Salesforce rolls out new low-code services for building AI-powered features

As an expansion of Einstein Platform Services, Salesforce admins and developers can now build AI-powered tools for translation and optical character recognition.

Salesforce on Wednesday announced an expansion of Einstein Platform Services, which is designed to help developers tap into Salesforce's Einstein artificial intelligence platform. Previously, Einstein Platform Services comprised APIs for vision, sentiment and intent. Now, it includes low-code services that let Salesforce administrators or developers build AI-powered tools for translation and optical character recognition (OCR). 

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

Einstein Translation enables developers and admins, using Salesforce's Apex programming language, to set up automatic language translation of any Salesforce object or field. For example, an admin could use it to translate inbound customer queries to the native language of the service agents. The service translates between 21 different languages, which should be especially helpful for Salesforce's customer base of multi-national corporations.  

Meanwhile, Einstein Optical Character Recognition leverages computer vision to analyze documents and extract relevant information. A tool built with Einstein OCR could, for example, make it easier for sales reps to extract information from a business card and send it to the correct Salesforce records.

Salesforce users can find other, third-party translation and OCR tools on the Salesforce App Exchange. However, using them means handing over your company's data to that service.

"As trust is our number one virtue, we'd like these developers to allow their users to keep that data in Salesforce," Marco Casalaina, VP of product management for Salesforce Einstein, said to ZDNet.  

The new tools are part of Salesforce's effort to offer a range of developer services, as well as low-code tools, so that organizations don't have to recruit new talent to build enterprise apps. Einstein Platform Services are geared toward "trailblazers" -- Salesforce administrators and developers trained on the company's online learning platform, Trailhead. By 2022, there will be 10 million trailblazers, Casalaina said.

In addition to expanding Einstein Platform Services, Salesforce announced Wednesday that it's launching the new Einstein Predictions Service. This lets customers bring prediction capabilities built in Salesforce to third-party services.

Salesforce previously rolled out Einstein Prediction Builder, which allows admins to predict a business outcome by building AI models on any Salesforce field or object -- custom or standard. With the new service, admins can take that capability to, for instance, an ERP or HR system outside of Salesforce.

Salesforce on Wednesday also announced new capabilities that help customers use AI responsibly. For instance, users can add "protected fields" when building AI tools to warn of potential harmful bias in datasets -- including second-order bias. This is useful if a user wants to ensure their AI models don't consider factors like race. They could remove race as a variable within Einstein, and they'll also get a notification of other variables that could be correlated to race, such as zip codes.