Answer
What are the 3 levels of AI adoption in a business?
Last updated
Why is a chat subscription not the whole answer?
Picture a company with fifty accountants. Ten of them, maybe only five, do the best work you have ever seen. The rest vary, and you pay everybody roughly the same, so one client gets a brilliant accountant and the next, under the very same brand, gets a very different experience. Now give all fifty a chat subscription. Does that fix it? Not really, and that gap is what the three levels are about.
If you already pay for chat tool subscriptions and train your staff to use them, you are ahead of most companies. A lot of gurus sell that as the silver bullet, and I am tired of hearing it. It is only the first of three levels, and that lead will not last.
The idea behind all three is the same: keep your headcount roughly where it is, with nobody let go, and make the same people more productive. You serve customers better, maybe charge more, maybe take on more of them. Revenue goes up while hard costs stay roughly flat, and that difference is profit.
- 01
Level 1: chat tools
Your team uses the subscriptions you pay for, trained to prompt them well.
- 02
Level 2: systems on top
Calls, leads and follow-ups run on a system, and nobody learns a new tool.
- 03
Level 3: a company brain
Your best five to ten people train it, so everyone else works at their level.
What is level 1 of AI adoption?
Level 1 is your team using chat tools. You buy the subscriptions, then show people that the tools can be trusted, how to prompt them properly and how to work with them.
The next step inside level 1 is letting staff connect those tools to what you already run, so the assistant can reach your email, your database or your CRM and do productive work there. This is very good, but it is quite limited, and that is where level 2 comes in.
What is level 2 of AI adoption?
Level 2 is automating jobs with AI: calling leads, taking inbound calls, and handling customer requests and disputes by email, messenger or phone.
Here is a day you may recognise. Your sales team takes inbound calls from people asking for a quote, and many of the callers are unqualified. I know what that does to a person, because I used to make a lot of cold calls myself: after three bad calls, the fourth is the real prospect, and you meet it less energetic, less positive and less prepared. The business loses an opportunity right there, because some leads only want to place an order, but some need persuading, and those are exactly the ones a tired team loses.
At level 2 a system takes that incoming traffic and filters it. It can also follow up with people who asked for something in the past and never heard back, and there is money sitting there.
Can it close the deal itself? For simple orders, where people know exactly what they want, yes. But that is not the goal. AI does not have a person's emotional read, it is a programmable system. Its job is to filter out the weaker and more negative callers, so your best reps get the best leads while they are fresh and prepared. That is where more sales and better clients come from.
Why does level 2 need no retraining?
A new tool or a new CRM means training and migrating everybody. Level 2 skips that: your sales people pick up the same phone and open the same CRM, and they do not need to learn anything new.
The work is outlining your processes. You explain to the system how each kind of conversation goes: is it a support ticket or a product sale, what are that product's criteria, and what must the customer answer before you can quote. In practice you draw one very big tree of every call flow, and the branches you do not want AI handling go straight to a person.
A call comes in
Is it a support ticket or a product sale?
Each kind of conversation gets its own branch of the tree.
A product sale
What are that product's criteria, and what must the customer answer before you can quote?
The system asks exactly those questions and nothing else.
A call you do not want AI handling
Anything you decided belongs to a person.
It is transferred to a human right away.
The point of the tree is to see everything that is repetitive, because that is the part a system can take. The free prompt linked in the video's description is a first pass at that map: paste it into a chat tool and it interviews you about your company's repetitive conversations and follow-ups.
What is level 3 of AI adoption?
Back to the fifty accountants. At level 3 the top five to ten train one AI system that everybody can use, one that prepares documents and communicates with clients. It is called a company brain. Once it is trained, the other forty or so work at your best people's level. Customers get a better experience, and you can charge more, take on more customers, or handle work you could not before.
It is not a closed box that works while you do nothing. It sits in the team chat you already use, or you email it. You tell it what needs doing, it follows how it was instructed, and when it has a question it asks. Tell it to do something differently and it corrects its own notes, so it gets more precise every time. It can also fill in your old CRM, go to websites and click buttons.
Access is set per person: accounting sees what accounting needs, sales what sales needs. Often only those top five to ten people can change its skills, and everyone else sends requests.
Where are most companies today?
Most are at level 0: not using it, not touching it, a little scared of it. I would expect that to change over the next two or three years as adoption moves on.
- 01
Level 0: not started
Nothing yet. Most companies are here.
- 02
Level 1
Chat tool subscriptions and training. The team learns to use the tools.
- 03
Level 2
Systems on repetitive work: inbound calls, lead calls, follow-ups. The team changes nothing, same phone and same CRM.
- 04
Level 3
A company brain. The best five to ten train it and everyone else uses it.
So which level is your company at, and which of your fifty accountants would you ask to train the brain?
Short answer
The 3 levels of AI adoption in a business are: level 1, your team using chat tools, with training on how to prompt them; level 2, systems built on top of the work you already do, such as answering and qualifying calls and following up, so nobody has to learn anything new; and level 3, a company brain that your best five to ten people train so everyone else works at their level. Most companies have not started yet. All three aim at the same result: the same team serves more customers at roughly the same cost, so profit goes up.
Justas Butkus is a fractional AI officer based in Vilnius, Lithuania, working with mid-market companies and scale-ups across the UK, EU and US at two to four days a month. He builds and operates the production AI systems he advises on, and takes no commission, referral fee or revenue share from any vendor.
If you are not sure which level you are at
Describe the conversations your team repeats every week in a short email. That is usually enough to see which level the next step belongs to.
Frequently asked questions
Does adopting AI mean cutting staff?
Not in this approach. You keep headcount roughly the same and make the same people more productive, so costs stay flat while revenue rises.
Can AI close sales on its own?
For simple orders, where the customer knows exactly what they want, yes. It is not the goal, though: at level 2 its job is to filter and qualify, and hand the best leads to your best people.
Do my staff need to learn a new tool for level 2?
No. Unlike a new CRM, nobody is retrained or migrated. Your team keeps the same phone and the same CRM, and calls you do not want AI handling go straight to a person.
Who trains a company brain?
Your best five to ten people, the ones whose work you want everyone else to match. They shape its skills, and everyone else uses it and sends requests for changes.