- Meta and Sierra proposed an open standard for how personal AI agents talk to businesses, with Walmart, Shopify and Stripe on board.
- B2B marketing tools now track how AI assistants describe and recommend a brand.
- Agent builders and training programmes aim to let non-developers build agents.
- No standard is final yet. The basics (clear information, an easy way to book or buy, clear rules) are work you can do for clients now.
Meta and Sierra propose a standard for agents that deal with businesses
On 6 October 2026, Meta and Sierra announced the Personal Agent Protocol, a proposed open standard for how a person's AI agent connects to a business. It builds on OAuth, the same sign-in technology behind "Log in with Google". An agent starts with guest access and can move to signed-in read-only or write access when the customer allows it. The coverage describes three ways to connect: navigating the website, using the company's APIs (such as MCP or OpenAPI), or talking to the company's own agent.
The founding partners named at launch are Walmart, Shopify, Stripe, Rocket, Genesys and Instinct. Amazon, OpenAI and Anthropic were not on that list. A first v0.1 specification was expected later in October, and payments are described as a future extension. Sierra's Bret Taylor called the current situation "kind of chaos until such a standard exists".
Your angleMost small businesses have never asked "what happens when an agent visits us?". That question alone opens a conversation. Start with the website path, because it works for every client today.
Sources: Channel Insider, Constellation Research
Brands start tracking what AI assistants say about them
At its GO conference in New York on 20 October 2026, Demandbase launched three AI agents for B2B marketing. The AEO Agent (answer engine optimisation) tracks how AI assistants describe and recommend a brand, flags outdated information and helps identify anonymous traffic that comes from AI tools. The Campaign Agent and Audience Agent work on targeting and campaigns. In its announcement, Demandbase cites a Gartner prediction that by 2028 90% of B2B buying will be intermediated by AI agents. That is a forecast, not a measured result.
Your angleBig companies pay for AI answer tracking. A local business can get a simple version from you: ask the main assistants about the business every month and report what is wrong or missing.
Agent builders now start from plain English
Flozic launched an AI Agent Builder on 20 October 2026 that creates an agent from a plain-English description. According to the company's press release it comes with more than 80 prebuilt templates, voice agents, computer-use agents and execution logs, and it integrates with tools such as Salesforce, HubSpot, Gmail, Slack, Shopify and Stripe. It lets you pick models from OpenAI, Anthropic, Google, Meta and Mistral. These are the company's own claims, so test them before you build a client offer on top.
Your angleWhen building gets this easy, the value moves to choosing the right task and checking the logs. Sell the setup and the weekly review, not the button click.
Source: PR Newswire
Consultancies train business teams to build their own agents
Consultancy Sia launched AgentLaunch on 22 October 2026, a 10-week part-time programme that teaches business employees (not developers) to build AI agents. Participants work on their own business cases and present them at a Demo Day. Sia says it has developed more than 1,000 agents itself. The signal: large companies now treat agent building as a business skill, not only an IT project.
Your angleSmall businesses will not run a 10-week programme. They need someone who does the same thing for them in a few sessions. That is a service, and it can be packaged.
Source: Consultancy-me.com
Your client's next visitor might not be a person. It might be that person's agent.
The B2A checklist: 5 things to set up for a client
None of the new standards are final. These five steps work today, whatever standard wins:
- Ask the assistants. Ask ChatGPT, Gemini and Claude what they know about the business, its prices and its opening hours. Write down what is wrong or missing.
- Make the basics readable. Services, prices, hours, location and contact details as plain text on the site, not only inside images or PDFs. Add structured data where it fits.
- Give agents a clear route. One booking link, one contact form or one product page per offer, so an agent can finish the task without guessing.
- Decide the rules. What may an agent only read, and what may it do (book, order, cancel)? The Personal Agent Protocol uses the same split: read-only first, write access later.
- Check every month. Repeat step one and keep a short log. That log is the report your client sees, and the reason they keep paying for the service.
v0.1 is a first draft. Clear information and a clear booking route help with any agent, today.
Read versus write access is the core of the new protocol. Clients need someone to decide this for them.
Forecasts like Gartner's are forecasts. Sell what you can check and show, month after month.
Get the next briefing in your inbox
One short email when there is AI agent news that matters for people who sell to businesses. Sources included, no hype.
FAQ
What does B2A (business-to-agent) mean?
A customer sends an AI agent to search, compare, book or buy on their behalf. The business then has to serve the agent as well as the human: readable information, a clear way to connect and rules for what the agent may do.
What is the Personal Agent Protocol?
A proposed open standard that Meta and Sierra announced on 6 October 2026 for how personal AI agents connect to businesses. It builds on OAuth, moves from guest access to signed-in read and write access, and launched with Walmart, Shopify, Stripe, Rocket, Genesys and Instinct as founding partners. A first v0.1 specification was expected later in October 2026.
What can a small business do about AI agents today?
Check what AI assistants say about the business, make prices, hours and services easy to read on the website, offer a simple booking or contact route an agent can use, and decide what an agent is allowed to do. None of the new standards are final yet, so start with the basics.
Recorded training · real agent screens
Watch a team of AI agents do real client work
In the free training you will:
- See agents handle leads, follow-up, content research and spreadsheets, step by step.
- Learn how to use it for your own business or whitelabel it for clients.
- Leave with a next-step plan you can start the same week. No coding required.
Free training. No income guarantee: results depend on your own work and clients.
What you learn to deliverReady to start this week?
Skip the wait: get the full course
€497 one-time · instant login after paying
- 6 modules: build agents for leads, follow-up, research and reporting
- Price calculator + proposal and outreach templates
- Progress tracking, lifetime access, no coding needed
Not sure yet? Watch the short video first · compare all 3 options
Secure checkout via Stripe (+2% service fee). No income guarantee: results depend on your own work and clients.
All sources
- Channel Insider: Meta, Walmart, Shopify and Stripe back an AI agent standard
- Constellation Research: Meta and Sierra pitch the Personal Agent Protocol
- Demandbase launches AI agents (press release, 20 Oct 2026)
- PR Newswire: Flozic launches AI Agent Builder (20 Oct 2026)
- Consultancy-me.com: Sia launches AgentLaunch (22 Oct 2026)