This syndicated post originally appeared at No Jitter - Recent posts by Zeus Kerravala.

Democratizes artificial intelligence services so any
customer can easily build AI-powered CRM workflows.

This week at Dreamforce 2017, Salesforce is launching its latest and greatest innovations aimed at enabling businesses to do more with the data stored in its CRM platform. The most interesting to me are the new capabilities it’s introduced for its year-old artificial intelligence (AI) platform, Einstein.

Salesforce, despite its size and regardless of how innovative it’s been in the past, was late to the AI game. Rather, it had left innovation in this area up to partners such as Altocloud, Five9, and Talkdesk. But now, with a services platform it calls myEinstein, Salesforce has moved to democratize AI so that companies of all sizes can leverage the technology.

myEinstein services let users and developers build applications that can change the way organizations interact with their customers. As well documented on No Jitter and elsewhere, modern, digital businesses will compete on the basis of customer experience. However, the definition of great customer experience is changing. Having polite, well-trained contact center agents is table stakes, not a differentiator. Businesses need to rethink customer service and create experiences that make customers go “Wow!” — and myEinstein is intended to do just that.

Designed to be easy to use, myEinstein features a setup guide that takes users through the process of building and deploying AI features using structured and unstructured Salesforce data. Admins build the logic and structure, and myEinstein does the heavy lifting, automating the process of building models and scoring data. These models, embeddable into Salesforce workflows, are constantly learning — the more they’re used, the better they get.

Predicting Business Outcomes

The first myEinstein service is Einstein Prediction Builder, which lets admins create AI models that can predict business outcomes. For example, a bank could monitor Salesforce data for a change in activity, a link to a new account, or other factors that indicate a customer is about to close an account. With this predictive data, a bank representative could reach out to the customer in an effort to prevent the churn. Or, a mobile phone provider could monitor for changes in calling patterns or data usage, and proactively offer new plans to customers rather than running commercials saying, “We’re only 1% worse than the other guys!”

With the declarative graphical Prediction Builder setup tool, users can define predictions and identify which fields to build models on and which data to use. They can set an attrition score, based on predetermined signals, and embed it into the Salesforce account page. From there they can create a task alerting reps to take action for high-risk customers.

Automated Outreach

myEinstein also includes Einstein Bots that let admins quickly build, train, and deploy custom service chatbots. Like all bots, the Einstein ones use natural language processing to communicate with customers on basic requests, in theory freeing up agents to focus on more complex or higher value inquiries. Businesses can tie Einstein Bots to workflows and train them with historical data for instantaneous response to customer inquiries or hand off to an agent.

I have mixed feelings on bots. I used “in theory” above as I’ve talked to many businesses that have had bad results from initial bot deployments — not because the technology is bad but because they’ve used it in the wrong way. Bots should only be used for simple interactions that are highly repetitive. Anything unique or complex should be routed to a human agent. Good use cases for bots are tasks like password resets, tracking orders, or answering basic information. Bots have their place, but they’re not for everything.

Training Models

Lastly, myEinstein includes the Einstein Language and Einstein Vision APIs that Salesforce introduced in June. With Language for Intent, developers will be able to train a model to classify an inquiry’s intent and automate routing to the right place. For example, an inquiry on pricing could lead to an interaction with a salesperson, whereas a more general inquiry would lead to a marketing page. Language for Sentiment will let developers classify tone of a text as positive, negative, or neutral so businesses can react appropriately.

Einstein Vision for Image Classification will bring visual search, brand detection, and product identification competencies to CRM apps. Using Vision of Object Detection, developers will be able to train models to identify different objects in images.

The myEinstein news brings the number of AI-focused announcements in a little over a week to four, the others being Cisco’s Spark Assistant, Avaya’s AI developer platform, and Talkdesk for Sales. In case you had any doubt, the AI era has arrived — and, as these announcements show, with such variety that businesses have the opportunity to apply AI any way they want. With myEinstein, any Salesforce customer can easily build AI-powered CRM workflows.

The following two tabs change content below.

Zeus Kerravala

Zeus Kerravala is the founder and principal analyst with ZK Research. Kerravala provides a mix of tactical advice to help his clients in the current business climate and long term strategic advice.

Latest posts by Zeus Kerravala (see all)

Share This Post:
No Interactions
 

Be the first to comment!

Post a Comment:

You must be signed in to post a comment.

Insight and Influence Through Social Media
ZK Research: Home
Google+
Twitter
LinkedIn
Facebook
RSS Feed
ZK Research is proudly powered by WordPress | Entries (RSS) | Comments (RSS) | Custom Theme by The Website Taylor