Not all that long ago, I found myself trying to find an answer to a question about the PSU I might need for a PC accessory. I was alarmed to see that my site wasn’t in the results for the query, because that information wasn’t featured in a review I’d done about a year before.

A hard, but important lesson: Be useful, complete in your writing and have empathy for your user’s needs, or disappear.

AI isn’t a threat to good content.

As a content marketer, I look at the past year through slightly jaded eyes, “AI slop” (which is a term I don’t particularly enjoy using), felt like it was overwhelming every marketing channel we work in.

Prolific as it is, using AI to write content is not a sustainable practice. It’s a bad idea.

I’ve done the tests; AI generated content works in search for long enough to give you a false sense that everything is working well, but, if you don’t stay on top of the output or you’re lazy, your rankings will go south really quickly. In any case people instinctively know that they’re reading AI content.

You lose your brand tone of voice and start sounding like a LinkedIn post.

Creative debt.

The creative debt of having a large site littered with AI written copy is problematic too; at some point you’re going to need to rewrite it or constantly keep a close eye on what’s performing instead of focusing on growth.

Modelling natural language is tough (check out my Voice Analyser open source repo here if you’re interested).

The problem is that making writing truly feel human can’t be replicated. Your sentence rhythm, vocabulary patterns, argument flow, narrative, those micro-rhythms we weave into prose are unique and unpredictable. AI uses language in a mathematical, formulaic way – it is by its nature, measurable.

Your content people are your best asset, so in 2026 I’m focused on working with my clients to help them get more productive and streamlined.

I’m obsessed with this idea:

AI enables human productivity, it’s not a replacement for human creativity.

I own my own content sites for experimentation and revenue (of course!) so I think a lot about moving my workflows along faster, without compromising on quality. I really enjoy working with AI for content research, evaluating strengths and weaknesses of competing content, updating old content with new data, and so on.

This is where Gemini MCP is so powerful in my content workflow.

Imagine this scenario: You’ve nailed down the idea and you’re doing the research. The content research process can take a long time. This is where I might use Claude Desktop with Gemini. You can connect all manner of MCP servers to your AI assistant and get access to data or grounded insight in seconds. If you’re a large organisation, you’ll have a measure on where the time is being spent in your creative pipeline

The other focal area for us all in 2026 is of course, AI search visibility. I really enjoyed Metahan’s research on Harmonic Centrality as an authority signal in search and I recommend you read his article.

Information rich content that covers specific recommendations, nails down the key data points and, can present demonstrable first hand experience with unique observations is so important. A lot of writers (myself included) are a little guilty of vagueness in their prose. What I mean is: “pretty good”, “and so on”, “they’re reasonable” – this is not useful to AI or humans, and I think that’s a powerful takeaway.

All of us should be aware of when, and when not to use AI in our work.