Facebook Inc. is taking aim at Google’s key advertising asset: the search engine’s knowledge about what a consumer is looking for.

To do so, Facebook is expanding the scope of dynamic ads, the ad format that  lets a retailer upload its product catalog to the social network, enabling Facebook to serve shoppers remarketing ads based on the products they looked at on the merchant’s site or app.

The social network today announced that it is rolling out a dynamic ads update that uses on-Facebook and off-Facebook signals (gathered from retailers that advertise with Facebook and have the Facebook Pixel and/or App Events, both of which enable the social network to gather consumer information on retailers’ websites and apps, respectively). Those signals are aggregated from multiple websites and apps to enable a merchant to target ads at potential customers whom Facebook’s model believes are likely to respond to specific products—even if those signals occurred on a competitor’s website. Among those signals are the content a consumer looks at and  his Facebook activity, such as his likes and ad clicks. That means that dynamic ads, which had largely been a remarketing tool, are now positioned to be a customer-acquisition tool.

For instance, if a consumer is browsing for furniture across multiple e-commerce sites and apps, Facebook may infer that he’s looking to buy a couch. The social network’s model may then determine that he’s a good candidate to receive a dynamic ad for furniture from Wayfair, if the merchant has set up a campaign to target consumers that the model determines are good a fit for products.

“We’re looking to automate the process of showing the most relevant products to the most relevant people,” says Maz Sharafi, Facebook’s director of product marketing. “Today, if a retailer has 1,000 products, it would have to create 1,000 different ads to reach prospective customers. This automates that process.”

In some ways, the model that determines the products a consumer is likely to be interested in is similar to Facebook’s Lookalike Audiences tool. That tool uses a model that draws on data gathered about customers included in a merchant’s Custom Audiences data set—a customer list a retailer shares with Facebook based on off-Facebook information, such as its email database—to find similarly behaving Facebook users that a retailer can then target with ads. While retailers using Lookalike Audiences still need to create an ad aimed at a Lookalike Audience and establish parameters such as the interests and demographics that might be interested in a particular product, the dynamic ad creation process is automatic.

“Rather than require retailers to take part in the ad-targeting guessing game, we’re automating that rather tedious process,” says Gabriel Francis, Facebook’s direct response product marketing lead. “We’re taking the burden off of them to save them time and deliver better results.”

Wayfair Inc., No. 24 in the Internet Retailer 2016 Top 500 Guide, is one of a “limited set” set of North American retail and e-commerce companies  using the new tool, which is available via Facebook’s Ads Management API (or application program interface, which enables a retailer to create ad campaigns, provide bids for ad auctions and manage ad creative) and its Power Editor tool that enables large advertisers to create, edit and publish multiple ads at once and  control their campaigns. The new dynamic ads functionality is not yet available via Facebook’s Ads Manager tool, which is the tool typically used by smaller merchants. When a retailer sets up a campaign, it uploads its product catalog and then can select specific parameters, such as “only present ads to women” or “consumers in Texas”.

By attempting to provide marketers with ads based on consumer intent, Facebook is taking aim at Google at a time when the search giant also is facing a challenge from consumers increasingly starting their shopping-oriented searches on Amazon.com Inc.’s sites and apps. 55% of consumers in a recent survey by internet marketing firm BloomReach Inc. said they go to Amazon first when searching for products online, an increase from 44% a year earlier. The survey showed search engines, such as Google and Yahoo, and retailers losing ground to Amazon (No. 1 in the Top 500). Search engines were the starting point for 28% of those surveyed, declining from 34% a year earlier. Specific retailers were the starting point for 16%, down from 21%.

Facebook, despite its massive growth, faces its own challenges because some of that growth has been the result of the social network increasing its ad load, which it can no longer sustain, David Wehner, Facebook’s chief financial officer, said during its third quarter earnings call.

“We continue to expect that ad load will play a less significant factor driving revenue growth after mid-2017,” he said. “Over the past few years, we have averaged about 50% revenue growth in advertising. Ad load has been one of the three primary factors fueling that growth. With a much smaller contribution from this important factor going forward, we expect to see ad revenue growth rates come down meaningfully.”

As its ad load growth slows, Facebook has to figure new ways to drive growth—particularly given that social networks continue to represent a small share of e-commerce sales, at least when measured by last-click attribution. For instance, data gathered by Adobe Digital Insights shows just 0.9% of holiday sales from Nov. 1-Dec. 5 stemmed from shoppers who clicked from a social network.

Francis suggests that while studies that rely on last-click attribution often undersell social’s influence on sales, he acknowledges that retailers are still figuring out how to understand what drives someone to make a purchase.

“It’s a challenge for many marketers,” he says. It’s also a challenge for Facebook, which is why the social network is betting on this dynamic ads expansion to attract new ad dollars.