AI meets CRM and BI: 15 Salesforce Einstein and Einstein Analytics announcements from Dreamforce 2019

Salesforce is pushing natural language and machine learning capabilities beyond the boundaries of its sales and service clouds. Here’s a closer look at the latest Einstein releases.

Today's workers want to be heard, says Salesforce's SVP of employee engagement The modern workforce wants to be part of a values-driven company where they're allowed to speak their mind, says Salesforce's Jody Kohner.

Nearly half of companies now say that they're using some form of artificial intelligence, according to McKinsey research that Salesforce shared at Dreamforce 2019, yet seven out of ten companies report they're seeing little to no value from their AI projects so far.

This gap exists, Salesforce executives asserted, because too many companies are struggling to overcome data-management and data-science technical hurdles, so they have yet to put AI into production at scale. Salesforce's alternative is Einstein and Einstein Analytics, both of which offer prebuilt, CRM-embedded predictions and recommendations along with the ability to build AI apps and answer company-specific business questions without coding or data science expertise.

There's no question that AI is of interest to organizations and that the Einstein and Einstein Analytics promise of pre-built capabilities and declarative, no-code/low-code app and model development is compelling. Yet Salesforce executives frankly acknowledged that Einstein and Einstein Analytics have only scratched the surface of potential adoption, with Sales Cloud use cases leading the way. Thus, what we heard about during this year's Einstein and Einstein Analytics keynotes was more out-of-the-box features, more built-in platform capabilities and prebuilt industry apps, more proven use cases, and more training and support options. Here's my rundown and take on 15 announcements.

Einstein Steps Up on Voice and Recommendations

your-smart-crm-assistant.jpg

Salesforce

The two themes emerging around Salesforce Einstein at Dreamforce where voice and recommendations, with new capabilities introduced or announced for the platform and specific clouds. Here's the rundown:

  • Einstein Voice. This platform-level service was introduced last year, but the Einstein Voice Assistant (mobile and web) user interface has been revamped and many more pre-built skills have been added across the Salesforce platform and specific applications. Sales Cloud users, for example, can talk to Einstein to qualify a lead, create an account, update an opportunity, log a call or create a contact. The Einstein Voice Skills component of Einstein Voice enables administrators to build and embed custom voice skills into any Salesforce app. Einstein Voice skills were demonstrated at Dreamforce on a smart speaker created in the likeness of the Salesforce Einstein character, but the device was a limited prototype that was just for show. Salesforce (along with dozens of other companies) has joined a voice interoperability initiative with Amazon through which its own "Hey Einstein" wake word will be enabled on Echo devices (with users adding the capability through the Amazon Alexa app). Once a user pairs their Salesforce profile to the device, they could then interact (securely and directly) with Salesforce using Einstein Voice though the Echo devices.
  • Einstein Call Coaching. This new feature for the Sales Cloud harnesses Salesforce natural language and AI capabilities to study call data at scale (with opt-in permission), extract insights on conversation trends. Sales and service managers and employees can use these insights for coaching, drawing on recordings from specific calls and outcomes to build a library of best practices that can be used to prep sales reps for calls related to the same topic.
  • Service Cloud Voice. Salesforce and AWS have integrated the Amazon Connect omnichannel contact center service with Service Cloud Voice, a new product designed to brings together phone, digital channels, and CRM data into a unified console. The console is a single workspace in which agents can manage customer data, review interaction histories, and deliver service across phone, email, chat and messaging. The Amazon Transcribe, Amazon Translate, and Amazon Comprehend AI services are employed to surface sentiment analysis, speech-to-text transcription, and translation into preferred languages through Service Cloud Voice. Salesforce Einstein then mines these real-time transcripts to give the agent recommended answers, contextual knowledge articles, and next best actions for the customer within the Service Cloud console.
  • Einstein Search and Translation. Extending Einstein natural language capabilities, the omnipresent Salesforce Search bar on most interfaces is now Einstein Search, with better contextual understanding and filtering of typed questions and more personalized results tailored to what matters at your company and how you work as an individual. Einstein Translation is a Salesforce platform-level service now entering pilot testing that will translate text from one language to another.
  • Einstein Recommendation Builder. Developers can already take advantage of platform-level Einstein Prediction Builder and Einstein Next-Best-Action services to create custom predictions and prescriptive actions. Einstein Recommendation Builder, now in pilot testing, will enabled developers to build custom recommendation engines while abstracting the complexities of the behind-the-scenes data science.
  • Einstein Article Recommendations. This recommendation engine for the Service Cloud helps resolve common service issues by recommending the best-fit article for each case.
  • Einstein Content Selection. This new recommendation service for the Marketing Cloud personalizes the content shown to respondents when they click through on email messages.
  • Einstein Designer. This new service for the Commerce Cloud, currently in pilot testing, helps firms develop and present variations of website content personalized for customers based on their purchase histories and click-through patterns. The idea is to boost response by moving  away from generic, one-size-fits-all website experiences.  

MyPOV on the Einstein announcements

Voice was clearly the highlight of the Einstein keynote, and the Einstein smart speaker demo (whether real or not) gave it a novel, anthropomorphic twist. Nonetheless, I expect that the vast majority of Einstein Voice interactions will take place on smart phones through Salesforce mobile apps.

One surprise announcement that deserves mention was the offering of 1 free prediction (beginning in February) for every organization with an enterprise license to Einstein Prediction Builder. The results of this prediction can be exposed to all licensed users within the organization. This is clearly a "try it, you'll like it" offer meant to promote adoption. In fact, many Salesforce customers are entitled to included Einstein capabilities or bundled licenses but are not yet using them. The impediment is often not knowing where to start. Salesforce's growing Trailhead community and educational offerings, such as the new "Einstein's Guide to AI Use Cases," are geared to demystifying AI and promoting adoption. But for some organizations, licensing costs are a concern and the reason for hesitation. My advice is to let business value be your guide. Take advantage of any built-in capabilities and bundled licenses to prove initial business value and see where it takes you.  

Einstein Analytics Blends BI and AI

bi-and-ai.jpg

Salesforce

It's early days for the Salesforce acquisition of Tableau Software, so the simple explanation of the company's analytics portfolio presented at Dreamforce (and a week earlier at the Tableau Conference) is that Einstein Analytics is for CRM-embedded analytics while Tableau is for Enterprise-Wide Analytics (while last year's Datorama acquisition is for marketing intelligence). In their keynote, Einstein Analytics execs noted that the embedding platform provides a single stack for BI and machine learning that focuses on optimizing for business outcomes rather than exploring data in hopes of finding relevant insights. Here a rundown on the biggest Einstein Analytics announcements.

  • Watchlist. Just as you pick the stocks you follow on you financial apps and web sites, this new Einstein Analytics feature, due in February, lets you pick the key performance indicators and metrics that you want to follow and Einstein automatically comes up with a dashboard view accompanied by ongoing trend analysis.  
  • "Ask Einstein." The formal name has yet to be finalized, but this feature, due for pilot testing in February, lets you type natural-language questions through web and mobile interfaces and get personalized responses relevant to your behavior and others like you in your organization. The feature learns from the way you work, including all the data you touch and all the business- and industry-specific metadata hardcoded into your Salesforce deployment. The promise is the ability to query without coding, taking advantage of learned context to better understand the patterns and intent of your natural language questions.   
  • Einstein Analytics Industry Applications. These analytic apps for specific industries include all the platform and AI licenses you need together with prebuilt industry-specific KPIs and best practices. The apps also deliver business-specific predictions and actionable recommendations. Analytic apps for Healthcare, Manufacturing and Financial Services are generally available while an Einstein Analytic App for Consumer Goods due in February.
  • Einstein Predictions. This new licensing bundle includes 10 mix-and-match predictions from either Einstein Discovery or Einstein Prediction Builder (the components of the top-tier Einstein Analytics Plus licensing level). Salesforce also announced an API for Einstein Discovery that makes it possible to embed predictions into external systems such as portal and bespoke applications.
  • Einstein Analytics What-if Scenario Analysis. This feature was not announced during the keynote, but sources tell Constellation that work is underway on scenario modeling and planning capabilities.
  • Einstein Analytics Data Catalog. Catalogs are all the rage. This one for Einstein Analytics includes the requisite shopping cart user interface and governance and data-lineage-tracking features.
  • Einstein Analytics Direct Data. The Einstein Analytics team touted its own auto-scaling, distributed data store, but they also acknowledged that many companies use third-party services. Direct Data access capabilities start with Snowflake and will soon extend to Amazon Redshift and Google BigQuery.

MyPOV on the Einstein Analytics announcements

Like the Einstein team, Einstein Analytics leaders pointed to multiple training and support assets, including Trailhead, a new Learning Map and upcoming Adoption and Learning Academies as ways to promote adoption. Here, too, I think cost is also an impediment for some firms, with the $150 per-user, per-month fee for Einstein Analytics Plus being an impediment, particularly when organizations have existing investments in BI and analytics tools and data management infrastructure. The new mix-and-match, $75 per-use, per-month Einstein Predictions SKU and Direct Data offerings are clearly designed to overcome these barriers and jump start adoption. The bet is that a taste of predictive insight embedded directly into Salesforce apps will win teams over to the combined BI and ML Einstein Analytics platform.

einstein-analytics-packaging-and-pricing.jpg

Salesforce

MyPOV on Dreamforce and the Salesforce Economy

Dreamforce is turning into two events in one. The larger stage highlighted Salesforce as the exemplar of what Marc Benioff has called "the new capitalism" in opinion pieces published in The New York Times and elsewhere. This was the side of the event that featured talk of values, sustainability goals, and corporate responsibility – not to mention all the celebrities and sessions on diversity, equality, and climate change. Not that I don't identify with many of these values or believe they're not genuine, but I suspect another motivation for highlighting these topics at Dreamforce is to attract employees.

Salesforce is now the largest employer in San Francisco, according to Benioff, but it has fierce competition for workers. Studies show that younger workers, in particular, want to know that their careers have a higher purpose and that they identify with the values of their employers. Further, many customers are also choosing to do business with companies that focus on more than just profits. Salesforce is certainly setting itself apart in a Silicon Valley culture that's too often focused on funding rounds and market valuations.  

The second, more traditional aspect of Dreamforce is the product-, technology- and innovation-focused CRM event, which is the part I was there to see. The Einstein and Einstein Analytics keynotes delivered plenty of announcements tweaks to packaging and pricing. Both presentations were also laced with real-world examples and validations featuring, or even presented by, customers including State Street Global Advisors, PWC, Schneider Electric and Indeed.com. There clearly is a Salesforce economy and Dreamforce has become more than just an annual gathering for customers and partners building on the platform of CRM.  

Related stories: