Salesforce on Wednesday announced Einstein Search, a new search feature that aims to make sales and service teams more productive. The new, intelligent search bar is currently being piloted by some customers, and early results show it can reduce the number of clicks it takes to accomplish a task by 50 percent to 80 percent, Salesforce says.
At scale, Salesforce says, this will result in significantly more productivity, given that search is Salesforce's most-used feature. There are more than a billion searches a month.
Einstein Search makes searching within Salesforce's CRM platform more efficient in a few different ways, starting by returning personalized results. The new search tool also responds to natural language queries, with an understanding of natural language specifically as it applies to Salesforce. For instance, a salesperson could search for "my open opportunities in New York."
Einstein Search also improves efficiency with an actionable search bar. For instance, a user could edit their sales records, or connect a contact to an opportunity, directly from their search results.
In a blog post, Salesforce VP of Product Management Will Breetz explained why creating an efficient search tool for the enterprise is more challenging than it sounds. First, different users within an enterprise can have completely different goals -- such as trying to close a sales deal, responding to a customer service case or creating a new marketing campaign -- which makes it harder for the search engine to understand user intent.
Additionally, different Salesforce customers customize their CRM environments in different ways, organizing data to fit their needs. This can make it difficult to build one, single search model that's optimal for all customers.
On top of that, Breetz noted, Salesforce does not look at customer data. Maintaining its privacy standards means Salesforce can't rely on traditional machine learning techniques to improve its search tool.
Salesforce says it uses innovative data mining and machine learning techniques to deliver personalized results while keeping specific user information anonymized.
Einstein Search will be generally available next year, while customers can get early access to the beta release this winter.