AI is improving customer service experiences — here’s how

IBM Think AI

Enhancing customer experiences continues to be a major priority for businesses, as only 36% of contact centers are currently able to track a customer journey that spans multiple channels.[1] And the stakes are high — 70% of consumers say they prefer messaging over calling customer support. Yet of the 270 billion service calls made annually, 50% of first calls go unresolved.[2]

Fortunately, companies like Autodesk are using technologies like artificial intelligence (AI) to make poor customer service a thing of the past.

Autodesk employs a staff of 350 internal and external customer support agents who manage more than 1 million customer and partner interactions every year. Roughly half of those questions are simple customer queries that can easily be identified and answered — such as code requests, changes of address, contract problems, technical issues, activation code requests, and so on. Unfortunately, every time a customer service agent spends time helping one customer with a simple task, another customer must wait for them to finish that task in order to receive service.

With the AI capabilities of IBM Watson conversational technologies, Autodesk is automating the resolution of simpler tasks, freeing up customer service staff to focus on more complex questions. Autodesk used Watson to build their own Autodesk Virtual Agent (AVA) to quickly identify and answer their most common customer queries — all day, every day. AVA recognizes the intent behind customer queries with Watson’s natural language capabilities, uses keywords and phrases to parse the conversation’s context, and returns high-confidence answers quickly and effectively. Since implementing AVA, Autodesk has increased their response time by 99% while increasing customer satisfaction 10 points. That means they’ve reduced their average response time from 38 hours to just 5.4 minutes for most inquiries.

For every future customer interaction, AVA will essentially be the first stop. Since AVA has been trained extensively on Autodesk’s consolidated body of knowledge, AVA can easily recognize different types of customer queries and either provide an answer or redirect the question to customer service staff. AVA’s original catalog of knowledge came from chat logs, use cases, and forum posts. But as of today, AVA has analyzed more than 14 million sentences for keywords, entities, phrases, clusters, and other telling speech and language patterns.

As it operates, AVA will continue to process conversations and expand its understanding of the nuance, meaning, and context associated with its customers’ queries. That’s the beauty of AI — it never stops learning. Over time, AVA will become more and more capable of handling different customer questions by drawing from the expertise of an ever-growing number of internal resources.

These days, we have near-immediate access to just about anything we could want, but useful help has always been a bit more elusive. Finally, with the AI capabilities of Watson, Autodesk is helping to change the way companies answer our most pressing questions.

This is just the tip of the iceberg. AI is actively changing the world in so many ways right now. At Think 2018, you’ll have the chance to talk directly with AI specialists about these capabilities and so much more.

Think 2018 takes place March 19-22 in Las Vegas. To learn more about AI at Think — plus cloud and data, IT infrastructure, security, IoT, blockchain, and more — visit the conference website.

This post is sponsor content from IBM and was created by IBM and Insider Studios.


[2] US “Switching Economy” Puts Up To $1.3 Trillion of Revenue Up for Grabs for Companies Offering Superior Customer Experiences, Accenture Research Finds


Join the conversation about this story »