In this episode of the Unscripted SEO Podcast, Jeremy Rivera sits down with Chris Garrett, a digital marketing veteran who connected a 2,000-student UK college to the internet in 1994 — on a 128k line he had to learn TCP/IP from scratch to configure, while his colleagues took a step back and left him to it. Thirty years of watching the industry change from the inside gives the conversation a grounding that most SEO discussions lack.
Topics include why “pure SEO” doesn’t exist, why Bing now matters to anyone who cares about LLM visibility, what the Helpful Content Update actually did and who it hurt, why AI is most dangerous in domains where you’re not already an expert, and why human creativity retains one specific advantage that AI will not replicate.
The Myth of the Pure SEO
Chris’s challenge to the label: ask any “pure SEO” to describe a typical week of work. The answer will include writing link bait articles, pitching websites, engaging on Reddit, and generating traffic from social. That’s content, PR, and social media. The pure SEO doesn’t exist — the label describes someone who hasn’t realized how many things they’re doing. The more limiting version is the practitioner who restricts themselves to a single game plan: every excluded channel is a missed opportunity.
Bing + LLMs: The Reason to Care Now
The reframe that makes Bing worth optimizing for in 2025: at least three major LLMs — including ChatGPT — pull from the Bing index. Bing is no longer just Bing. It’s the underlying index for a meaningful slice of AI-assisted search. Chris’s practical Bing playbook, built with Jeremy: register in Bing Webmaster Tools; install the IndexNow WordPress plugin (pushes new content to Bing within hours of publication); verify your Bing Business listing for local map visibility inside ChatGPT; build links from sites that Bing already favors; and lean into social signals, which Bing has officially acknowledged as a factor since 2010.
Schema for AI: What It Says, Not Whether It Validates
Chris’s observation on schema’s changing role: Google trained the industry to be technically rigid — invalid schema was ignored. LLMs work differently. They care about the semantic content of the schema, not its technical perfection. If it communicates what the page is about clearly, it gets read. FAQ schema is the best example: it gives the LLM a question and an answer in one clean unit. Practitioners who dismissed schema for years should revisit it — the audience has changed.
HCU: Google Became the Destination
Chris’s read on the Helpful Content Update: it targeted made-for-AdSense content farming and caught quality publishers in the crossfire. But the deeper issue is that Google’s incentives changed. They want to be the destination, not the directory. Featured snippets were the first piece. AI overviews are the current piece. The publishers who built the content ecosystem Google monetized for two decades are now being displaced so Google can be the endpoint. They were boiling frogs — and now they’re soup.
The AI Training Loop Problem
Chris flags a structural issue in the AI content ecosystem: if AI companies create virtual-voice podcasts with hallucinated content, and those get transcribed and summarized by other AI tools, the training data is poisoning itself. There’s no information gain — just a loop of AI generating content about things AI got wrong, which then becomes the training data for the next round. The people who provided genuine insight are being squeezed out of the economic model that would have funded their work.
Human Creativity: The Cross-Domain Connection
Chris’s specific claim for where human creativity still wins: unexpected cross-domain connections. A person who read Scientific Advertising 30 years ago and a Seth Godin book this year will make a connection that even the best AI tool won’t find — because it’s a human brain making random firing connections, possibly overcaffeinated. AI synthesizes what it was trained on. It can’t make a leap from a 1920s ad book to a 2024 brand challenge in the way a human who has lived both decades can.
RAG Database as a Content Repurposing Tool
Chris’s practical AI application: he trained a retrieval-augmented generation database on his own transcripts and writing. When he asks it a question, it gives accurate answers about his own material — because the retrieval is searching his own content, not hallucinating. He uses the outputs for LinkedIn posts, editing them to maintain his voice. His framing: it’s not plagiarism. It’s a search engine for his own accumulated knowledge — repurposing the expertise he’s already produced rather than generating new content from scratch.
Vibe Coding: Expert Use Only
Chris’s warning on AI coding tools: they work well in domains where you’re already expert, because you can spot the hallucinations. In unfamiliar domains, you can’t — and the errors compound. He cites a vibe-coded WordPress plugin that introduced a security vulnerability the developer didn’t catch because they didn’t know what to look for. AI is highly effective for experienced practitioners who can verify the output. For everyone else, it’s a way to get confidently wrong answers in a language you can’t audit.
Key Quotes
“AI tools are really helpful in domains that you’re well versed in, but don’t try to use it in a domain that you don’t know very well because you don’t spot the inaccuracies and the hallucinations.”
— Chris Garrett
“A human that read Scientific Advertising 30 years ago and reads a Seth Godin book this year is going to make a connection that even the best AI tool isn’t going to find — because it’s a human brain making random firing connections, possibly over caffeinated.”
— Chris Garrett
“Where are the people in all of this? If you look to just Google and Gemini, they create podcasts using virtual voices with hallucinated content. Then those get transcribed and summarized by AI. Where does the money come from?”
— Chris Garrett
About Chris Garrett
Chris Garrett has been in digital marketing since 1994, when he connected a UK college to the internet and taught himself TCP/IP from scratch to make it work. Over 30 years he has worked across content, SEO, email marketing, and digital strategy, and brings a long-view perspective to questions the industry tends to treat as brand new.
