Skip to main content

Microsoft brings Visual Search to smartphones — and it’s a mixed bag

Watch all the Transform 2020 sessions on-demand here.


A keyboard isn’t necessarily the fastest way to find cat pictures — or furniture, for that matter. Today, Microsoft’s Bing launched a mobile version of its Visual Search tool that pulls up relevant pictures, product listings, and text snippets through photos your phone captures.

The new and improved Visual Search, which is built into the Microsoft Launcher for Android, the Bing app for iOS and Android, and Edge for iOS and Android, works more or less like Google Lens, Google’s image recognition service for smartphones. It’ll scan images (from either your camera or photo gallery) for barcodes and objects such as flowers and animals, plus notable landmarks like the Empire State Building and Statue of Liberty. And if you point your camera at a piece of clothing, furniture, or home decor, it’ll show visually similar products and their prices, along with links to stores where they’re in stock.

It’s fair to say that in this domain, Microsoft is a bit late to the object recognition game. Google Lens came to Pixel-branded devices in 2017 and, thanks to native integration with the Google Photos app, non-Pixel devices (including the iPhone) in March. Meanwhile, Bixby Vision, Samsung’s in-house visual search feature, rolled out alongside the Galaxy S8 and S8+ in summer 2017.

So better late than never? Not exactly. Annoyingly, Bing’s Visual Search doesn’t detect barcodes automatically — in the Bing app, after you tap the camera shortcut that launches Visual Search, you’re forced to select a second button that activates barcode scanning. Worse still, Bing’s algorithm misses the mark more often than not. In my brief testing, it mistook the cover of Woodward and Bernstein’s All The President’s Men as “Bob Woodward,” which obviously isn’t quite right. And it couldn’t ID a box of Starbucks Keurig cups.


June 5th: The AI Audit in NYC

Join us next week in NYC to engage with top executive leaders, delving into strategies for auditing AI models to ensure fairness, optimal performance, and ethical compliance across diverse organizations. Secure your attendance for this exclusive invite-only event.


It’s also a bit bare-bones at the moment. You won’t find a feature like Google Lens’ text selection, which lets you copy and paste words from printouts, brochures, and business cards. And unlike Bixby Vision, Visual Search can’t read wine bottles and serve up information about them.

That’s not to suggest it’s completely useless, though. Visual Search had an easier time finding lookalikes for a lamp and painting, and it picked out a Diet Pepsi can from a crowded shelf in my refrigerator without any trouble. But as with Google Lens, Bixby, and the rest of the computer vision apps before it, there’s clearly room for improvement.

Here’s hoping it’s just the start.