AI wrote half the internet. A consultant's byline just got scarcer.

Start with the number that should worry anyone who thought AI made publishing pointless: 74.2%.

In April 2025, Ahrefs sampled 900,000 newly created English-language web pages and ran them through an AI content detector. 74.2% contained AI-generated text. Only 25.8% were written entirely by a person. Separately, Graphite tracked 55,400 articles back to before ChatGPT existed: about 10% of new articles were AI-generated then. By the first quarter of 2026, that share had climbed to 49.9%, and it has held near that line for five straight quarters.

That's the flood. Half of what gets published now is generated, not written. And the common read on that number is exactly backwards.

The flood is real, and it isn't slowing down

The instinct, when a resource gets this cheap this fast, is to assume it gets less valuable to have more of it. If a machine can write a 1,200-word article on your topic in nine seconds, why would a client pay for your take on the same topic?

Read the breakdown again, though. It isn't that AI writing replaced human writing one-for-one. It's that AI writing got added on top, at a volume no consultant, coach, or agency was ever competing against before. The 25.8% of pages written entirely by a person didn't shrink because people stopped writing. It shrank as a share, because the denominator exploded.

Every search result, every feed, every "top 10" roundup a prospective client scrolls through now carries roughly even odds of being machine-written. That changes what stands out. It doesn't matter how good the machine gets at sounding confident. What it cannot do is have actually run the engagement, sat in the room, or made the call that went wrong and learned from it.

What gets scarce when content gets free

This is basic supply and demand, applied to something most experts never had to think about as a supply: words. When the supply of generic, plausible-sounding content approaches infinite, its price approaches zero, and it did. Nobody pays for a blog post anymore. Nobody was going to pay a consultant for one either.

But nothing about that flood adds a single hour of real client work to the machine's résumé, or a single verifiable name to its byline. The thing that was always scarce, a specific person's judgment, tested against real engagements and signed with their name, just got more scarce relative to everything competing for attention around it. Not because it changed. Because everything around it multiplied.

Buyers already told researchers they trust published thinking more than marketing materials, 73% of them, in Edelman and LinkedIn's B2B survey. That was true before the flood. It matters more now, because "published thinking" is no longer competing against silence. It's competing against an ocean of AI-assisted text that reads fine and proves nothing about who wrote it or whether they've done the work.

Buyers already distrust the robot, and they say so

This isn't theoretical. It shows up directly in how B2B buyers behave. TrustRadius surveyed 1,862 technology buyers in January 2026 for its B2B Buying Disconnect Report. 63% had used AI somewhere in their purchase research. Of those, 94% said they fact-check what the AI told them, at least some of the time. And the share of buyers who trust online research resources less than they did a year earlier rose from 39% to 47%.

Buyers are using the tools and not believing the tools, at the same time. That combination is exactly what makes a real name attached to real work more valuable, not less. A generated answer is where a buyer's research starts. A verifiable byline, work they can trace to a person with a track record, is what a buyer now has to go find to actually trust what they read.

The audit a machine can't pass

Here's the practical version of that shift. A buyer who has just read three AI summaries of your field is, by their own admission, more skeptical than they were a year ago. When they finally land on a piece with a name on it, they run a fast, mostly unconscious audit: does this sound like it came from somewhere real? Specific numbers, a stance the writer would have to defend, a client detail that couldn't have been guessed. That's the check a generated article fails and a genuinely published one passes. It's the same audit buyers run for months before they ever call, now running against a noisier field of candidates.

The byline doesn't need to be dramatic to pass. It needs to be consistently, verifiably a real person with real engagements behind it, which is a lower bar than it sounds like and one almost nobody clears on purpose.

Your practical step

This week, make your byline pass the audit:

1. Put your name, a real one, on every piece you publish, on a page you own, next to a specific client detail or a number you can defend if someone asks.
2. Cut anything in your last three published pieces that reads like it could have been written about any firm in your field. If a competitor could paste their name over yours, rewrite it.
3. Link your byline to a page that shows more of the same work, so the audit has somewhere to go past the first piece.

None of this requires writing more. It requires making sure what you do write couldn't have been generated, and is easy for a skeptical buyer to verify came from you.

Nashra keeps your byline attached to a real, growing archive instead of a scattered feed: every piece you publish lives on a page under your name, and every reader it earns lands in your list, not a platform's. The flood doesn't go away. Your name just has to be findable inside it.

Sources: Ahrefs AI content study, 900,000 pages (April 2025); Graphite State of AI Content report (Q1 2026); TrustRadius 2026 B2B Buying Disconnect Report, 1,862 buyers (January 2026); Edelman–LinkedIn B2B Thought Leadership Impact Report (2024).