TL;DR
Putting “AI-era relevant” on a profile is only as strong as the work behind it. I ran two real AI-assisted content engagements and the same systems-building instinct shows up in both: write the spec yourself, keep the human check in the loop, and keep the source traceable. One happened just before the AI shift went mainstream, one just after; the claim is scoped to exactly those two contexts.
How “AI-era relevance” gets claimed
The usual way a senior profile answers the AI question is a line: tools named, a promise of AI comfort, or a defensive silence that reads to you as “pre-AI.” A claim is cheap to make, and the problem is what happens when it meets a screen built to remove un-evidenced statements.
AI, operationally, is another system to design: you write the spec, you calibrate, you test, you keep the human check in the loop, you make the source traceable. That is the register this conviction belongs to. It is also the register of the systems that scale I build.
Where the claim breaks
The concrete failure I watched happen, not hypothesized: even where automation was wanted, AI relevance had to be evidenced and bounded to be trusted.
At Simplébo, a SaaS web agency, I ran a proof of concept for AI-assisted page-content production. The director was a firm advocate of automation yet still feared losing the human side of production, and ending up with similar content across client sites. The deliberate decision was to keep the human at the heart of the machine, using part of the tool’s potential rather than a specialist orchestrating AI-validated content alone.
Why the claim carries no signal
A claim is cheap to make and cheap to dismiss, so it proves nothing on its own. Two angles a reader can hold:
- A line about AI is no signal: anyone can write it, so it does not show whether the person can actually work with the technology.
- An operations profile written in a pre-AI grammar can be out of date, so the burden is real, not manufactured, and the trustworthy signal is the method that produces the work, not the claim about it.
What dismissing the pre-AI builder costs
The instinct that makes systems work is not new. It did not appear because AI became mainstream, and the operator who learned it before that moment still has it; dismissing them outright throws it away. Accepting un-evidenced AI claims admits noise into the screen. Both errors are expensive, and evidence is what resolves the trade-off.
The same instinct, twice
Two engagements, argued as the same systems-building instinct on each side of the moment AI went mainstream: one just before it, one just after. Three shared behaviors, each visible in both.
You write the spec yourself. At Simplébo, after auditing specialized editorial AI tools, I let the content teams pick the tool they were most comfortable with, then put a process and a matrix in place to show capacity and size the time needed. At Aprovall, a compliance-software business, I wrote precise test-and-learn scenarios myself, using the tool’s spaces as a file system of specifications and instructions, and calibrated them until the first articles held.
You keep the human check in the loop. At Simplébo, that was the human-at-the-heart decision. At Aprovall, the first articles were validated by the marketing lead and the legal content referent, the internal re-reader had very few remarks, and every piece of feedback fed back into the specifications, so the process improved as it ran.
You make the source traceable. At Aprovall this mattered most. In a compliance sector, the truthfulness of arguments, texts, figures, and sources was imperative. The traceability history meant any delivered article could be justified or corrected, even on a late review, and that rigor was never once questioned.
At Aprovall the collaboration ran about six months and roughly 100 articles were produced; the legal content referent praised their quality, and I offered the client my production procedures at the end. At Simplébo, velocity absorbed the flow of client requests without compromising the more critical content work.
This is neither a pivot nor a novelty. The same instinct shows in older, pre-AI work: the editorial priority algorithm and the fraud detection algorithm predate the AI moment. The systems-building instinct is not new, which is why these two AI engagements belong on the same ground in who I am.
What these two engagements prove, and do not
The strongest objection is honest and must be said plainly: two engagements is thin evidence for a broad claim of AI-era relevance. It is.
The claim is therefore deliberately narrow. It covers AI-assisted content production in exactly two contexts: a legally constrained compliance sector, and a SaaS page-production context. It is not a claim to build AI systems, and not a claim to be a general-purpose AI expert. The proof is narrow, real, and repeatable; naming the boundary keeps it an argument, not an assertion.
A claim about the current AI moment, evidenced by work, survives the screen. If the AI-era operational question is live in your own context, let’s talk.