Salesforce debuts Einstein Voice Skills for customizing your CRM voice assistant

Meanwhile, Salesforce is helping sales and service teams get more from their call transcripts with features based on natural language processing.

Kicking off Day One of the Dreamforce conference in San Francisco, Salesforce on Tuesday announced a series of new voice-related capabilities across its platform. With the new Einstein Voice Skills, developers and admins can build custom voice-powered apps for employees, regardless of their role or industry. Meanwhile, Salesforce is adding new features powered by natural language processing that will help sales and service teams get more value out of call transcripts.

executive guide

What is AI? Everything you need to know about Artificial Intelligence

A guide to artificial intelligence, from machine learning and general AI to neural networks.

Read More

The new capabilities come a year after Salesforce first introduced Einstein Voice and about five years after Amazon's Alexa brought voice UIs to the mass consumer market. Salesforce is positing that voice and natural language processing are about to make a big impact in the enterprise space.  

"Voice is a huge shift for the industry and will be as impactful in businesses as it's been in our homes," Salesforce President and Chief Product Officer Bret Taylor said in a statement. "With Einstein, Salesforce is bringing the power of voice to every business, giving everyone an intelligent, trusted guide at work."

According to Salesforce research, 75 percent of business buyers say emerging technologies such as voice assistants and chatbots are changing expectations at work. Meanwhile, IDC forecasts that one-third of enterprises will use conversational speech technology for customer engagement by 2022. 

Einstein Voice Skills builds on last year's rollout of the Einstein Voice Assistant, which can help sales reps with routine CRM transactions like updating customer records. With Einstein Voice Skills, companies can build voice-powered apps to replace any type of manual data entry or manual Salesforce navigation. 

Starting at the set-up page, a user or developer can choose the CRM action they want to build an app for, such as updating a field or creating a task. From there, they can select the Salesforce fields or objects that inform each action. For instance, an admin could build an app for field technicians, enabling them to use a voice UI to review a customer's service history. 

Admins can control how information will be read to the user. The voice assistant can offer next steps or follow-up tasks as part of its response. Admins can also control the channels and devices from which the app is accessible. Earlier this year, Salesforce joined the Voice Interoperability Initiative with Amazon and several other leading tech companies (though Google, Apple and Samsung are not participating) to encourage compatibility between voice assistants and devices. 

Meanwhile, Salesforce is also applying natural language processing (NLP) to call transcriptions in new ways, helping sales and service teams get more value out of their transcripts. For instance, a new Sales Cloud capability called Einstein Call Coaching uses NLP to help managers glean insights and spot trends from conversational data. For instance, the feature can identify keywords in sales call transcripts to alert a manager if there's a spike in competitor mentions. 

NLP also comes into play in Salesforce's new Service Cloud Voice, which integrates telephony natively inside Service Cloud. The new product can integrate with transcription services that convert voice to text in real time, enabling Einstein to provide recommended responses, knowledge articles and next best actions to service agents while they're on calls with customers. 

Prior and related coverage: