Lisa Larson-Kelley, a friend and the CEO of the B2B marketing firm Quantious, published a piece in Fast Company last week with a word I wish I had thought of first. The piece is "The 'enslopification' of B2B marketing is here," and her definition is the whole argument in one sentence: "'Enslopification' is what happens when content trends toward the same AI-flattened, meaningless mush."
She borrowed the shape of the word from Cory Doctorow's enshittification, and the borrowing is fair. Doctorow described platforms getting worse on purpose. Lisa is describing marketing getting worse by default.
I read AI output all day. Ours, our clients', and every month the answers AI engines give about the businesses we work for. So I want to do two things with her piece. Agree with it in public, and add the part I can see from where I sit, which makes her argument more urgent than she had room to make it.
I. What She Got Right
Lisa's team ran an internal workshop to find out whether they could tell AI-made text and images from the real thing. They could. "Everyone was able to spot the AI giveaways," she writes, and the verdict on the synthetic work was that it was functional and nothing more.
That matches what I see. A slop paragraph is rarely wrong. It is unowned. Nobody in particular could have written it, and a reader can feel that by the second sentence even when they cannot say why.
Her five responses are the right five:
- Share a point of view only your organization can hold.
- Make content that earns attention instead of assuming it.
- Build trust early by showing your work, with concrete examples instead of claims.
- Keep human judgment in the loop.
- Build community, so a buyer can tell there is a person on the other end.
If every B2B marketing team did those five things, most of the slop would disappear. And the line in her piece that I keep coming back to is a question: "Your team might use AI, but do they think with it?"
That question is the hinge. Everything below is an attempt to answer it.
II. Why the Mush Is Always the Same Mush
Lisa's workshop could spot the giveaways because the giveaways are consistent. The same cadence, the same hedges, the same three-item lists, the same "in today's fast-paced landscape." That consistency is not a bug in one tool. It is the whole class of tools working as designed.
In September I wrote a response to Aaron Horwath's Noema essay on the same symptom, and the mechanism is short enough to repeat here. Every language model learned from the same object, human language, and every one of them converges on the same underlying map of meaning. Ask a model for "a blog post about cybersecurity for mid-market CFOs" with nothing else, and it walks to the most probable spot on that map. The most probable spot is the average of everything anyone has written on the subject. That is what slop is: the center of the map, delivered on request.
Slop is not a malfunction. It is what the tool does when nobody shows up.
Drew Thomas Hendricks
So "do they think with it?" is exactly the right question. A model with nobody across the table produces the average. A model with someone across the table who brings facts, customers, judgment and a position produces something that belongs to that person. The tool is the same in both rooms. The chair is not.
III. The Second Reader
Lisa's piece is about the human reader. Her team's test for every post is a good one: "Would this content stop me mid-scroll?" If it would not, it does not ship.
There is now a second reader, and it does not scroll.
When a buyer asks ChatGPT, Perplexity or Google's AI Mode which firm to shortlist, the engine reads the web and assembles an answer. We built Eidomai to read those answers for our clients every month, and to check whether what the engines say about a business is true. The pattern is consistent. An engine does not need your page for the average answer. It already has the average answer. It was built from it.
What an engine can use a page for is the part it could not produce on its own: a specific fact, a verified number, a clear position, a named thing only you did. That is what gets attributed. Slop has none of it, so there is nothing to attribute, and the business goes unnamed.
Slop now costs twice. Once with the person who scrolls past it, and once with the engine that never names you.
Drew Thomas Hendricks
This is the part Lisa's argument needs. Enslopification was always going to lose readers. It now also loses the recommendation, in the one place a growing share of buyers go to ask for one.
IV. Her Five Fixes, Read by the Second Reader
Here is what happens to each of Lisa's five responses when the engine is also in the audience.
A distinct point of view. For a human reader, a point of view is what makes a piece worth finishing. For an engine, it is the only thing that separates your page from the thousand others it already summarized. A business is a region on that shared map of meaning, and a region needs edges. A position gives it edges.
Content that earns attention. For the engine, attention is a citation. What earns one is a passage that answers a specific question in a self-contained way, with something in it the engine could not have said itself. The short, visual, conversational work Lisa's team favors does well here for the same reason it stops a scroll: it says one thing clearly.
Show your work. Lisa asks for concrete examples instead of vague claims. The engine is stricter. A verified number with a source is the single most citable thing a page can carry. "We help companies grow" cannot be attributed to anyone. A named client, a real number and the quarter it happened in can.
Human judgment in the loop. Lisa tells a story about a client brief that looked organized and was not. "The deliverables list didn't make sense, the due dates were unreasonable," and it took nearly two days before anyone caught it. Notice who caught it. Not a reviewer checking grammar. Someone who knew the client. Checking whether a paragraph is well-formed is cheap and getting cheaper. Checking whether it is true of this client, this month, requires a person who knows the business. That was always the job, and it is the Cookbook Problem again: the recipe was never the hard part.
Community. Lisa suggests advisory boards and office-hours calls. I would add that community is where everything in the first three fixes comes from. The question a customer asked on Tuesday is not in any training set. The objection that comes up on every third sales call is not either. Community is the supply line for the facts, the positions and the examples that a model cannot have, because nobody ever wrote them down anywhere a model could read.
V. Two Tests Before It Ships
Lisa's test, for the first reader: "Would this content stop me mid-scroll?"
Mine, for the second: could anyone else have written this?
If a competitor could put their logo on the piece and nobody would notice, the engine has no reason to put your name on it either. Run both tests. A piece that passes the first and fails the second will get read and will not get recommended. A piece that fails both is slop, however good the prompt was.
VI. The Threat to Small Agencies Is Real, and So Is the Way Out
Lisa does not soften this part: "This is the biggest threat to B2B marketing, and dare I say small agencies."
She is right, and I run one. A small agency that sells output is now competing with a twenty-dollar subscription, and it will lose that comparison every time, because the client can see the output and cannot see the difference.
The way out is to sell what the subscription cannot have. A small agency that owns the context, meaning the client's real facts, their customers as they actually talk, the history of what worked and what did not, and the thing that happened on Tuesday, is selling the one input every model lacks. That is the bet I have made for Nimbletoad: own the context, not the tasks. It is also, I think, what Lisa's five fixes add up to when you read them as a business model instead of a checklist.
VII. Stop the Slop, Twice
Lisa closes with a line I will borrow outright: "For the future of our industry, let's stop the slop."
I would add the reason she did not have room for. The engines are reading too. They already hold the average, and they have no reason to name a business that gave them nothing beyond it. Every piece that passes her test and mine is a piece that can earn a reader and a recommendation. Everything else is mush, and the mush is the same mush everywhere.
Lisa Larson-Kelley is the founder and CEO of Quantious, a B2B marketing firm she built from a two-person shop into a team working with Meta, Salesforce, Shopify and Adobe. Her piece is worth reading in full at Fast Company.