The Brand Lost Control of the Room and the Gossip Got a Search Box

​

Unleash this on:

Original insights by Swan Sit

30-second rundown

Key learnings:

  • AI systems assemble your reputation from many voices: Your website matters, but reviews, retailers, media, forums and other sources can shape whether your brand is recommended.
  • Distinctive facts give AI something worth repeating: A memorable origin, product advantage or customer proof is more useful than another bucket of polished marketing porridge.
  • Consistency turns a claim into consensus: Repeat the same accurate brand facts across your pages and credible third-party sources so the story does not fracture in transit.

This week: Write five realistic customer questions about one important product, ask them in ChatGPT, Gemini, Claude and Perplexity, then record the brands recommended, claims made and sources cited. Compare the answers with your intended positioning and correct the largest gap on the relevant product page using a clear, self-contained paragraph.

Coming 3 months: Use an AI Search monitoring tool or scheduled workflow to repeat those questions daily across major platforms. Give one person ownership of checking incorrect claims, weak source coverage and missing recommendations, then track whether corrections appear and your brand becomes more consistent.


Swan Sit walked onto the Dreamforce stage with an ugly message wrapped in nightclub gossip: your brand is what people say when you are not in the room, and now an LLM is asking everyone questions. The brand story is no longer yours to dictate. It is being assembled by customers, publishers, retailers, critics and whatever lunatic posted a confident paragraph on Reddit at 2:13 a.m.

Sit’s answer is to make the brand distinctive enough to remember, consistent enough to verify and accessible enough for AI systems to read. That turns GEO into reputation management with receipts.

1. Your Website Is Only One Witness

Owned pages still matter, but they are the friendly witness in a cheap suit. AI systems can also encounter reviews, retailer pages, journalism, forums and research, then use that evidence to decide whether a product is reliable, sustainable or worth buying.

Sit describes these sources as the loyal best friend, measured teacher, town gossip and smiling frenemy with a knife under the table. A claim repeated only by the brand is fragile.

Compare what the company says with what the wider web supports. A running-shoe company may call one model ideal for marathons, while retailers describe it as a daily trainer and reviews complain about durability. The company does not get the final vote merely because it owns the product page.

Different AI systems may search, retrieve and cite different sources. Treat Sit’s descriptions of their preferences as a working model, not permanent technical specifications. The useful principle survives: your reputation must hold together beyond your domain.

2. Distinction Gives the Robot Jury Something to Remember

Generic praise is industrial fog. Every brand claims quality, innovation and excellent service in the same beige language. If nothing is specific, AI has nothing memorable to carry into its answer.

Sit points to her work with Elizabeth Arden. The team revived the founder’s history as a wellness pioneer and suffragette, then reimagined her as a contemporary character called Liz. Sit says traces of that distinctive campaign still surface in LLM answers years later.

Distinctive does not mean louder or stranger. Give customers and independent sources a truth they can repeat: an unusual origin, measurable product difference, service policy, original research or experience competitors cannot claim. The desert is knee-deep in adjectives. Bring evidence.

For a running-shoe seller, “premium performance” is mush. “A wide toe box for long-distance runners, supported by wear-test data after 500 kilometres” can be checked and remembered. The first evaporates under the casino lights. The second leaves footprints.

3. Consistency Builds Consensus, but the Facts Must Be Readable

Sit uses Nike’s “Just Do It” as the blunt instrument of consistency. The words, athlete choices and belief that anyone with a body can be an athlete have reinforced one another for decades. The central story kept arriving in the same battered vehicle.

Smaller companies need a short set of accurate facts used consistently across product pages, retailer listings, review responses, PR and partner profiles. Repetition across credible sources helps turn positioning into consensus. Contradictions create an evidence locker where every drawer contains a different alibi.

The facts must also be machine-readable. Product details, prices, availability, delivery terms, returns and use cases should appear as clear text, not remain trapped in images, scripts or slogans. An answer about whether a shoe suits wet winter runs should make sense even when retrieved without surrounding paragraphs.

Start with a baseline. Ask realistic questions across several AI systems, record the answers and cited sources, and measure the gap between your intended positioning and what comes back. Fix the most commercially damaging mismatch first.

The old law of reputation has acquired a search box and a frightening appetite. Earn a story worth repeating, make the evidence easy to find and keep the facts consistent. Otherwise the gossip will write the verdict, and she has never waited for legal approval.

Unleash this on:
Avatar photo
WH.

All the paranoia of a field correspondent. None of the plane tickets.

WH. has spent 14 years inside the SEO machine and started The Vector Gazette, because he got tired of watching entrepreneurs make catastrophic decisions based on advice from people who discovered GEO last Tuesday.