Home AI Meta’s New Business AI Offers Agentic Customer Support For Brands

Meta’s New Business AI Offers Agentic Customer Support For Brands

SHARE:

Naming a new product “Business AI” feels a bit on the nose for Meta’s current artificial intelligence efforts. But in this case, it’s certainly applicable.

Tuesday saw the early rollout of this new AI agentic tool. Business AI is part of an invite-only pilot program, which was originally tested with a select group of advertisers but is now available for any business to apply to be invited.

Business AI is intended primarily for small and midsize organizations that lack the infrastructure to offer immediate customer support or program a dedicated chatbot.

The tool sits atop a brand’s existing social media presence on Facebook, Instagram and WhatsApp, where it can provide real-time answers to customer questions over direct messages.

Head of Business AI Clara Shih told reporters at a press event in New York, “It’s really about delivering that 24-seven personalized service that every consumer, every person, wants from the businesses.”

Acclimating to AI

Business AI reportedly trains itself on existing brand-uploaded content across Meta’s platforms, as well as whatever else the business is able to provide from its own knowledge base – even uploaded pictures of handwritten material, Shih said.

Their team has also been working with new-to-Meta brands to develop a guide for how much baseline content the AI needs to operate effectively right out of the gate.

Once the AI has all the data it needs, account managers can choose what kinds of functions the agent is permitted to perform, including sharing business information, recommending products and gathering customer data. It can even handle purchases, shipping or returns directly through the existing shop ads framework.

If the AI doesn’t know the answer to a question or isn’t permitted to handle a particular task, it will redirect the customer to a human member of the brand’s support team to take over.

So far, said Shih, most businesses have been slow to open the AI up to autonomous functions as they start to acclimate themselves to the tool, mostly starting with answering easy questions about shop hours and product costs.

The concierge treatment

So far this all sounds a bit like a beefed-up version of Instagram’s existing auto reply feature, which business owners can turn on to answer up to four frequently asked questions in one go.

According to Shih, however, Business AI’s underlying LLM structure allows it to better interpret oddly worded or incomplete sentences with more natural-sounding responses, even if it’s explicitly limited in what kinds of information it can communicate to customers.

Pilot program aside, Meta already has big plans for Business AI’s future development, including a voice-activated chat function that will directly answer spoken questions and even complete purchases à la Amazon’s Alexa voice assistant. The thinking, Shih said, is that busy, multitasking users will be less likely to abandon a cart before completing their transaction if the AI is able to assist them in the moment.

Ultimately, the goal is for Business AI to develop a rapport with returning customers, approximating the kind of white-glove, concierge service associated with luxury retailers.

And then, of course, there are the potential resell and upsell opportunities that come with maintaining a conversation with customers. If someone’s already cool with ordering a necklace through an AI agent on WhatsApp, what could recommending a matching pair of earrings hurt?

Tagged in:

Must Read

Sports Publisher On3 Tries AI Recommendations To Keep Engagement In Its Home Court

Mula’s AI native content feed helps On3 keep its engagement and RPS consistent amid traffic drop-offs to publisher sites and the growing scarcity of online attention.

Comic: Race To The Bottom

Hearst Built A Unified Ad Marketplace To Simplify Omnichannel News Buys

Hearst is stitching together its far‑flung news properties into a single programmatic marketplace to simplify buying local news and shore up its business as the ad market shifts.

Northbeam Adds The Third Leg Of The Attribution Stool With Incrementality Testing

There’s MMM and MTA, but no single ad measurement works for brands with multiple points of sale. On Tuesday, Northbeam launched an incrementality tool to complete what it calls “the trifecta of digital attribution.”

Privacy! Commerce! Connected TV! Read all about it. Subscribe to AdExchanger Newsletters
Comic: The Great Online Privacy Battle

What Regulators Talk About When They Talk About Ad Tech

If you want to know what privacy regulators think about online advertising, it’s not a mystery. Just listen to what they’re saying.

Keyword Blocking Demonetized More Than Half Of Reuters’ Brand-Safe Stories

The effect wasn’t just limited to news content. The Reuters.com/lifestyle vertical also had some of its brand-suitable pages blocked.

The Agentic Marketplace Is Here. Where Does That Leave DSPs and SSPs?

Swivel and Olyzon’s new partnership brings buy-side and sell-side agents together as early examples of an agentic marketplace.