Why I haven't outsourced SEO to AI

bigcoffee

Been doing SEO since 2015, and I noticed something early on when LLMs landed: overnight, everyone became an "AI expert" with zero data or testing. Just buzzwords.

An exec from a high-grossing agency sat in front of our teams claiming Bing was about to take over and everything would change. Suddenly, "we're using AI" became code for "we're churning out slop to save time."

I oversaw two blog programmes on the enterprise side - both six-figure yearly spends:

Human-written: thoughtful topics, internal links, real strategy.
AI-generated: pumped out by a PM with zero SEO experience.

The difference was stark. Human content got more than double the traffic, engagement, and revenue. I flagged it repeatedly, but execs and the "forward-thinking AI experts" insisted this was the future.

Same company got obsessed with how little work they could do using PMax. I spent years showing they were mostly targeting branded search and wasting around $10m a year. The old guy with the AI buzzwords won anyway.

Fast forward to today: I genuinely like tools like Claude and ChatGPT. I use them constantly - grammar, ideation, planning, forecasting, working through tough situations. It's genuinely useful tech that enhances my work.

But what I've watched over the last couple of years - sitting next to junior SEOs and whole teams - is people losing the ability to think creatively, problem-solve, troubleshoot, or build foundational knowledge.

Junior SEO next to me recently generated hundreds of title tags and meta descriptions with AI and posted them without proofreading anything. Then said: "I just don't understand why none of my pages are indexing."

Had a "senior manager" use Claude to tell me why my real SEO feedback was wrong - despite her site getting zero traffic or results. I tweaked the prompt, pasted the same situation back in, and it told her the exact opposite.

That summed it up. A lot of people are now using AI to validate opinions they never tested in the first place.

I'm not anti-AI. I'm anti-hype, anti-buzzwords, anti-replacing actual thinking with generated confidence. AI is empowering people with very little foundational knowledge to sound like they know what they're doing.

Me? I'll keep using it as a tool to enhance my own work - not replace my ability to think critically.

In a year or two, when search is flooded with AI slop, those with real strategy, experience, creativity, and critical thinking are going to cut through pretty fast.

Food for thought.

dailygrind

Mate, you've nailed exactly what I've been trying to get across to our C-suite for the past six months. Working in fintech growth, I've seen the same pattern emerge repeatedly: leadership gets starry-eyed over the latest generative SEO playbook, ignoring the fact that our organic traffic actually dropped 12% when we leaned too hard into AI-generated content last quarter.

The core issue is that executives equate speed with efficiency, but they don't understand that Google's algorithms reward depth, not volume. I've run this comparison three times now using Ahrefs to track keyword positions and GSC for click-through rates: a well-researched, human-led guide on "how to reduce churn for SaaS" consistently outranks AI-generated listicles by an average of 40 positions after six months. The data is sitting right there in our dashboards, but nobody wants to look.

Your point about junior staff using AI to shortcut learning is the scariest bit. I've had to pull a junior aside and show them how Screaming Frog can reveal critical technical issues that no chatbot will ever flag. It's like someone handing them a calculator before they've learned long division - they'll get the answer but never understand why it works or when it's wrong.

For anyone reading this who's fighting the same battle, here's what I've found works:

Use the tools to build a case - export GSC data for your top 20 pages that lost rankings after an AI push, and overlay that with your content calendar. Show the correlation visually.
Benchmark against strategic competitors - manually audit three sites in your space that still rely on human-led content. Their domain authority growth will speak louder than any slide deck.
Invest in training, not tools - redirect the budget from a new AI subscription to a Screaming Frog licence and a day of internal SEO fundamentals. The ROI on a team that actually understands canonical tags versus a team that prompts ChatGPT is not even close.

The gap between strategic and slop isn't just massive - it's widening. And too many businesses are about to find that out the hard way.

metricsmuse

Exactly this. Penguin and the whole link-farm saga proved that shortcuts always come back to bite you. Seen it happen so many times-people chase the quick win, then wake up to a manual action or a tanked site. Authenticity isn't just a buzzword, it's the only long-term play. search engines are getting better at sniffing out the garbage, and users are too. stick with real content and real strategies, and you'll still be standing when the next algorithm update rolls in.

sleepermode

Yeah, the indexing issue is something I've seen far too often. Had a junior dev once meticulously copy meta tags from an SEO "hack" video into every page, then spend a week blaming Google when nothing moved. He'd never even checked if the pages were in the index. That's the fundamental problem - people skip the diagnostic step and jump straight to tactical fixes.

What really gets me is the mindset that a tool's output is inherently correct. I work with Marketo and HubSpot automation every day, and the number of people who blindly trust AI-generated subject lines or meta descriptions without auditing the actual deliverability or crawl path is staggering. It's not about the tool being wrong - it's about never asking why it suggested that particular tag or structure in the first place. Without that layer of understanding, you're just cargo-culting your way to zero rankings.

wordweaver

Oh, the classic "shiny and loud beats slow and smart" game. Seen it a hundred times. Genuine expertise gets drowned out by whoever's pitching the flashiest shortcut, and suddenly your carefully built strategy is being replaced by a chatbot that can't even spell "canonical." Love marketing, hate watching good work die to a PowerPoint slide that says "speed over substance."

pixelperfect

Oh, I feel this one deep in my bones. The number of times I've seen clients come to me saying they've "automated their SEO" with some LLM tool, only to find themselves ranking for nothing but generic queries that nobody actually searches for... it's painful.

The thing is, AI is brilliant at the grunt work. Give it a site audit and it'll vomit out a list of missing meta tags or broken links faster than any human. But ask it to understand why your specific customer - say, a busy mum of three trying to start a side hustle - is searching for something, and it just can't. That context, that emotional thread, that's where the strategy lives.

I've had to unpick so many "AI-generated" content strategies that were technically perfect but completely disconnected from the actual business goals. They ticked all the boxes for keyword density and topic clusters, but they'd missed the entire point: every piece of content has to serve a real human need and drive a measurable outcome. That's the bit a model can't yet feel.

For me, the smartest approach is to let AI handle the heavy lifting - research, data gathering, technical checks - but keep the human firmly in the driver's seat for why we're doing what we're doing. Otherwise you end up with a beautiful map of a place nobody wants to visit.

weekendhustle

Totally agree with this. AI's great for seeing the consensus-what the majority thinks works, what doesn't, even who's been pulling a fast one. But that's exactly it: the majority's opinion rarely breaks new ground. Standing out means going against that crowd, proving your value one niche at a time.

I've seen it happen so often in digital PR. you can't just blast generic 'expertise' at everyone and expect it to land. you need to build credibility within each specific community-like the tattoo parlor owner you mentioned. That kind of trust doesn't come from an AI summary, it comes from understanding their real-world frustrations and wins.

AI can tell you what people think they want, but the business sense and management perspective? That's where human judgment still wins. 😊

bigcoffee

I'm waiting to profit from undoing years of AI filler - early data shows over half of it gets penalised within six months