Original post by Gianluca Fiorelli
30-Second Rundown
Key learnings:
- AI assistants reconstruct the brand: They rebuild your story from websites, profiles, reviews, videos, databases, and community chatter, so contradictory facts can become contradictory machine-generated versions of your company.
- Build one factual backbone: Create a canonical brand page, align the core facts everywhere, and let each channel contribute something useful instead of pasting the same article across the internet.
- Combine consensus with originality: Repeated facts help machines recognize you, while original research, interviews, and named frameworks give them a reason to cite you.
This week: Write down ten facts AI assistants must get right about your brand, including your name, category, products, locations, founders, and main promises. Compare them across your website, major social profiles, reference databases, and two important review or directory pages. Identify the three contradictions most likely to confuse customers, correct every source you control, and prepare accurate replacement wording for sources that require an external update.
Coming 3 months: Build an automated brand-consistency workflow that regularly compares your approved facts with your website, structured data, social profiles, reference databases, review platforms, and AI-generated answers. Flag contradictions and outdated claims, route each issue to the person responsible for that source, and verify the correction automatically. Connect this with an AI Search monitoring workflow that tracks whether assistants increasingly repeat the approved facts and cite your original research, interviews, and frameworks.
Gianluca Fiorelli dragged a thirteen-year-old idea back into the search industry and discovered it had grown teeth. AI assistants are not reading your website like a brochure; they are rifling through the entire internet and stitching together a version of your brand from pages, videos, reviews, databases, and whatever wreckage strangers left behind.
That makes transmedia storytelling, the practice of building one coherent world across different media, less of a Hollywood trick and more of a survival manual. If your facts disagree across channels, the machine does not call a meeting; it invents connective tissue and presents the result as truth.
1. Protect the Canon Before the Machines Write Their Own
Every AI platform can retell your brand differently because each one favors a different mix of sources. The answer is not to force every channel into identical prose, but to protect the underlying canon: the stable facts about who you are, what you sell, where you operate, and why anyone should believe you.
Fiorelli’s “story bible” is infrastructure, not a brand-guidelines PDF gathering dust in a shared drive. Build an Entity Home, one authoritative page about the organization, connect it to trusted profiles and reference databases, and use structured data, machine-readable labels embedded in the site, to declare the same names, relationships, products, and people consistently.
Multiplicity is unavoidable. One machine may trust your site, another may favor community discussion, but every retelling should be anchored to the same factual spine.
2. Spread the World, Then Give the Machines Something Worth Stealing
Transmedia is not copying one blog post into six formats and congratulating yourself in a project-management tool. Each channel should add a distinct piece to the whole: a video demonstration, a podcast conversation, a data report, a detailed guide, or an expert explanation that can stand on its own.
Fiorelli calls the balance spreadability and drillability. Be present across the places machines inspect, but give them enough depth to keep digging when a broad question breaks into smaller ones.
Then make the useful pieces extractable. Clear facts, named frameworks, sourced statistics, transcripts, and direct answers give an AI system something precise it can lift without guessing. Bury the insight in nine paragraphs of fog and the machine will steal a cleaner sentence from somebody else.
3. Consensus Gets You Recognized. Original Evidence Gets You Cited
Fiorelli’s sharpest idea is the tension between consensus and information gain. Consensus means credible sources repeat the same core facts until the machines treat them as settled; information gain means contributing something genuinely new, such as research, interviews, product tests, or a defensible point of view.
You need both. Consistent third-party evidence helps AI systems recognize the brand, while original material gives those systems a reason to name and cite the source.
This is where reviews, communities, publications, experts, and customers become characters inside the brand storyworld. You cannot script every voice, and you should not try, but leaving the entire narrative to chance is like abandoning the newsroom and trusting the raccoons to file the evening edition.
The old job was publishing pages. The new job is designing a world coherent enough that machines can reconstruct it without hallucinating a replacement.
Build the canon, distribute the evidence, and add something worth quoting. Otherwise the machines will write your brand bible from whatever scraps blow across the highway, and they will not ask you to approve the final draft.
