Google is injecting its search engine with new technology to better interpret the billions of Web queries it handles every day, a change top executives described as one of the most significant in the company’s history.

This shift, announced on Friday, moves the world’s biggest search engine from spitting out results based on keywords to “something closer to language”, said Ben Gomes, the search chief for Google. “We’re very far from solving the problem fully, but this is a huge step forward,” he said at a press briefing.

Google has no peers in Web search. But improvements to its core search technology are critical for keeping an edge in adjacent areas, primarily voice computing, where Google competes with Amazon.com. The new system relies on a Google artificial intelligence tool designed to parse long, complicated sentences, rather than just strings of words. In tests, Google executives said it produced far more precise results.

Any tweaks to Google search have ripple effects across industries that depend on it for Web traffic. Over the years, Google has moved further away from its old 10 blue links results page. It now shows more results from its own services for some queries, such as flight information, or pulls out blocks of texts from websites into what it calls “featured snippets”. Rivals have complained furiously to regulators that Google’s actions are anticompetitive.

The Google executives said the new system had produced more featured snippets in results outside the US. They insist that Google isn’t pulling eyeballs away from others, but that improvements to search results drive more searches and overall Web traffic.

‘More queries’

“If people are able to get more queries answered, they ask more queries,” Gomes said. “And that results in more traffic to the overall ecosystem.” He didn’t share data on those trends.