Does Content at Scale's AI Detector Actually Work? (It's Now BrandWell)

If you're searching for "Content at Scale AI detector," the first thing worth knowing is that the company renamed itself. Content at Scale became BrandWell in August 2024, and the detector you'll find today lives at brandwell.ai, not the old domain. The tool itself didn't disappear — it's still free, still live, and it's actually one of the older entrants in this space. The company states it built the first version of its AI detector in October 2022, before ChatGPT's public launch.

Content at Scale is now BrandWell

What started as a tool focused on generating and publishing long-form SEO blog posts expanded into a broader platform, and the rebrand reflects that. BrandWell now runs two product suites: RankWell, an evolved version of the original long-form AI writer with integrated keyword research and SEO analysis, and WriteWell, a suite of 40+ AI agents for shorter marketing copy like LinkedIn posts. The free AI detector carried over from the old brand and still works the same way it did under the Content at Scale name.

How the detector actually scores text

Paste text in, and the tool returns a color-coded read: green for a high chance of human-written, yellow when it's unclear, orange for likely AI, red for highly likely AI. It also flags specific sentences and phrases within a longer passage that read as more "robotic," rather than only giving one score for the whole document — useful if you've hand-edited an AI draft and want to know which parts still read as machine-written.

On its own detector page, the company states plainly that "accuracy can change with the ongoing evolution of artificial intelligence and large language models" and does not publish a specific accuracy percentage. That's worth noting on its own: several third-party review and comparison sites do cite specific numbers for this tool (we found claims in the 46-52% range attributed to it, against 80-84% for competitors like GPTZero and Originality.ai), but we couldn't trace those figures back to a named, credible primary source, and the company's own page makes no such claim. So we're reporting what BrandWell states about itself, and flagging — not repeating — the numbers we couldn't verify.

Why this matters if you're humanizing AI text

A detector that highlights specific sentences rather than scoring a whole document gives you a more actionable signal: you can see which parts of a draft still read as AI-generated after an edit, rather than getting one number and guessing at the cause. That's a genuinely useful design, regardless of what any single accuracy figure would tell you — the color-coded, sentence-level approach is the part worth understanding, more than any specific pass rate.

Refrazr takes a different approach at the input stage: our own free AI detector scores a passage across eight measurable patterns (sentence-length variance, AI-vocabulary density, transition-word frequency, passive-voice ratio, short-sentence presence, em-dash usage, parallel-structure density, contraction frequency) before you rewrite anything, and the rewrite itself runs as one clean pass rather than chasing that checklist item by item. Full mechanics on the methodology page.

The honest caveat, on both sides

Neither BrandWell's detector nor Refrazr's makes a claim about defeating the other's classifier, and neither should. Detector accuracy really is a moving target — models change, detectors retrain, and a number that held last quarter doesn't necessarily hold today. If a review site hands you a precise, confident percentage for any AI detector without naming where that number came from, treat it as marketing rather than measurement. That's true of claims about detectors, and it's true of claims about humanizers — including ours.

Related reading

• Does GPTZero actually work? 100-essay test
• Does Grammarly detect AI text? Official claims vs independent tests