What AI Makes Possible

A lot of the conversation about AI in marketing has focused on scale, especially when it comes to content. How can we create more articles, social posts, landing pages, or other assets without adding more people or hours? While there are certainly efficiencies there, producing more isn't what I find most interesting.

I've written before about using AI to scale content without sacrificing quality. The goal isn't to create 100 mediocre articles in the time it used to take to create 10 good ones. It's to rethink what's possible when parts of the process that used to require hours of manual work can happen faster. The more I've experimented with AI workflows and automation, the more I've realized that same idea applies well beyond content creation.

From Manual Work to Faster Analysis

I recently experimented with connecting Claude to Screaming Frog while looking for opportunities on a client's website. I started with a fresh Screaming Frog crawl and worked through the technical findings. Instead of simply working down a list of flagged issues, I was able to quickly separate meaningful problems from things that didn't require action.

For example, dozens of title tags and meta descriptions were over recommended lengths. Looking at those individually could have turned into more than 100 separate optimizations. Looking at the pattern instead showed that many of them shared the same underlying template issue. The better recommendation wasn't to edit every page manually. It was to fix the source of the problem.

From there, I wanted to understand a recent decline in organic traffic. My first data source couldn't provide the historical page-level information I needed because of account limitations, so I moved to Google Search Console and continued the analysis there. That ability to quickly change direction without rebuilding the entire analysis from scratch is where the workflow started becoming particularly useful.

Getting to Better Questions Faster

The initial question was simple: Which pages have lost the most organic traffic? A comparison of Search Console data gave me that answer, but it also surfaced something even better. Many of the pages losing clicks had improved in average position while simultaneously losing a significant number of impressions.

If rankings are stable or improving, a traditional rankings-focused optimization may not address the actual problem. The decline could instead be related to lower search demand, changes in how often queries trigger the site, changes to the SERP, or evolving search behavior (hello, zero-click world). So rather than continuing to analyze clicks, I shifted the analysis toward impressions.

That change surfaced issues the original analysis had buried. One declining URL was now redirecting across multiple variants, while another page had disappeared from the site entirely. What started as a traffic-loss report became a much more useful investigation into where visibility was declining, why it might be happening, and which losses actually warranted action.

AI Doesn't Replace SEO Expertise

None of these tools make strategic decisions on their own. They can process a crawl, compare datasets, organize information, and surface patterns much faster than I could manually, but someone still has to know what questions to ask and how to interpret the answers.

A list of 128 long title tags is just a list until someone recognizes the template pattern behind it. A table showing lost clicks doesn't tell you that rankings have actually improved unless you know to compare those metrics. And a rankings report can easily send you in the wrong direction if the bigger issue is disappearing impressions.

There are also still gaps in the data. In this case, the analysis couldn't tell me how the brand's visibility had changed in AI Overviews or LLM responses. Recognizing what the available data doesn't tell you is just as important as analyzing what it does.

Making Previously Impractical Work Practical

This is where I think the real opportunity with AI and automation gets interesting. There are analyses that SEOs have always been able to do, but doing them consistently across every account hasn't always been practical. Exporting multiple datasets, cleaning them, combining sources, comparing periods, checking URLs, categorizing changes, and packaging the findings can take hours before you even get to the strategic analysis.

When much of that work can be automated, the economics change. A detailed crawl analysis could become a regular process that focuses attention only on meaningful changes. Content decay could be identified before a page experiences a major traffic loss. Internal linking opportunities, redirect issues, metadata patterns, competitor changes, and declining visibility could become ongoing signals rather than projects someone remembers to investigate once or twice a year.

The result isn't necessarily more deliverables. It's the ability to do deeper work more consistently and spend more time on the parts that require human expertise.

The Next Phase of AI in Marketing

We've spent a lot of time asking what AI can create. I'm becoming much more interested in what AI makes possible.

For marketers, that means looking beyond content generation and asking which parts of our work have historically been limited by time rather than value. There are countless analyses, quality checks, research projects, and monitoring processes that would be useful to perform more often if they didn't require so much manual effort.

That's where I see the bigger opportunity. AI can reduce the tedious work between a question and an answer, but the value still comes from knowing which questions are worth asking and what to do with the answers.