Skip to content
Dixon Jones

Dixon Jones

CEO, Inlinks and Waikay

Inlinks and Waikay

At a Glance

Role
CEO of Inlinks and Waikay
Co-founded
Majestic, winner of two Queen's Awards for Enterprise
Career
In digital marketing since 1999
Latest tool
Waikay, short for What AI Knows About You
Dixon Jones is CEO of Inlinks, an entity SEO tool, and of Waikay, which asks AI models what they know about a brand and checks the answers against the facts. He co-founded Majestic, the search engine that maps the web's links, and has worked in digital marketing since 1999. Where most GEO tools count mentions, Jones asks whether the model has the brand right.

Majestic and the Link Graph

Jones was one of the founders of Majestic, a search engine built around links instead of keywords. Its index became a standard reference for SEOs, and the company won two Queen's Awards for Enterprise, which led to invitations to Buckingham Palace. He later became its brand ambassador and has said he is no longer part of the company, only an ambassador for it.

The link work still shapes how he thinks about AI. On the BuzzStream podcast he said he keeps hoping someone from the language model world will decide to use Majestic's data, “because Majestic has this brilliant proxy for PageRank.” He compares today's AI visibility tools with the rank trackers of the past: they report where a brand ended up, not how it got there.

Inlinks and the Idea of Entities

At Inlinks, Jones works on entity SEO, the practice of making clear what a page is about and how its topics connect. He explains it with an example. One set of pages might discuss mustangs as wild horses, tied to nature, and another might discuss the Ford Mustang, tied to petrol and engines. “The words surrounding the term ‘Mustang’ change its meaning and help determine what it refers to,” he told an SEO interviewer in October 2026.

The lesson for a brand is to stay on topic. A blog post about the company holiday party will not help a machine decide what the business does, he said, while a post explaining a specific feature will. At the Kalicube Summit in June 2026, hosted by Jason Barnard of Kalicube, Jones spoke from the knowledge graph side in the sessions on how engines annotate content.

What AI Knows About You

Waikay, short for What AI Knows About You, started from a simple test: ask a model what it knows about a brand in the context of a product, reduce the free-form answer to one-line facts, and let a person check each line. Jones told the BuzzStream podcast that internal research suggested hallucinations about brands run around 8 percent on average.

His own company supplied the example. An AI answer described Inlinks as a rank checking tool, which it is not. He said the team traced it back to one positive review of Inlinks that credited the product with a feature it does not have. The page publishing that review was wrong, so Jones contacted its owners, they improved the content, and Inlinks got a link out of the exchange. Our AI brand misinformation statistics show how common such errors are across models.

Prompt Tracking Is the Wrong Place to Start

Jones is skeptical of dashboards that run hundreds of prompts. “I think AI prompt tracking is the wrong place to start, but the right place to look at the results,” he told the SEO Confidential interview. “It tells you where you've ended up, but it's less useful for understanding how you got there.” He recommends tracking only the prompts that matter to the business and writing them neutrally, so that a question like “What is the best CRM for small businesses?” lets the model mention several brands instead of the three the asker already picked.

To make comparisons fair, Waikay also measures what he calls topical presence: the context in which a model recommends a brand. Booking.com might be called great for a wide range of options while EasyJet is described as a budget choice, he explained, and both statements can sit in one answer. That caution echoes the doubts Barry Schwartz and Rand Fishkin have raised about prompt tracking, and our AI answer volatility statistics quantify them.

Say It Plainly on the About Page

His practical advice is plain. Clarity beats cleverness: an About page that states who the chief executive is, where the headquarters sits, and what the business does helps a machine more than a story about a founder having the idea on a beach in Bali. He points to open standards such as EntityMap for going further, and says that when a model gets something wrong, the fix is to trace the source, whether an out-of-date page of your own or a third-party review, and correct it.

Jones has also written a book, Power in the Modern World, on the risk that models trained on their own output will favor the average answer. The founder who spent a career counting links now spends his days asking machines what they believe, and then checking whether it is true.

Links

← All people

Sources