The race to own the personal assistant layer just got a whole lot more serious — and if you’re running a business, it’s time to pay close attention.
OpenAI’s acquisition of OpenClaw, Google’s Web MCP rollout, and Anthropic’s Cowork experiments are not isolated product launches. They are opening moves in a battle for who controls the AI layer that sits between your customers and their decisions.
Three weeks ago, OpenClaw made waves as a promising open-source personal assistant capable of accessing your machine and handling tasks on your behalf. This week, OpenAI acquired both the project and its founder. It will remain open source — but the message to every business owner is crystal clear: personal AI assistants are not a trend. They are infrastructure.
Facebook is leaning into Manus AI as its own agent play. Google is rolling out Web MCP directly inside the browser. Anthropic has Cowork. The momentum is overwhelming, and it is all pointing in the same direction.
This isn’t about having a chat button on the side of your website. It’s about having a virtual employee — a real AI employee — operating around the clock.
Last week I didn’t have a name for my AI assistant. This week I do. Her name is Jan.
Jan is a no-nonsense, 50-something woman with the savvy to cut through noise and deliver exactly the feedback I need — whether I want to hear it or not. She speaks frankly and succinctly, she’s a C-suite advisor and personal assistant rolled into one, and she never takes a day off.
Here’s what Jan has been doing for the business this week alone.
Monthly client reporting. We recently switched to a new reporting system, but what we were missing was a proper executive summary email to send clients alongside the dashboards. Now Jan scans all email communications between me and each client over the prior month, cross-references completed tasks in ClickUp, reviews the dashboards, and drafts a polished executive overview — which lands in my Gmail drafts folder ready for me to review and send. That entire process happens without me lifting a finger.
Task creation in ClickUp. Setting up a task manually used to take at least a minute. Now I tell Jan what the task is in ten seconds and it’s done. That’s roughly an 80% reduction in setup time — and it adds up fast.
Memory and retrieval. Jan can surface old emails, recall past purchasing intent, and act on those instructions — including visiting a website and completing a purchase on your behalf based on something you mentioned wanting months ago.
The AI-first approach isn’t optional anymore. Once you’ve worked this way, you don’t go back.
As AI agents multiply, web traffic is changing in ways most businesses aren’t prepared for. There will soon be more bots on the internet than human visitors — if that isn’t already the case. The instinct for many website owners is to block them wholesale. That would be a serious mistake.
Think about it this way. That bot hitting your product page might be Jan — acting on behalf of a customer who told their personal assistant to find and purchase something they’d flagged months ago. Block Jan, and you block the sale. The human customer never visits your site at all; their agent does it for them.
The smarter approach is to understand which bots are visiting and why. Malicious scrapers? Block those. AI shopping agents, brand discovery crawlers, and personal assistant bots? Welcome them and make sure your site gives them exactly what they need.
This is one of the things we’re building into the SKAW tool — visibility into your bot traffic so you can make informed decisions rather than blunt ones.
A lot of people are asking the wrong question when it comes to AI search. “Why don’t I appear at the top of ChatGPT?” is the wrong obsession. The right question is far more precise: when my perfect customer asks the perfect question that only my business can answer, in the right location — am I there?
If the answer is no, you have a site problem. Not a content trick problem.
Writing a new blog post to game AI rankings is not the play. What actually matters is addressing your customers’ real pain points in your content and product descriptions, and making sure every piece of data-driven information on your site — anything with relationships, entities, and structured meaning — is properly flagged using schema markup. The vast majority of sites still don’t do this correctly, and it is costing them visibility with AI systems every single day.
SKAW is being built to help you see exactly where those gaps are and fix them. We’re testing on a number of sites at the moment and the data is promising — but I’m not releasing it until I have solid before-and-after case studies to show you. Watch this space.
I posed a provocative question on LinkedIn this week: is WordPress a solution looking for a problem that no longer exists?
If a personal AI assistant can build a website, update it on demand, and manage it continuously — why maintain a traditional CMS at all? The vision isn’t far off: an agentic webmaster that monitors your site, makes updates when you send a text via Siri, and handles routine maintenance in between. Whether or not WordPress survives this shift, the question itself signals how dramatically the operating model for web businesses is about to change.
It’s going to be a very interesting 12 months.
If you have questions about any of this, or if you’re interested in a workshop or webinar on building an agentic business operating system, reach out. You can contact me at [email protected]. If there’s enough interest, we’ll make it happen.
And as always — like, share, subscribe, and tell your friends. Thank you.

Jim’s been here for a while, you know who he is.