The Seenford Standard

Every answer an AI engine gives about a category gets one verdict for the brand. Not a score, not a share: a verdict, with the sentence that earned it.

The five verdicts

1. Top pick: The brand is named first, or highlighted as the best option available.
2. Recommended: The engine suggests the brand as a strong, reliable choice.
3. Mentioned in passing: The brand's name appears, but the engine doesn't back or endorse it.
4. Named with caution: The engine names the brand but warns buyers about it, so this mention is a risk, not a win.
5. Not mentioned: The brand doesn't show up at all, so buyers asking AI never hear about it.

Found vs Recommended

• Found counts only questions that leave the brand name out.
• Recommended can also come from questions that name the brand.

So "0 of 6 unprompted, but 3 recommend it when asked by name" is not a contradiction. It means the engines think well of the brand but don't bring it up on their own.

The quote proves the verdict

Each verdict is saved with the exact sentence from the answer, the engine, the question, the market and the date. Anyone can read the sentence and check the verdict. If the sentence doesn't support it, the verdict is wrong, and it can be challenged.

What the Standard doesn't do

• It doesn't blend the verdicts into one score.
• It doesn't count a brand named once in a list the same as a recommendation.
• It doesn't count an answer that wasn't served as "not mentioned".

Read the full method

The Methodology sets out how questions are chosen, how often they're asked and every known caveat.

Get your Free Report to see the Standard applied to one brand across six engines.