Emily Tan
Aug 9, 2017

Brands can now automate image search on Twitter with Einstein Vision

Einstein Vision for Social Studio is part of Salesforce Marketing Cloud's social-media solution.

Einstein Vision identifying and classifying cocktails
Einstein Vision identifying and classifying cocktails

Salesforce has launched an AI-powered image recognition service that will allow marketers to automate the discovery and identification of images on Twitter.

The product, Einstein Vision for Social Studio, is part of Salesforce Marketing Cloud's social media solution and currently only works with Twitter.

It will be available at the end of the month for Social Studio customers and according to Salesforce it can be used to for:

Gleaning consumer insights

For example, if a restaurant chain is considering adding a new menu item, it can use Einstein Vision to learn about trending options food influencers are sharing through images on Twitter. The tool is currently able to identify around 200 different foods.

Brand tracking

The tool can detect when a company logo is showing up on Twitter. So, if a brand wants to sponsor an event, for example, marketers can better quantify the brand lift associated with that sponsorship.

Social customer service

Marketers can monitor social feeds to understand where and how products are being used, and then respond. For example, if an auto manufacturer has a recall for a defective glove compartment mechanism, they can uncover cases through images on Twitter and offer customer service support.

Salesforce has also made two APIs available to developers to build their own AI-powered apps. The first, Einstein Image Classification, powers image recognition use cases including visual search, brand detection and product identification. The second, Einstein Object Detection, recognises objects within images and detects types and quantities of the objects at scale.

With around 1.8 billion photos shared daily, marketers need help, Tom Smith, product marketing manager, Marketing Cloud for Salesforce in EMEA, said.

Smith added: "By automating the discovery and identification of images marketers are able to respond appropriately, even when a post hasn’t specifically mentioned a brand, product or service by name. This level of detail is useful for developing highly personalised content, coordinated across each channel, ultimately making social deliver better results for marketers."

Campaign UK

Related Articles

Just Published

9 hours ago

Forrester CMO pushes for tighter integration of ...

Marketers need to tighten up oversight over all the touchpoints in the buying journey, which has increased from 17 to 27 on average during the pandemic, according to Shirley Macbeth.

9 hours ago

Baby Shark doo doo doo doo doo. LG Styler doo doo ...

Smooth out doo doo doo doo doo. Bad smell doo doo doo doo doo. Shake it out doo doo doo doo doo. LG Styler! [Something is very wrong with Ad Nut.]

11 hours ago

IAS explains ad verification in manga style

A young automobile-company exec learns how to mitigate brand risk and ad fraud in a free download from the ad-verification company.

12 hours ago

L'Oreal opens omnichannel concept store in Shanghai

Designed with AKQA, the store has a livestreaming set, face-scanning technology and personalised shopping via WeChat integration, as well as a chance to ride a bike through the streets of Paris.