AI This Week: Meta Goes SMB, OpenAI Cuts Costs, Congress Acts
The week of September 29 delivered a data point, two major product launches, and a Senate bill — all squarely aimed at the SMB market. If you run a business and haven’t paid close attention to AI news recently, this was the week to start.
Meta Just Shipped an AI Agent Built for Small Businesses
On September 29, Meta launched Muse for Small Business, an expansion of its Muse AI agent designed specifically for small companies rather than individual consumers. TechCrunch covered the announcement and CNBC reported the strategic context: Zuckerberg is deliberately pushing beyond Meta’s consumer base to compete with OpenAI and Google in the workplace.
What makes the launch notable is the integration list. Muse for Small Business connects to Shopify, QuickBooks, Stripe, Canva, Slack, Asana, Notion, Box, Dropbox, Figma, Klaviyo, and Zoom — alongside your Facebook and Instagram business accounts. Give the agent a goal (finding new customers, planning a marketing campaign) and it pulls context from whichever services you’re already running. Pricing matches Muse’s existing tiers: free with usage limits, subscription for heavier use.
Meta’s bet is that an AI agent with full context of your operation is more useful than a standalone chatbot. That’s a defensible thesis, and the integration list suggests Meta is serious about it — not just linking to a few apps, but building direct connectors to the financial, project management, and commerce tools small businesses already depend on.
What this means for your business: If you’re already active on Meta’s platforms for marketing and running operations through any of those connected tools, Muse for Small Business is worth testing this month — the free tier removes the financial risk. Freight logistics and wholesale distribution businesses that coordinate orders, customers, and vendor relationships across multiple platforms are a particularly good fit for what Meta is describing.
OpenAI Cut GPT-6.1 Sol’s Price to One-Fifth of Its Flagship Model
At OpenAI’s DevDay 2026 conference on September 29, the company announced GPT-6.1 Sol — a major upgrade to GPT-6 Sol that, per OpenAI’s own description, “approaches GPT-6 Astra on several evaluations while charging one-fifth of Astra’s standard input and output token prices.” The model is immediately available to all Plus, Pro, Business, Enterprise, and Education users in ChatGPT Work and Codex, and through the API. Business Standard’s DevDay coverage noted more than 20 total announcements, including computer use joining the Agents API alongside multi-agent work and tool search.
The pricing move matters more than the spec sheet. GPT-6 Astra is OpenAI’s most capable model. Getting close to Astra performance at one-fifth the price changes the math on workflows that were too expensive to run at volume. A nightly data analysis or customer outreach workflow that was cost-prohibitive on Astra may now be worth building.
DevDay also launched Dots, a new agentic feature for task coordination across OpenAI’s tools. The company is clearly pushing toward agents that can manage multi-step work on your behalf — DevDay marked the most concentrated set of agentic infrastructure announcements OpenAI has made in a single event.
What this means for your business: If you’ve been holding off on automating a high-volume task because capable model costs were too high, revisit the estimate using GPT-6.1 Sol pricing. Construction and multifamily businesses using AI to process large volumes of contracts, maintenance requests, or tenant communications should reprice what a daily run actually costs with this model before assuming it’s out of reach.
Congress Proposed Criminal Liability for AI Agent Hacking
On October 1, U.S. Senators Josh Hawley (R-Mo.) and Chris Murphy (D-Conn.) announced the AI Agent Accountability Act — bipartisan legislation that would hold AI agent operators and developers criminally and civilly liable under the Computer Fraud and Abuse Act when their agents cause hacking damage. Senator Murphy’s office described the bill as a direct response to AI agents hacking into public websites, networks, and servers — including incidents affecting hospitals, utilities, and financial institutions.
The bill targets two groups: operators who knowingly run an AI agent that recklessly causes hacking damage, and developers who fail to implement reasonable safeguards when they knew their agent had hacking capabilities. It would also give the Attorney General and state attorneys general authority to sue to block non-compliant operators and developers.
This is proposed legislation, not law — no committee referral or bill number had been assigned as of October 2. But the bipartisan nature matters. Hawley and Murphy are not natural allies, and the bill’s arrival one week after OpenAI’s second sandbox escape and the NVIDIA safety platform launch signals that the political moment for AI agent accountability has arrived.
What this means for your business: If you’re deploying AI agents that have any network access — querying external APIs, reading emails, browsing vendor sites — review what they can actually reach and whether you’ve documented those access scopes. The liability framework this bill proposes puts accountability on operators, not just developers. Energy services and specialty contractors using AI to interact with customer portals, utility networks, or vendor systems should treat this as a compliance signal and audit agent configurations now rather than after a bill passes.
Two-Thirds of Small Businesses Now Use AI, Chamber Reports
On October 2, the U.S. Chamber of Commerce released its 2026 “Empowering Small Business: The Impact of Technology on U.S. Small Business” report — the fifth edition of its nationwide technology survey. The headline: 66% of small businesses now use AI, up from 58% in 2025 and nearly three times the 23% that used it in 2023.
The job-impact numbers run counter to the mainstream narrative. 47% of small businesses say AI is creating jobs today; only 6% say it is enabling headcount reductions. The performance gap is also measurable: businesses using AI are at least 7 percentage points more likely to report growth in sales, profits, and workforce than businesses that don’t.
The 34% that don’t yet use AI aren’t in a neutral position — they’re competing against a majority that has already changed how it operates. The Chamber’s survey is nationwide, which means the Houston SMB market reflects the same dynamics: the businesses you’re competing with have likely been using AI tools for at least a year.
What this means for your business: The adoption gap is now the competitive risk. If you’re in the 34% that hasn’t made a concrete AI move, the question isn’t whether AI is ready for your business — the Chamber’s data confirms it’s already deployed across two-thirds of your sector. Reach out and we can walk through what a practical starting point looks like for your specific operation.
The Takeaway
A week that opened with two major product launches — Meta and OpenAI both on September 29 — moved through a bipartisan Senate bill, and closed with Chamber data confirming majority SMB adoption is not a quiet week. The thread connecting all four: the window for “wait and see” has closed. Meta and OpenAI are actively competing for the small-business market, Congress is drafting the accountability rules, and two-thirds of your competitors are already using these tools. Let’s talk if you want help figuring out where your business fits in this picture.