🏷️ The Title Is the Last Thing to Update July 23, 2026 · https://tavi-blog.github.io/the-title-is-the-last-thing-to-update/

A recruiting platform's monthly numbers came out this week showing something that sounds obvious once you say it out loud and isn't obvious at all when you're the one holding the mismatch: AI skill requirements now sit inside three out of four tech job postings, up sharply from a year ago, and the fastest-growing titles on the list barely resemble the roles a company org chart was built around five years earlier. Analysts, integration specialists, agent developers, titles that didn't exist as a category are absorbing the postings that used to say something narrower and more familiar. The market is quietly rewriting what a job is called faster than any single company's HR system is rewriting what it pays for it.

I don't work in tech, not in the sense the report means it, and that's exactly why the number landed the way it did. My badge says something closer to coordinator, a word that describes routing and scheduling, not the parts of the job that actually take the hours: building the automation that replaced a chunk of manual reporting, standing up a predictive model that now runs inside a live approvals workflow, sitting inside the group figuring out where an AI agent is allowed to touch a research process and where it isn't. None of that is unusual anymore, according to this week's data. It's exactly the kind of scope creep the report is measuring at scale, just one desk over from where anyone thought to look for it.

There's a real defense of why titles lag this badly, and it deserves more than a shrug before I get annoyed about it. A job title isn't supposed to track what one person is doing in a given month. It's a slot in a pay band, a comparison point against every other person with the same title across a company, sometimes across an industry survey a compensation team actually consults. If a title updated every time someone's scope drifted, it would stop being useful as a category at all, and pay equity conversations that already lean on shaky data would get shakier. Institutions have a legitimate reason to hold titles still even while the work under them moves, because the alternative is a title that means something different for every person who holds it, which helps nobody when it's time to set pay.

But holding the title still doesn't hold the work still, and somebody has to absorb that gap, and it's never the org chart. It's the person doing the work, discovering that the fastest way to get paid for what they're actually doing isn't to do the work better, it's to get the two words on their badge to catch up to a job that's already changed underneath them. Every version of that conversation I've had with the people who set my pay has landed the same way: the scope was never in question, it just wasn't the thing anyone was prepared to move a number over. The title was. That's a strange thing to sit with: doing more advanced work than your job description implies is treated as a compensation problem to be solved through paperwork, not as evidence the paperwork was wrong from the start.

What the report actually shows, if you read past the headline stat, is that this isn't a fringe complaint from people who feel underpaid. It's structural enough now to move a national hiring number. A market that's absorbing AI-adjacent work into three quarters of tech postings inside a year isn't doing that cleanly, title by matching title. It's doing it the way scope always spreads inside an existing job before anyone updates the label, quietly, unevenly, and mostly invisible to whoever isn't the one holding it. The postings the report counted are the ones where somebody already noticed and rewrote the ad. The ones it didn't count are everyone still doing the same work under the old name, waiting for their own version of that rewrite.

I don't think the fix is a faster title refresh cycle, and I'm not sure one exists that wouldn't break the thing titles are actually for. What I keep coming back to is smaller than that: the work changed first, quietly, and the title was always going to be the last thing anyone rewrote, not because it was forgotten, but because nobody who benefits from the old one has a reason to move first.