Salesforce Shuts down Einstein Voice Assistant; Confirms Strong Links to Amazon Connect

A very short release note issued in July 2020 marked a quiet end to Salesforce’s Einstein Voice Assistant. The end-of-life notice, effective as of July 10, applied to the mobile app for iOS and Android-based smartphones, as well as the “Setup for this feature in Lightning Experience.” The Einstein Voice Assistant was introduced with Salesforce’s characteristic fanfare at Dreamforce 2018. Its relatively quick demise surprised almost nobody with serious plans for voice-based Intelligent Assistants because, upon close investigation, it was a shallow UI that was little more than a conversational skin for existing applications.

In 2018, Einstein Voice Assistant was part of a much more expansive role that Salesforce described for Einstein Voice Services. At the time, Einstein was positioned as a “smart CRM assistant across sales, service, marketing commerce and more.” The expectations were very high. Reality, in the form of commoditization, set in by 2019. It became clear that automated speech recognition (ASR) and text-to-speech (TTS) are mere building blocks that virtually all developers expect to incorporate into their Voice User Interfaces (VUIs). The power shifted to tools and services that simplify the development efforts of enterprise employees, system integrators and business process outsourcers.

The Amazon Connect Connection

As the third quarter of 2020 approaches, Salesforce’s decision to deprecate Einstein Voice Assistant doesn’t mean that it is abandoning its “Voice Rules!” attitude. Quite the contrary. The giant of cloud-based CRM has realized that the future of both Conversational AI and UI reside in the giant of e-commerce’s own service cloud. To be fair, it had explicitly revealed this roadmap at Dreamforce 2019 in this document describing “Hey Einstein” as the wake up word that brings voice search and command-and-control “tailored to any role or industry”, giving every customer or employee his or her own voice guide. Salesforce’s Service Cloud Voice is the foundation for “Hey Einstein”. The tools and pre-configured “Einstein Voice Skills” provide the primitives for an admin or developer to create voice assistants to perform such tasks as updating a field, creating a task or reading out a forecast.

The developers using Salesforce Service Cloud Voice have come to appreciate deep integrations into Amazon Web Services (AWS), specifically its a la carte Contact Center as a Service (CCaaS), Amazon Connect took top billing in anticipation of enterprises’ need for services that combine old-guard voice processing, call processing and  NLU/ML with a portfolio of “AI powered speech analytics”, a set of capabilities that included transcription, translation, categorization and sentiment analysis… all included in AWS’ cloud.

Even though we live in a “multi-cloud” world, it is more and more obvious that Amazon Connect is Salesforce’s preferred CCaaS provider. This release makes it clear that, even though Salesforce sees value in providing enterprises with their choice of providers for National Language Processing, Machine Learning and Analytics, AWS has preferred position. In so many words, “Amazon Connect is also providing powerful AI-powered speech analytics, using Amazon Transcribe, Amazon Translate and Amazon Comprehend, to surface sentiment analysis, speed text transcription, and translation into preferred languages, directly to agents through Service Cloud Voice.”

The connection to Einstein is also explicit. “Salesforce Einstein uses information from these real-time transcripts to give the agent recommended answers, contextual knowledge articles, and next best actions for the customer, all within the Service Cloud console. Clearly, AWS has agreed that Salesforce had won the battle for real estate on each agents screen and that Einstein has a role to play as a source of information and advice to agents.

Viewing Customer Data Through Amazon’s Contact Lens

Meanwhile AWS has stepped up development of a new AI-based analytic offering called Contact Lens, designed to provide access to such functions as transcription, insights (key phrases) or redaction with the click of a “check box” in the Amazon Connect Control panel. Instead of having to invoke Amazon Lex (for speech recognition, text-to-speech and NLP) or shopping for individual “Lambda Functions” to carry out specific tasks, Contact Lens combines selected functions to serve specific use cases.

One of the promising use cases that Amazon execs described to us is “call disposition.” It is a horizontal application for post-call processing of transcribed content. Rather than taking up an agents time to compose descriptions of the purpose of a call (usually selecting from a checklist of outcomes), Contact Lens prepares one based on its transcription. It can provide a more accurate account of the topics covered, assurance of compliance with laws, rules and policy, and a listing of any commitements that an agent has made on behalf of the company.

TTEC Acquiring An Amazon Connect Accelerator: VoiceFoundry

As if on cue, one of the longest-standing Business Process Outsourcers (BPOs) in the world is acquiring a company that is dedicated to simplifying and speeding  Amazon Connect integration for global enterprises. Denver-based TTEC enters an agreement to purchase VoiceFoundry to integrate into its Humanifytm CX Cloud. VoiceFoundry had been part of Waterfield Technologies portfolio of companies. In was among the first communications consultants and integrators to focus on developing personnel, best practices, tools, APIs and connectors specifically to simplify and accelerate migration to Amazon Connect.

Waterfield’s experience with building connectors from enterprise contact centers to AWS goes back more than three years, which means it pre-dates the formal launch of Amazon Connect. It became one of the first Amazon Connect resellers on a global basis and built a staff of designers and consultants responsible for some of the first integrations with Salesforce, ServiceNow, ZenDesk and other popular platforms. It employs a set of migration tools that, according to the TTEC press release, no comprise an enterprise-grade API integration and deployment platform, including Humanify’s intelligent administration, orchestration, analytics, and automation solutions.

Most important is the new path to market and the CRM cloud giant’s role. TTEC’s Humanifytm Cloud will “seamlessly leverage the power of Salesforce’s Service Cloud + Amazon Connect to offer out-of-the-box contact center solutions sold by Salesforce” (the italics are mine).

It’s All About CXaaS and CSA Now

There’s a new competitive landscape emerging in the “CX as a Service” World. Opus Research will be issuing a quick succession of posts about the competitive responses from Genesys (Engage Cloud), Cisco, NICE, Five9 and others. As the focus rightfully moves to resources that speed up delivery of automated Intelligent Assistants that span multiple channels and modes, expect both competition and innovation to come from the Conversational Service Automation (CSA) domain, including Uniphore and Thrio.



Categories: Intelligent Assistants