Apple Inc. reportedly acquired a startup last year called Laserlike Inc., which built software that uses machine learning to search the web for information and deliver personalized results to users.

According to a report in The Information today, Apple confirmed that the acquisition, unnoticed until now, closed late last year. Apple is likely to integrate Laserlike’s technology with its digital assistant Siri in order to improve its user-specific results.

Before the acquisition, Laserlike offered an application that was billed as an “interest search engine” that dug up news and videos related to users’ personal interests. It would also recommend new websites to to users based on their browsing habits. The company was founded by ex-Google LLC engineers Anand Shukla, Srinivasan Venkatachary and Steven Baker four years ago, The Information reported.

Apple didn’t say why it had acquired Laserlike, but the reasons are obvious enough. Apple’s progress in artificial intelligence technologies such as machine learning has left much to be desired, and it has been taking steps to rectify this.

For example, last year the iPhone maker hired John Giannandrea, the former head of search and AI at Google. Months later, it promoted him to its executive leadership team as its new senior vice president of machine learning and artificial intelligence strategy.

Giannandrea likely sees Laserlike’s technology as a way to improve Siri’s ability to fetch pertinent information from the internet. Siri has long been the butt of jokes for its sometimes nonsensical responses to requests for information and the mediocre search results it turns up.

“Apple has a lot of catching up to do as Siri’s shortcomings are now often more an amusement than helpful automation,” said Holger Mueller, principal analyst and vice president of Constellation Research Inc. “This should help Apple make Siri better at understanding the web, which she has never been particularly good at.”

Still, Apple will need to do more than just improve Siri’s underlying technology. One of its main problems is that it doesn’t have reliable compute infrastructure of its own to power its machine learning services. Instead, it relies on compute infrastructure from partners such as Google, Amazon Web Services Inc. and Microsoft Corp., Mueller said.

“It will be interesting to see if Apple will ramp up on that internally, or if it will continue to rely on its three iCloud partners for the compute needs,” Mueller said. “But first Apple needs to do some product development. Millions of Apple users will be happy if it can close the substantial gap between it and Android when it comes to speech recognition, and the ability to return more meaningful results to a user request.”

That product development is likely to be well underway, if Apple’s recent acquisitions are an indicator. In recent months its snapped up the technology and assets of several AI companies. For example it acquired a smart thermostat maker called Silk Labs Inc. last November. Also last year, it acquired several patents from the failed smart security camera company Lighthouse AI Inc.

Image: Laserlike