Personal AI Adoption Is Outpacing Institutional Adoption

Workers are already using AI on their own. A lot of the companies they work for still don’t have a plan for it.

Today I want to talk about something I keep seeing. AI adoption is happening at the personal level, and it hasn’t reached the institutional level yet.

What does that look like? It looks like the worker at XYZ company who is more than likely already having some kind of interaction with AI, whether or not the company has said anything about it.

The spectrum of personal use

If you lined up how people use AI on their own, I think it would look something like this, from simplest to most advanced:

1. Basic chatbot questions. You open a chat window and ask it something.
2. Everyday tasks. “Make me a spreadsheet.” “Edit this PDF.” “Make me a slideshow.”
3. Producing software. The easy end of this is a script that makes some part of your day easier.
4. An actual software factory. At the high end, people are building systems that produce an outcome.

People are using AI at scale along that whole spectrum. Some are only asking questions. Some are building real tools.

Why personal adoption is easier

I think it’s much easier to adopt AI at a personal level, and there are a few reasons for that.

The first is security. As an individual, you aren’t the one who has to sign off on security decisions for the whole company. You can just start using it.

The second is cost. A personal plan is cheap, and you can sign up today without asking anyone.

The third is the lack of bureaucracy. There’s no committee and no approval process. Nobody has to hold a meeting before you try something.

The fourth, and I think the biggest, is leverage.

The leverage math

Here’s how I think about it. Say you pay $100 to $200 a month for an AI plan, and it handles about 20% of your workload. If you’re getting paid $7,000 a month, 20% of that is about $1,400 a month worth of labor. That’s a lot of return for a plan that costs a fraction of that.

I know of one real example. There was an administrative worker in Russia, paid about $3,000 a month, who had already offloaded about 15 to 20% of her work to Claude. She was doing it on a $20 a month plan.

As I had more conversations with her and talked through what was possible, she moved closer to 20 to 25% of her workload. Most of that was the rote administrative work. Claude was running scripts, producing forms, and doing a lot of the heavy lifting on the administrative side.

Do the math on that. 20 to 25% of $3,000 is about $600 to $750 a month of work, handled by a $20 plan. She may have upgraded since then. I haven’t talked with her lately. Either way, it shows how much one person can get done on their own when they know what’s possible.

Why institutions are late

So what I see is that the institution is late to this. And I understand why.

Adopting AI across a company is costly and difficult. It takes a lot of decisions about permissions, security, ingress and egress, and a lot of other pieces. Many parts of that conversation are very technical.

Now think about the typical 5 to 20 person company. They don’t have a CTO. Nobody is in charge of AI or information systems. IT is probably outsourced to a managed service provider, an MSP.

A company like that may not be able to make any real decision about AI. The only decision available to them is something like: “Okay, we’re going to give everybody $20 or $100 ChatGPT plans. Go to work.”

There’s no implementation plan behind that. The message to the staff is basically “figure it out yourself, and maybe watch some YouTube videos.”

Why it takes more than a subscription

Getting AI working at the institutional or enterprise level is a lot harder, because it actually takes a plan, a team, a leader, and capital.

A lot of businesses don’t see the advantage of investing in AI at that scale. They think the solution is $20 plans for everybody, or $100 plans for everybody. Good luck, go at it, watch some YouTube, and add some value to the company.

I think that’s a fair summary of why AI diffusion has stalled a bit at the institutional and enterprise level. Individuals are moving ahead on their own, and the companies they work for are behind. There’s no one click button that fixes that.

There is an easy button

Thankfully, this is what we specialize in at Agints. We build custom AI agents for business, and we help companies put AI to work across a whole team.

If you’re part of an organization, or you head a team, and you’d like to implement AI, and maybe agents too, because you want to reduce the busy work, this article is meant to wake you up. You’re not alone, and an easy button does exist.

---

Go to agints.com/free and talk to Gini, Agints’ 24/7 AI phone receptionist and chief of staff. Tell her where you’d like to implement AI. We’ll put together a plan and tell you how quickly and at what price we can get you set up.

The goal is to make your life easier and your staff’s life easier, and to get you the productivity gains that don’t come from a $20 a month ChatGPT plan alone.

Get the next one in your inbox → agints.substack.com