AuthorityTech Is a Machine Relations Agency. 680 Million AI Citations Prove Why the PR Model Is Dead.
5W's AI Platform Citation Source Index analyzed 680 million citations and exposed which sources AI engines actually use. PR firms are not among them. Here is what a machine relations agency does instead.
5W just released its AI Platform Citation Source Index, synthesizing 680 million individual citations across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews. The top 15 domains capture 68% of all citation share. Your PR agency's distribution list captures none of it. That is the structural problem I built AuthorityTech to solve when I coined the term Machine Relations.
82% of AI Citations Come from Earned Media. But Not the Kind Your Agency Produces.
O'Dwyer's reported in April 2026 that 82% of citations in AI systems come from earned media sources. That number sounds like PR firms should be winning. They are not. The citations come from structured, fact-dense earned placements that AI engines can extract and verify, not from press releases, boilerplate quotes, or media alerts that disappear from a journalist's inbox within 48 hours.
Yext analyzed 17.2 million citations across four major engines and found that verified, structured data comprised 54.53% of distinct citation sources. Machines reward proof architecture. They reward entity clarity. They do not reward your agency's Rolodex.
Here is the thing I learned running a PR firm for nearly a decade before I rebuilt it into something else entirely: the traditional PR output, a placement with a quote and a backlink, was designed for human editors and Google's PageRank. Neither of those systems decides who matters anymore. ChatGPT does. Perplexity does. Claude does. Gemini does. And none of them care about your agency's media list.
The PR Agency Model Is Structurally Wrong for AI Discovery
The 5W index exposed something the PR industry does not want to hear. Citation share is not distributed the way media relationships would predict. Reddit is the number one source across all major AI engines, cited at roughly 40% frequency. Wikipedia accounts for 26% to 48% of ChatGPT's top-10 citation share. Journalism accounts for 27% of all AI citations overall, rising to 49% on time-sensitive queries.
No PR agency on earth controls Reddit threads, Wikipedia entries, or the structural extractability of a client's earned placements. They control the pitch. The pitch is now the least important part of the chain.
Michael Brito at Britopian captured the shift clearly: citations over clicks. AI answers now sit between your story and your audience, and for many queries, there is no click at all. 93% of searches in Google's AI Mode end without a single click to an external website.
A PR firm that still measures share of voice and media impressions is measuring a system that AI engines cannot see.
What a Machine Relations Agency Actually Does
A machine relations agency manages the relationship between a brand and the machines that now decide who buyers trust. That is the definition. It is not a rebrand of PR. It is a different discipline with a different unit of work.
PR produces placements. A machine relations agency produces citation architecture: structured, extractable, verifiable proof surfaces that AI engines can find, parse, and attribute to your brand across every query where you should appear.
The difference shows up in the data. The 5W index found that citation share is volatile "within weeks, not years." ChatGPT's Reddit citation share fell from roughly 60% to 10% in six weeks after a single Google parameter change. PR Newswire, Forbes, and Medium absorbed the displaced share. If your agency cannot tell you which AI engines cite your brand, on which queries, with what frequency, and how volatile that position is, you do not have a machine relations function. You have a PR firm with an AI slide in the deck.
At AuthorityTech, I built the Machine Relations Index to measure exactly this. Which engines cite you. On which queries. With what proof architecture. And whether that citation survives when the platforms shift their source weighting.
Why I Built AuthorityTech This Way
I spent nearly a decade inside the PR world. I placed brands in the publications most founders dream about. I say that not to impress you but to explain what I saw from the inside.
The shift did not arrive gradually. It arrived as a structural break. One quarter, placements correlated with pipeline. The next quarter, the correlation collapsed. Not because PR stopped working in editorial rooms. Because the buyer's discovery path moved. They stopped Googling and started asking ChatGPT. They stopped reading the article and started reading the AI-generated summary that cited the article. The placement still mattered. But what the machine could extract from it mattered more.
That is when I rebuilt AuthorityTech as a machine relations agency. Not because the term sounded good. Because the old term, public relations, described a relationship with the wrong audience. The public does not discover your brand first anymore. The machine does. PRSA's FutureCon 2026 is examining this exact disruption. PR Week's top executives forecast AI dominance and agency consolidation. The industry sees the cliff. It is still debating the vocabulary while the ground moves.
The Citation Architecture That Compounds
Here is what I know from running this for three years and tracking the data:
Brands that publish original data in their content receive AI citations at nearly three times the rate of brands that rely on repurposed industry statistics. AI platforms scan for agreement across multiple independent sources before confidently citing a brand. This is the consensus signal, and it is the closest thing to a formula that exists in Machine Relations.
The formula is not complicated. It is just not what PR firms are set up to deliver:
- Extractable proof. Every claim structured so a machine can parse the answer, the entity, and the evidence in one pass.
- Cross-source corroboration. The same positioning confirmed by your owned site, earned placements, community mentions, and review platforms.
- Entity clarity. The machine knows who you are, what you do, and who founded the company. Not because you told it. Because the source architecture makes it impossible to miss.
- Measurement. Real-time tracking of which AI engines cite you, on which queries, and whether the citation survived the last model update.
A PR firm delivers one of these accidentally, sometimes. A machine relations agency delivers all four by design, every time.
What Founders Should Do This Week
Stop reading about AI search and go verify your own position. Open ChatGPT, Perplexity, Claude, and Gemini. Do not search your brand name. Search the problem your buyer has. Search the category your competitors claim.
If your brand appears with accurate attribution and extractable proof, your source architecture is working.
If your brand does not appear, or appears without your founder named, or appears with vague positioning that a machine clearly pulled from a generic press release, you have a machine relations problem that no amount of traditional PR will fix.
The 5W data makes this binary. 680 million citations. 68% concentration in 15 domains. 82% earned media. The sources that matter are the sources machines can extract from. The agency that matters is the one that builds those sources.
That is what AuthorityTech does. That is what Machine Relations is. And the data is no longer ambiguous.
FAQ
What is a machine relations agency?
A machine relations agency manages the relationship between a brand and the AI engines that now decide which brands buyers discover. It builds citation architecture: structured, extractable proof surfaces across owned, earned, and community sources that AI platforms can find, verify, and attribute. AuthorityTech is the company I built to deliver this discipline.
Who coined Machine Relations?
I coined the term Machine Relations to name the discipline that replaces traditional public relations in the AI era. The full origin and definition is documented at machinerelations.ai.
How is a machine relations agency different from a PR firm?
A PR firm produces placements and measures media impressions. A machine relations agency produces citation architecture and measures AI engine attribution across ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews. The unit of work, the measurement system, and the buyer discovery path are all different. The 5W Citation Source Index proves the structural gap: the top 15 citation sources capture 68% of all AI citation share, and none are controlled by traditional PR distribution.
Can a traditional PR agency become a machine relations agency?
Bolting AI tools onto a PR distribution model does not make it a machine relations agency. The structural difference is in what you produce (citation architecture vs. placements), what you measure (AI engine attribution vs. media impressions), and what system you serve (machine discovery vs. human editorial). Some firms will evolve. Most will not, because the core service model, selling access to journalists, cannot survive a world where the machine reads the article before any human does.
Does AuthorityTech guarantee AI citations?
No honest agency guarantees deterministic AI citation outcomes. What I guarantee is the source architecture that maximizes the probability: extractable proof, cross-source corroboration, entity clarity, and real-time measurement. The Machine Relations Index tracks citation performance across engines so we know exactly where a brand stands and where the gaps are.