How we research and update these guides

Who writes these guides

The advisible team writes them: the people who build the product and write its documentation, not a byline of invented contributors.

We publish under one name because there is no ghostwriting here. Where a guide reflects a judgement call rather than a checked fact, it reads as such.

Where the facts come from

Vendor pricing, feature and documentation pages, engine documentation, primary research we link at its source, and our own product work for anything about how advisible behaves.

• Vendor pages first. A claim about a competitor comes from that competitor's own pricing page, docs or feature pages.
• Engine documentation for how ChatGPT, Perplexity, Gemini, Grok and Claude capture sources, rather than someone's summary of a changelog.
• Primary research for external studies: we link the paper or dataset and paraphrase rather than lift sentences.
• Our own product work for measurement claims about advisible, such as daily runs, sample sizes and how confidence intervals are shown.

A fact we cannot reach at its source gets deleted rather than softened. That rule removes tempting lines every month, usually numbers recycled between listicles.

How competitor facts are checked

On the vendor's own site, on the day of writing, with the date printed on the page. Every comparison page carries a “Checked on YYYY-MM-DD” line.

Prices come from the vendor's own pricing page, never from a listicle, an affiliate roundup or a forum post. If a vendor does not publish a price, the guide says the price is not published.

Feature claims follow the same path: the docs or product page is the source, screenshots beat memory, and anything we could not open on the checking date is cut. Our own product's limits get the same treatment, since a comparison where only one side is scrutinised is marketing.

How guides are updated and corrected

Guides are re-checked when engines or pricing change, dateModified is bumped whenever content changes, and corrections come in through the feedback link or our contact address.

1. Report it, with the sentence and, where you have one, the source that contradicts it.
2. We verify against the source named in the guide or the vendor's current page.
3. We fix the text and bump the date. A wrong fact left standing ages the whole page.

The loop also runs without reports: when an engine adds or drops a source type, or a vendor republishes pricing, the affected guides get re-read and re-dated rather than trusted from memory.

Our limits, stated plainly

We make the product we compare. That biases us, so the page says it out loud and keeps every competitor fact checkable in one click.

No studies we have not run, and no statistics without a sample behind them. Measured figures carry their sample size, approximations are marked est., demo content is badged Sample data. Coverage claims stay inside what the product does today: five engines run daily, one market in English, unlimited team members, and access on your own provider keys.

For the product side, start with how to choose an AI visibility tool for the criteria we would apply to ourselves, or advisible vs Peec AI for the head-to-head with its Checked on date.