Of the 80 sites whose page text we could rescore, 20 use somewhere. A separate 20 use . Both are on a quarter of the sites in our sample, and both sit on SiteTell's ruleset as vocabulary tells.
The word is not rare, then, and it is not wrong in any grammatical sense. The problem is narrower and more useful to know: it makes a promise about something being absent (friction, a gap, a join), and it almost never says what that something was. A reader has no way to check it, so it slides past as noise.
- Rule
- Type
- vocabulary · AI cliché
- Severity
- medium
- Share of sites
- 25% (20 of 80)
- Instances
- 297 on 219 pages
- Per 1,000 words
- 0.037
Why does “” sound like AI wrote it?
Because it has become the default adjective for any two things working together. Integrations get it, onboarding flows get it, checkouts get it. When one word does that much work across that many unrelated products, it stops describing any of them.
Our data adds a second reason, and it is the stronger one. Pages carrying the word also tend to carry a whole set of stock vocabulary beside it. Measured as lift (how much likelier a pair is to share a page than chance alone would predict), a page containing it is about 20 times likelier than an average page to also contain , and nearly 18 times likelier to contain . Those pairings hold across at least five separate sites each, so they describe a habit, rather than one site's template repeated.
That clustering is what a reader picks up on, even without naming it. One stock word is a slip. Several, arriving together, read as a register: copy produced to sound like marketing, whoever or whatever produced it.
Where on a site does it show up?
Mostly where the pitch lives. Normalised per 1,000 words, the rescored corpus puts it at 0.289 on product pages and 0.196 on homepages, against 0.029 on blog posts. That is roughly a tenfold gap between product copy and editorial copy.
Two caveats travel with that ratio. Only five sites have a product page using the word, and the heaviest of them contributes about a third of those hits, so treat the gap as a direction rather than a measurement. And "product page" is a guess from the URL (anything under /product, /features, /solutions or /platform), which is right for most sites and wrong for some.
The adverb follows a similar pattern, with product pages again the densest bucket at 0.239 per 1,000 words.
- Rule
- Type
- vocabulary · buzzword
- Severity
- high
- Share of sites
- 25% (20 of 80)
- Instances
- 355 on 254 pages
- Per 1,000 words
- 0.044
- Plain swap
- smooth(ly) / easy
What does it look like, and what replaces it?
Each pair below is written by us for this page, since quoting a real site's sentence could identify it. Each one is also tested on every build of this site: the flagged version has to trip the rule named on it, and the rewrite has to pass every rule SiteTell can judge from a single passage. So "flagged" and "passes" are measured, not asserted.
The rewrite answers the question the adjective dodged. What exactly used to be painful? Retyping. So it says that.
Here the replacement is a number, which is the most convincing thing a sentence about smoothness can carry. Five seconds can be checked. A promise of smoothness cannot.
The last one trades a vague reassurance for a specific offer. It also admits that something might not carry over, which a nervous buyer will believe far sooner than "quick and painless".
Is it ever fine to use?
Yes, and the rule will flag it anyway. It matches the word wherever it appears, with no sense of context, so a textile brand selling knitwear, a manufacturer of steel tube, or a design tool generating textures will all see it in their report.
That is a deliberate trade. SiteTell shows the exact sentence and the rule that fired for every flag, so a literal use is obvious at a glance and costs nothing to ignore. Marking it ambiguous, as the ruleset does for words like and , would undersell how rarely the literal sense turns up on a software or services site.
What should you write instead?
The product's own swap for the adverb is "smoothly" or "easy", and neither is a real fix on its own, because each is just as unverifiable. Better to replace the claim with the fact behind it. Three questions usually get you there:
- What step does the reader no longer have to do? Name it.
- How long does the thing take now? Give the number, even a rough one.
- What could still go wrong? Saying so is often the most persuasive line on the page.
If none of those has an answer, the sentence probably had no content to begin with, and deleting it is the honest edit.