agsandrew - Fotolia

Salesforce upgrades include heavy dose of Einstein

Salesforce introduces image recognition software and upgrades its cloud products with new Einstein-related features.

Salesforce's AI product, Einstein, is now available across the company's core cloud products, as part of its spring 2017 product release.

The Salesforce upgrades include Einstein Vision, which brings image recognition software into the Salesforce suite of products. Einstein Vision is a set of APIs that allow developers to find insight from images, which could be used to detect inventory levels and product quality.

The upgrades follow Salesforce's introduction in 2016 of Einstein, a platform powered by artificial intelligence (AI) that exists across all Salesforce products. Einstein was developed through several AI acquisitions over the past couple of years.

Each Salesforce cloud received features related to Einstein, including Einstein Opportunity Insights for Sales Cloud, Einstein Supervisor for Service Cloud, Einstein Journey Insights for Marketing Cloud, Einstein Commerce Insights for Commerce Cloud, Einstein Data Discovery for Analytics Cloud, and Einstein Recommendations and Feed Insights for Community Cloud.

Second platform-wide update for Einstein

This is the second platform-wide update for Salesforce Einstein, with the coming in the wake of its release in October 2016. The Salesforce upgrades promise smarter insight and look to enable everyone to reach productivity levels directly in the applications they use. But, according to analysts, it takes some in-house cleaning and organizing to gain benefits from AI capabilities.

"What we see happening here is organizations will have AI systems, and [they] still all get together in a room and sit there and play with spreadsheets, and not bring [AI capabilities] to the decision-making process," said Ed See, principal at Deloitte Consulting. "You need to restructure the processes to integrate with those capabilities."

In addition to Opportunity Insights, Sales Cloud users will now have greater insight by combing news stories, financial results, and merger and acquisition activity with Einstein Account Insights. Scheduling automation will also be integrated into Sales Cloud with Einstein Activity Capture, connecting email and calendars to Salesforce.

Einstein Supervisor provides service supervisors with real-time, omnichannel views of the contact center and reroutes customer calls to open service agents, improving customer experiences.

Salesforce upgrades for Marketing Cloud include Einstein Journey Insights, which analyzes millions of data points to help marketers gain a complete picture of what events led to a buying decision to better optimize campaigns. These insights are better gleaned from a company's own warehouse of data, however, which can take time to incorporate into Einstein.

"When you have AI, you need to have something to train that machine algorithm, and that's going to take time," said Sheryl Kingstone, research director for customer experience and commerce at 451 Research. "Salesforce is jump-starting the algorithms, but to really improve the accuracy, you need to give it the right data."

Salesforce's newest cloud product, Commerce Cloud, is gaining intelligence with Einstein Commerce Insights. The tool is designed to analyze order history, as well as product and customer data, to better understand purchasing patterns and offer better product bundling or website optimization and campaign design.

Einstein Data Discovery, according to Salesforce, is integrated into every cloud to provide users with actionable insights and recommendations inside Salesforce. The insights can quickly be converted into presentation slides with charts and graphs, enabling communication of the discovered insights.

Community Cloud's Einstein Recommendations and Feed Insights allow for better networking of content to Community Cloud members.

"[With AI], you have to rethink how you do things today, and what you're willing to alter in the current decision-making process," See said. "Process structuring is extremely important, and people think of [AI] as a tool you can just drop in, as opposed to a tool that needs to be tightly integrated."

The addition of image recognition software

One of AI that continues to grow is image recognition software; tech giants like Google and Facebook have been working on the technology for years. The increase in uploaded photos from social media and other interactive areas has spurred interest in this technology. And with the introduction of Einstein Vision, Salesforce hopes its users can build image recognition software tailored to their own business.

The capabilities could help companies create better product databases for customers to comb through and search much easier, or help businesses better detect how consumers interact with their brand through user-generated images in social media or other web communities.

The benefits could also be useful for service agents to evaluate product issues and manage inventory.

"We created the opportunity to help bring the power of computer vision into the CRM application space," said Jim Sinai, vice president of product marketing for Salesforce Einstein. "It's about letting customers train Einstein on their customer brands and products."

Salesforce has a library of general stock imagery, but allows for companies to build out their own image recognition software to better identify images that pertain to the business. For example, a solar panel company can separate images of flat roofs and pitched roofs to create a clearer picture of the customer.

AI for everyone?

While Salesforce is touting its "democratization of AI" and how Einstein is meant for every user, despite their technological abilities, there are still questions regarding its benefits for companies on the smaller side that may not have the breadth of data available to large enterprises. That, combined with the debate between one all-encompassing AI platform versus third-party AI applications that have a distinct job, may make some users hesitant to take the AI plunge at .

"Some firms declare they have a monolithic solution, and it will work over all things," said Ian Barkin, co-founder and chief strategy officer for London-based Symphony Ventures, a professional services firm. "It makes for a good story, but when has there ever been a single tool that dominated?"

It makes for a good story, but when has there ever been a single tool that dominated?
Ian Barkinco-founder and chief strategy officer, Symphony Ventures

The data conundrum is another concern for smaller and medium-sized businesses wondering if revamping their organizational processes to include AI capabilities is worthwhile.

To help with this issue, Sinai said Salesforce can leverage data from other, similar-sized companies that can generate a best-practices AI engine for companies that don't have enough working data.

"Not only is Einstein learning across different applications, but when customers want, we can learn across customers and deliver a generalized model," Sinai said, adding that customers can choose to opt in on contributing to the generalized model.

Still, the best data to feed into AI software such as Einstein is always going to be your own.

"A lot of this depends on companies that are, at least initially, big enough to have volumes of data to supply into an AI-led system," said Joe Stanhope, vice president and principal analyst at Forrester Research. "The quality of data is a big hurdle."

Pricing for the new Salesforce upgrades varies depending on the cloud, with Community and Commerce Cloud upgrades baked into the cloud at no extra charge. Einstein products for Marketing Cloud are priced differently, depending on the edition of Marketing Cloud the customer has, while Einstein capabilities for Sales Cloud, Service Cloud and Analytics Cloud are individually priced.

Next Steps

Inside the Salesforce cloud platform

How Salesforce AI looks to change business

The souring of the Salesforce-Microsoft partnership

How AI revolutionizes customer analytics

Dig Deeper on AI for sales