Meta Just Put Its AI to Work for America's Small Businesses — and the Software Giants Are Already Bleeding

Meta CEO Mark Zuckerberg speaking at Facebook's F8 developer conference in April 2019. This week he declared enterprise AI the company's "next major pillar." Photo: Anthony Quintano / Wikimedia Commons (CC BY 2.0).

Meta spent the last decade selling ads to small businesses. Now it wants to run them.

On Tuesday, the company expanded its Muse AI agent into a full small-business operations tool, wiring it directly into the software that millions of American owners already use to keep the lights on: Shopify, QuickBooks, Stripe, and Canva. With the move, Meta is no longer just helping businesses find customers — it is offering to do their bookkeeping, analyze their ad campaigns, draft their marketing, and flag the emails that need a reply.

The announcement, which broke Tuesday morning via Reuters, is the second half of a one-two punch Meta threw this week. On Monday it launched the Meta Enterprise Platform, a dedicated business unit for selling its AI tools to large corporations, and Meta's surprise hire of MongoDB CEO CJ Desai to run the new unit as chief enterprise platform officer. Wall Street's reaction was brutal: Meta fell 4% on Monday, MongoDB cratered by roughly 18% in morning trading, and the sell-off spread to Salesforce, Microsoft, and SAP. Investors are pricing in a land war for the enterprise — and Meta just fired the opening shots.

What Meta announced Tuesday: an agent that reads your books

Muse for Small Business is not a new app. It is an expansion of Muse, the personal AI agent Meta launched earlier this month across iOS, Android, the web, and WhatsApp in the United States and Canada — an app that has already topped the free-app charts, with market intelligence firm Sensor Tower estimating about 2.8 million downloads in its first two weeks, according to Reuters.

The small-business version plugs Muse into the operational stack of a typical owner. Beyond the headline integrations with Shopify, QuickBooks, Stripe, and Canva, Meta says the agent connects with Instagram professional account analytics, Facebook Pages, and Meta ad accounts, plus a long tail of third-party tools: Asana, Box, Dropbox, Figma, Granola, HighLevel, Klaviyo, Lovable, Notion, Slack, and Zoom. Owners can also build custom connectors for services not yet supported.

The pitch is that Muse stops being a chatbot and starts being a chief of staff. Meta says the agent can use a company's business context — its brand, storefront, books, and customer records — to analyze sales and campaigns, review expenses, draft ads and customer messages, and help manage work across connected services. Ask it which ads are working, whether an expense looks unusual, or which emails need a reply, and it answers from your actual data rather than from general knowledge.

Meta is also drawing a hard line on autonomy that its rivals will be watching. The company says Muse will not publish content, send messages, or make purchases without a user's approval — a guardrail that matters enormously when the agent holds the keys to your storefront and your bank feed. And it is free for most uses, with subscription plans reserved for users seeking additional features, per Reuters.

The bigger bet: enterprise AI as Meta's "next major pillar"

Tuesday's small-business push only makes sense inside Monday's bigger announcement. Meta Enterprise Platform is a new, dedicated business line packaging the company's AI models, agents, developer APIs, and data-center compute for sale to outside businesses. CEO Mark Zuckerberg has called it the "next major pillar" of Meta's business — language that puts it alongside advertising and consumer products in the company's long-term structure.

The initial product stack, per TechRepublic's reporting on the launch, includes Muse itself, Meta Business Agent, the Muse API, and Muse Code. Meta Business Agent is the corporate-facing sibling: Meta says more than one million businesses already use a Business Agent on WhatsApp and Messenger to respond to customers, giving the new enterprise unit an installed base most AI startups would kill for.

The hiring tells you how serious this is. Chirantan "CJ" Desai, the sitting CEO of MongoDB, resigned effective immediately to become Meta's chief enterprise platform officer, reporting directly to Zuckerberg. MongoDB confirmed former CEO Dev Ittycheria would step in as interim chief executive. Desai's background is pure enterprise software: before his brief MongoDB tenure, which began in November 2025, he served as president and chief operating officer at ServiceNow and led product and engineering at Cloudflare. Meta is not staffing this like an experiment.

This is also a story about monetizing infrastructure Meta already paid for. The company has spent years building AI capacity to serve its consumer platforms — data centers, models, developer tooling — and Enterprise Platform is the vehicle for selling that investment beyond advertising. Meta's guidance calls for up to $145 billion in capital expenditure this year, much of it on data-center infrastructure and advanced models. Turning that sunk cost into someone else's subscription is the entire logic of the new unit.

Wall Street's verdict: sell first, ask questions later

Markets did not greet the enterprise offensive with applause. Meta shares fell 4% on Monday, reversing a sharp rally that had been driven by Muse's consumer success. MongoDB took the worst of it: shares plunged roughly 18% in morning trading on the sudden leadership change, with some intraday reports showing losses as deep as 27%.

The sell-off then spread across the software sector, per Dow Jones reporting: Salesforce fell 4.7% at the open, Microsoft dropped 2.3%, and SAP, Germany's most valuable company, lost 3.4%. Investors read the Desai hire the way Meta intended it to be read — as a declaration that Meta is coming for enterprise software budgets, starting with the vendors who sell AI tooling to businesses.

The skepticism has a number. Meta's free cash flow fell 91% year over year to $784 million in the second quarter of 2026, from $8.55 billion a year earlier, even as revenue rose 28% to $60.8 billion, per market coverage of the results. The AI spending that Enterprise Platform is meant to monetize is currently consuming nearly all of the cash Meta generates. Investors have punished this before: in October 2025, Meta shares fell more than 11% in after-hours trading after the company warned that 2026 expenses would rise sharply on infrastructure investment and pay for new AI hires. The enterprise unit will need to show revenue traction relatively quickly to justify the pace.

Why small businesses are the real prize

Strip away the stock-market drama and Tuesday's announcement is arguably the more consequential of the two. Big enterprises buy AI through long procurement cycles and pilot programs. Small businesses — tens of millions of them in the US — adopt whatever shows up inside the tools they already open every morning. Meta already owns several of those tools: Facebook Pages, Instagram professional accounts, WhatsApp. By embedding Muse in that daily surface and connecting it to Shopify and QuickBooks, Meta is distributing an AI agent through channels its enterprise rivals cannot easily replicate.

The competitive framing is explicit. This puts Meta more directly into an enterprise AI market already crowded with Microsoft, Salesforce, Google, and ServiceNow — all racing to make AI agents part of everyday business operations. But Meta's wedge is different: rather than selling to the CIO, it is selling to the owner who does the books at midnight. Muse for Small Business is an expansion of the existing consumer agent, not a separate app built just for business, which means every consumer download is a potential business conversion.

There is also a defensive logic. Shopify — one of Tuesday's headline integrations — has been aggressively building its own agentic commerce stack, including WebMCP checkout support that lets AI agents complete purchases on merchant sites. By making Muse the agent that operates inside Shopify stores, Meta keeps its own assistant at the center of the transaction rather than ceding the relationship to someone else's agent.

What to watch

Three questions will decide whether this week is remembered as a turning point or a spending spree. First, monetization: the announcement included no pricing details for the enterprise products and no customer commitments. "Free for most uses" is a distribution strategy, not a revenue model, and investors will want to see the subscription tiers and enterprise contracts that justify up to $145 billion in capex.

Second, trust. Asking small businesses to connect their books, their storefronts, and their customer records to one company's agent is an enormous ask — especially from a company whose core business is advertising. Meta's approval-before-action guardrails are a start, but the data-access questions from privacy advocates and regulators will be relentless.

Third, the competitive response. Microsoft has Copilot embedded in the office suite every small business runs. Salesforce owns the CRM. Google owns the inbox. None of them will watch Meta walk into their accounts unopposed — and Monday's stock moves suggest the market already expects a fight. Zuckerberg is expected at Tuesday's White House AI lunch with tech CEOs, where the balance between AI innovation and oversight is the agenda; Meta's enterprise ambitions will be part of the subtext whether or not they are on the menu.

For now, the message from Menlo Park is unmistakable: the ad business built this company, but the next decade gets built on selling AI to the businesses that buy the ads. Tuesday was the day Meta started collecting.

Sources: Reuters (Sept. 29, 2026); RuntimeWire; TechRepublic; Wealth Professional via Dow Jones; StockTi; AutoNAI News. Market figures as reported by the cited outlets; Meta product claims attributed to the company.