Amazon is set to launch a raft of new machine-learning-powered search bar features which it expects will drive more than $8 billion in extra sales.
According to leaked internal documents seen by Business Insider, Amazon is investing heavily in becoming a “product discovery engine”.
These documents suggest that Amazon has been testing a range of new smart features for its search bar, including displaying pictures and links within its autocomplete search predictions.
This would offer customers links to individual product or brand pages, including Prime-eligible product pages, before they have even hit enter.
Amazon’s visual search technology is also set to play a significant role in this transformation, with the company predicting its StyleSnap feature will generate 10 million search queries by December.
StyleSnap, a feature Amazon launched last year which allows customers to search for products by taking photos, will also see video functionality added which it expects to drive $1.6 billion in revenues this year.
Other improvements will include greater natural language search capabilities and multi-language search functionality.
It appears to be part of a three-year strategy laid out by Amazon’s vice president of search Srikanth Thirumalai last year, aiming to “transform search from a page with search results to a shopping and discovery page”.
According to the leaked documents, these new features could net Amazon an extra $8.8 billion in sales this year alone.
It’s not surprising that Amazon is focusing so heavily on improving its search bar functionality, as according to research it is now where 60 per cent of customers start their shopping journeys.
Despite this, customers still favour other platforms focused on inspiration and discovery when they’re not sure of the exact product they want.
Last month Morgan Stanley published research showing that just over a quarter of online shoppers start with Amazon for more casual products searches, heavily favouring its key rival Google instead.