving bot builders back time, so they can focus on high-value work and continue to drive exceptional CX. 

Reduce customer effort with Smart Detect

Let’s say you meet a new coworker who introduces himself and tells you he’s from Barcelona. You continue to chat and later in the conversation he tells you he’s due for a visit to see his family back home. You wouldn’t need to ask him where that is - he’s already told you! In fact, he might feel frustrated that you haven’t been listening to him.

Your customers feel that same frustration when they have to type the same information again and again as they converse with your bot.

That’s why we’ve invested in Smart Detect. This feature automatically and preemptively identifies key data—like names, locations, dates and times— that your customers have already mentioned in their conversation with you.

Smart Detect ML Blog

Introducing natural elements of our everyday human conversations continues to improve the customer experience. Customer effort is reduced when they don’t have to continuously repeat themselves, resolution is accelerated, and they get the more personalized experiences they’ve come to love.

Fill content gaps fast with Training Insights

As you may have realized, the CX team behind your bot is incredibly important. While marketing, sales, and customer success professionals put their brains together to anticipate their customers’ every need, it’s inevitable that a customer will still pop up with a unique inquiry. When the bot isn’t trained for this, a pre-built “Not Understood” answer is returned to the customer and they’re handed off to an agent for additional help. Ada stores and tracks the questions that aren’t understood by the bot.

Any good bot builder wants to adapt and learn to best serve their customers. At Ada, we try to help this by presenting ‘Not Understood’ questions for bot managers to assess what content gaps exist, and what new flows should be prioritized to meet customer needs.

Training Insights ML Blog

We wanted to make it even easier to find patterns, and fill content gaps faster, without ever leaving our platform. By introducing Training Insights, similar ‘Not Understood’ questions are automatically clustered by intent. These groups of questions can then be easily trained to a new or existing Answer.

This plain view of content gaps makes it easy for CX teams to stay on top of customer trends, unique to their own business. The ease at which builders can now train these missing flows means they can also act quickly to improve their customer experience.

New call-to-action

Lynn Pine
Lynn Pine

Lynn’s career has spanned across different kinds of content, from copywriting, to journalism, to marketing, and even mystery puzzle games. She brings facets from all these disciplines into her work at Ada. Outside of that, Lynn loves playing games, hiking, and reading about trees.

More info about Lynn Pine: LinkedIn

Share This Post