For the last six weeks I've been reading job postings the way I used to read ratings books. Not casually. I built a board of 92 open roles and kept it current, and 37 of them are companies hiring somebody to handle AI.
Banks. Restaurant franchisors. Hospitals. Ad agencies. Insurance brokers. A state government. Companies with nothing in common except that somebody in the building finally said "we should be doing something about this."
Here's what surprised me. The 37 postings are basically one job description with different logos on it.
If you run a company and you're trying to work out what your own AI plan should look like, that's useful. Thirty-seven companies with real budgets have already answered the question. You can just read their homework.
The six things every posting asks for
Strip out the logos and the job titles, and almost every one of these roles is the same six responsibilities:
- Find where AI actually helps. Go department by department and figure out which work is worth changing. Not all of it. Most postings are explicit that this comes first.
- Build a champion network. Find one person on each team who's willing to go first, and support them. This shows up over and over, usually by name.
- Train people by role. Not one all-hands demo. Different training for the sales team than for finance, because they do different work.
- Write the rules. What's allowed, what isn't, what never goes into a chat window. Almost always described as responsible use or governance.
- Vet the tools. Decide what the company buys and what it doesn't, and keep people from signing up for twelve things with a credit card.
- Measure whether it stuck. Adoption numbers. Hours saved. Proof, in a format an executive will actually read.
That's the job. All six, in that order, at a bank and at a chicken franchise and at a hospital.
What's not in them
Here's the part that should get your attention.
These postings go out of their way to say you don't need to be an engineer. Not "engineering optional." They say it on purpose, in plain language, because they're trying to stop the wrong people from applying.
One company wrote that it's "not a software engineering role." Another said no engineering background is required, full stop. A third said you don't need to be a developer. A fourth asked for familiarity with AI concepts and risks while specifically ruling out hands-on technical building. A restaurant company described the job as the connector between what the business needs and what the technology can do.
Read that again. Thirty-seven companies are spending real money to hire somebody, and a big share of them are saying the quiet part out loud.
The hard part isn't the technology. The hard part is getting people to change how they work.
Anybody who's ever rolled out a new phone system already knows this. The phones were never the problem.
Nobody knows what it's worth yet
The salaries tell their own story.
For what is recognizably the same job, the range I've tracked runs from the mid sixties to the high two hundreds. Same six responsibilities. Roughly a four times spread.
That's not a market with a price. That's a market still arguing about whether this is a training coordinator or a member of the leadership team.
My read, and I'll admit it's a read: the companies paying at the bottom think they're hiring somebody to teach software. The companies paying at the top have already tried it the cheap way once.
The one line worth stealing
Buried in one of these postings was the sharpest sentence I've read on this subject all year. The company wanted somebody who would redesign workflows around AI, instead of automating the workflows they already had.
That's the whole thing, in one line.
Most companies do the second one. They take a process built in 2015 for people, bolt AI onto the slowest step, save forty minutes a week, and quietly conclude the technology is overhyped.
The companies getting real results are asking a harder question first: if we were building this process today, knowing what these tools can do, would we even build it this way?
Usually the answer is no. That's uncomfortable, and it's also where all the value is.
If you're a founder-led company, you probably shouldn't hire this person
Now the part that matters if you're running a 40-person or 400-person company and reading this wondering whether you need to go find one of these people.
Probably not. Here's why.
Every one of these 37 companies is big enough to justify a full-time salary for this. You likely aren't, and that's fine, because you have three advantages they'd trade for.
You can actually decide. At a company with 2,000 people, step four on that list, writing the rules, takes a committee four months. You can do it in an afternoon. Founder-led companies move on this faster than anyone, and the speed is worth more than the headcount.
You already know where the time goes. Those big companies are paying somebody six figures to spend three months discovering which work is painful. You know already. You've complained about it out loud this week.
Your champions already exist. Somebody on your team is already using these tools on their own, probably without telling you, possibly on a personal account. That's step two, done. You just haven't given them permission or a job description.
What you're missing isn't a person. It's the playbook those 37 companies are paying to have written. The six steps, run in order, with someone who's done it before keeping you honest about the fourth one.
The uncomfortable part
I'll be straight about the other side of this.
The reason those companies are hiring instead of buying help is that a full-time person is there on Wednesday when somebody's confused. An outside partner isn't, unless the handoff is built properly. Anyone who tells you otherwise is selling.
So if you go the outside route, the test is simple: at the end of it, does your own team run this without the consultant? If the answer is no, you didn't buy adoption. You rented a dependency.
That's the standard I hold my own work to, and it's a fair one to hold anybody to.
Dave Clapper is President of AI and Marketing Systems at Taptico. Thirty years in marketing leadership, including Cox Media Group and his own venture-backed company, and now helping mid-market and founder-led businesses work out where AI actually earns its keep. If you want the six steps walked through against your own company, email dave@taptico.com.