AI news & training for business owners & operators. One email. One clear next step.

THIS WEEK IN AI

This week the biggest AI labs stopped pretending. They are not selling chatbots anymore. They are selling the people, the workflows, and the operating layer that go around them. If you run a business, the question stopped being whether to use AI. It started being whether your work is ready for it.

What happened:

On May 13, Anthropic launched Claude for Small Business. It is a toggle inside Claude Cowork that connects to QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace, and Microsoft 365. It ships with fifteen ready-to-run workflows for things like planning payroll, closing the month, chasing invoices, triaging leads, and reviewing contracts. Owners approve the plan before anything sends, posts, or pays. There is no extra charge beyond your Claude license and whatever tools you already pay for. Alongside the product, Anthropic launched a free AI fluency course built with PayPal and a 10-city training tour that started May 14 in Chicago.

Why it matters to your business:

This is the first major AI release built specifically for businesses with 10 to 200 employees. Until now, the playbook was big enterprise pilots or solo founders cobbling tools together. That gap just closed. The piece worth paying attention to is not the workflow list. It is the fact that Anthropic shipped training alongside the product. They know the tool by itself does not work. Your team has to know what to point it at. If your books are messy, if your HubSpot is half-used, if your invoices live in three places, plugging Claude into that stack is going to surface the mess faster than fix it.

Source: Anthropic

What happened:

On May 11, OpenAI announced the OpenAI Deployment Company. It is a standalone business backed by $4 billion in investment from TPG, Bain Capital, Brookfield, SoftBank, Goldman Sachs, and consulting partners like McKinsey, Bain & Company, and Capgemini. The job: send "Forward Deployed Engineers" directly into client companies to redesign workflows around AI. As part of the launch, OpenAI also acquired Tomoro, an applied AI consulting firm, picking up about 150 engineers on day one. A week earlier, Anthropic announced a similar $1.5 billion venture.

Why it matters to your business:

This is OpenAI admitting in public what every operator already knows. Selling access to a model is not enough. Businesses need help fixing how the work actually gets done before AI can plug in. The biggest AI company in the world just spent $4 billion to stand up an in-house consulting arm. Read that as a signal, not a sales pitch. The companies that pull ahead this year will not be the ones with the best model. They will be the ones whose workflows are clean enough for any model to plug in.

Source: OpenAI

What happened:

On May 12 at The Android Show, Google unveiled Gemini Intelligence. It is not a new chatbot. It is an agent layer baked into Android that can read what is on your screen, move across apps, and complete multi-step tasks. The demo: long-press the power button on a grocery list in your notes app, and Gemini builds a shopping cart in your delivery app and hands it back for approval. Other examples include pulling event details from a flyer photo and finding tickets, or auto-filling complex forms across apps. Rolling out this summer on Pixel and Samsung Galaxy first, then expanding across Chrome, Android Auto in 250 million vehicles, and a new "Googlebook" laptop line later this year.

Why it matters to your business:

The phone is becoming an agent. That changes how customers will find you, buy from you, and reach you. When a customer says "find me a bookkeeper near Hampton" to their phone, an agent is going to pick one. The question is whether your business is set up to be the one it picks. That is a different game than SEO. It rewards clean websites, structured data, fast response times, and reviews that an agent can actually read. None of that is new. It is just about to matter a lot more.

Source: TechCrunch

Look at the three stories above as one story.

Anthropic just put Claude inside your books, your CRM, and your contracts. OpenAI just stood up a $4 billion company to send engineers into your office. Google just made the phone an agent. None of that requires you to do anything technical. You toggle a setting. You sign a contract. You upgrade your phone.

That is the part that should worry you.

For two years the friction was the technology. You had to figure out which model to use, how to write prompts, how to chain steps together. The friction protected you. It forced a learning curve. You could not move faster than your understanding.

That friction is gone now. You can drop an agent inside your QuickBooks tomorrow morning. You can hire a Forward Deployed Engineer next month. You can let your phone book your customer's appointment without you touching it.

So here is the question I keep coming back to with clients. If the one person who just knows how it works stepped away tomorrow, what breaks first?

Because that is what gets automated. Not the work you wrote down. Not the SOP you keep meaning to write. The actual work, the version that lives in someone's head, the version where Sarah knows to call the customer back before sending the invoice because last time we forgot and they bounced the payment.

AI does not know about Sarah. It will run the workflow you can describe. And it will run it very fast.

That is the trap. Speed without clarity is just faster mistakes. A bad invoice process automated is a bad invoice process at scale. A messy lead pipeline plugged into AI is a messy lead pipeline that closes fewer deals more efficiently.

The companies that win this year are not going to be the ones that move fastest. They are going to be the ones who slow down for a week, write the work down, find the gaps, and then plug AI in. That is not exciting. It is not what the announcements are selling. But it is the difference between AI that pays for itself in 90 days and AI that costs you a year and a client.

I am not fully convinced every business is ready for what just shipped this week. The tools are ready. The training is the gap.

Pick one workflow in your business. Just one. The first one that comes to mind.

Now ask yourself:

If I had to write this workflow down on a single page so a new hire could run it on day one, could I?

Not a flowchart. Not a Notion doc with twelve nested pages. One page. Top to bottom. Inputs, steps, decisions, outputs.

If yes, that workflow is ready for AI. Plug Claude or Gemini or whatever tool you use into it this week and start small.

If no, that workflow is the one to fix first. Not the one to automate first.

Try this in your next leadership meeting. Ask each person on the team to name their most painful workflow and tell you if they could write it down in a page. Watch what happens to the room.

If this is the year you actually want AI to pay off, the bottleneck is not the tool. It is whether the work is ready. I help businesses map their workflows and find where AI actually fits, before they buy anything. Two weeks. $2,500. Reply to this email and I will send details.

If this issue was useful, forward it to one person on your team who is about to plug AI into something.

Book the call

Thanks for reading,

Andrew Keener

Operations & AI Strategist

Keen Alliance Consulting