NEW YORK (AP) — Major retail chains and tech companies are offering new or updated artificial-intelligence tools in time for the holiday shopping season, hoping to give consumers an easier gift-buying experience and themselves an augmented share of online spending.

Although AI-powered purchases are in early stages, the shopping assistants and agents rolled out by the likes of Walmart, Amazon and Google can do more than the chatbots of holidays past.

The latest versions were designed to provide personalized product recommendations, track prices and to place some orders through unscripted “conversations” with customers.

Those features are on top of shopping updates from AI platforms such as OpenAI’s ChatGPT and Google Gemini.

In one of the season’s most talked-about launches, Google last month introduced an AI agent that can be instructed to call local stores to ask if a desired product is in stock.

San Francisco software company Salesforce estimated that AI would influence $73 billion, or 22 percent, of all global sales in one way or another from the Tuesday before Thanksgiving through Monday after the holiday, according to Caila Schwartz, Salesforce’s director of consumer insights.

The figure, which stood at $60 billion a year ago, encompasses everything from a ChatGPT query to AI-supplied gift suggestions on a retailer’s website, Schwartz said.

Despite the advancements, AI’s impact on holiday shopping will be “relatively limited” this year since not every shopping site has useful tools and not every shopper is willing to try them, said Brad Jashinsky, a senior retail industry analyst at information technology research and consulting firm Gartner.

“The more retailers that launch these tools, the better they get, and the more that consumers get comfortable and start to seek them out,” Jashinsky said. “But customer behavior takes a long time to change.”

The use of AI by shoppers means bypassing the search bar. AI’s potential to simplify the search for the perfect present is most apparent so far in tools that promise to give shoppers faster and more detailed results than a web browser with fewer clicks.

OpenAI upgraded ChatGPT with a shopping research feature that provides personalized buyers’ guides. The information comes from product pages, reviews, prices and a user’s previous interactions with the chatbot. The tool works best for complicated products like electronics and appliances, or for “detail-heavy” items such as beauty or sporting goods, OpenAI said.

Then there’s Rufus, the shopping assistant that Amazon rolled out last year. It now remembers information customers previously fed it, such as having four children who all like board games, for example. A user’s browsing and purchase history and reviews are used to personalize recommendations.

Google upgraded its AI Mode search tool to provide answers to detailed questions composed in natural language. For example, users can tell the agent they want to buy a casual sweater to wear with skirt or jeans in New York in January that goes with a skirt or jeans,

Responses are pulled from Google’s 50 billion product listings. The tool can also produce charts with side-by-side comparisons of prices, features, reviews and other factors. Previously, shoppers had to use keywords, filters and product links to find the information they needed.

There are also new pricing tools and alerts.

Tools for tracking online prices have been around for years, including CamelCamelCamel, a third-party service for Amazon prices, as well as Paypal’s Honey browser extension for monitoring thousands of online shops.