Salesforce debuts two new AI voice assistants atop its Einstein platform, Einstein Voice for enterprises and consumer-facing Einstein Voice Bots
Amazon's got Alexa. Microsoft has Cortana. And Google has the Google Assistant. It seems like practically every tech giant has its own voice platform …
Context & Ripple Effects
This is Salesforce extending the Einstein AI platform it built in 2016 beyond embedded scoring and predictions into a voice layer — one assistant aimed at enterprise users inside its clouds, and Einstein Voice Bots aimed at companies deploying consumer-facing voice apps. The framing matters: rather than building a home device to challenge Alexa head-on, Salesforce is putting voice where it already owns the customer data.
The move also reads differently in hindsight. A year later Salesforce revamped the assistant and folded telephony into Service Cloud via Cloud Voice, and by 2023 the voice push had evolved into Einstein Copilot, a chatbot pilot across its apps — making this 2018 debut the first step in a steady shift toward conversational interfaces over CRM data.
First-order effects
- Enterprises running Sales Cloud and Marketing Cloud gain native voice access to their CRM data without buying into a separate assistant ecosystem, while brands can spin up consumer-facing voice bots on the same Einstein stack.
Second-order effects
- Amazon, Microsoft, and Google — whose assistants compete mainly for the home and device surface, per the related Echo/Alexa coverage — now face an enterprise rival whose advantage is sitting directly on the customer records businesses already maintain.
Third-order effects
- If the pattern holds, business voice assistants consolidate around the platforms that hold the underlying data rather than around consumer hardware makers — a trajectory Salesforce itself followed as voice gave way to chat-based copilots across its product line.
The trend: Enterprise software vendors are embedding conversational assistants directly atop their own data platforms, contesting the assistant layer with consumer giants on workflow ground rather than in the living room.