The biggest shift this week: AI agents no longer wait for you to type. Anthropic, OpenAI and Meta now all have agents that watch your inbox and apps in the background, prepare the work, and ask for approval before they act.

Here's what happened, and what it means if you run a service business.

1. Anthropic launches Claude Duet, an agent that works while you don't

On October 21, Anthropic introduced Claude Duet, a research preview of a proactive agent. It runs in the background on its own virtual machine, sends a morning brief, and shows exactly what it plans to do. It waits for your approval before it sends a message, makes a payment or deletes anything. It's available by waitlist to Claude Pro and Max subscribers in the US. Coverage: MacRumors, iPhone in Canada.

Duet joins OpenAI's Dots and Meta's Muse agents. Anthropic also says Claude's "auto mode" becomes the default for Pro, Max and Team plans starting the week of October 26, with automated checks that block risky actions.

What it means for you: the "approve before it acts" design is the right model for any business using AI. It's the same rule we use for AI receptionists: the AI answers, books and drafts, and anything with money, a contract or a sensitive client decision goes to a person.

2. Banks and Microsoft warn: AI agents are a new security risk

Axios reports that banks including Morgan Stanley and Bank of America are working to contain risks from third-party AI agents that can access customer accounts. The hard part is telling whether an action came from the customer or from their agent (Axios).

Microsoft introduced Guardian Agents, a system that watches other AI agents in real time and limits what they can do, for example keeping read access while blocking write access, or requiring human approval before a high-impact step (TechRadar).

What it means for you: before you connect any AI agent to email, payments or client files, set three rules:

  • It gets the least access it needs: read-only wherever possible.
  • Anything involving money, contracts or private client data needs a human click.
  • It never asks for passwords, card numbers or ID numbers in a chat or call.

3. New AI video and image tools for marketing

  • HeyGen Video 1 ranked first for text-to-video in a blind test by Artificial Analysis (announced October 23).
  • Canva launched its first image-generation model, available to all users (TechCrunch).
  • Ideogram launched a model that creates editable vector (SVG) graphics and converts images to vectors.

What it means for you: the tools keep getting better and cheaper, so the advantage is no longer access to the tool. It's direction: a clear idea, a real story about your business, and an ad that matches your brand. That's what we focus on in LeverageFirst Studio.

4. AI keeps getting cheaper

One AI price index fell 54% from its May peak to under $1 per million tokens, according to Silicon Data figures reported this week. For a small business, that means AI receptionists, follow-up and content tools cost less to run every month, which is why automation now makes sense for a 2-person office, not just large companies.

The takeaway

AI is moving from "a chat you open" to "a helper that works in the background." The businesses that benefit set the rules first: what the AI may do alone, and what always needs a person. If you want that set up for your calls, texts and follow-up, book a strategy call.

New to this? Start with missed-call text-back vs. an answering service or how an AI receptionist works for accounting firms.

Frequently asked questions

What is Claude Duet?

Claude Duet is a proactive AI agent from Anthropic, announced October 21, 2026 as a research preview. It works in the background, sends a morning brief and asks for approval before it sends messages, makes payments or deletes anything. It is available by waitlist to Claude Pro and Max subscribers in the US.

Are background AI agents safe for a small business?

They can be, if you limit their access, require a human approval for anything involving money, contracts or private client data, and never let them collect passwords or card numbers in a chat or call.

What are Microsoft Guardian Agents?

Guardian Agents are a Microsoft security system that monitors other AI agents in real time and can restrict them, for example by blocking write access or requiring human approval before a high-impact action.