Supercharging SEO with AI: My Journey and 5 Actionable Strategies

The short version: Three years ago at TubeBuddy, we went from one blog post a week to daily publication and grew traffic 500%. The playbook was a hybrid one. AI for speed and data, humans for voice and judgment. Here are the five strategies that worked, updated for 2026 where AEO and GEO are now part of the job.

Let me say this plainly. I used to think SEO was mostly a willpower game. Pick the keywords. Write the posts. Publish consistently. Wait.

That worked in 2019. It stopped working around the time content production costs collapsed and AI engines like ChatGPT, Perplexity, and Google AI Overviews started chewing up a real slice of search traffic.

Here’s the thing. The fundamentals of SEO did not change. What changed is the economics of doing them well, and the surface area you have to cover. In 2026 you are optimizing for Google, for AI answer engines, and for the humans who still click. The same playbook has to serve all three.

This is the playbook I ran at TubeBuddy as CMO, refined during growth work at Bloomberg DTC and MyRecipes (People Inc), and now run across client engagements at The Remarkable Agency.

What Did I Actually Learn at TubeBuddy?

Three years ago, as CMO of TubeBuddy, a tool helping YouTubers optimize their content, I was staring at a growth plateau. We were publishing one blog post per week. We had 300,000 monthly visitors and the trajectory had flattened. The challenge was blunt. Scale content dramatically, hold quality, stay in budget, keep technical accuracy for a specialized creator audience.

What I didn’t know until recently was how much of the job was pure data work that a good AI workflow could eat. Keyword clustering. Outline generation. Meta descriptions. Alt text. Schema hints. Internal link suggestions. None of that was where the strategic value lived. All of it was where the hours went.

We rebuilt the engine around a hybrid methodology. AI handled the data-heavy starting point. Human writers owned voice, nuance, and original insight. SEO tools optimized against top-ranking competitors. Human review gated every publish.

Traffic grew 500% from a 300,000 monthly baseline. Trial signups grew 30%. Revenue grew 10%. Performance held for two more years. That outcome is what convinced me the hybrid model was not a trick. It was the new default.

How Should You Rethink Keyword Research With AI?

Use AI to collapse the research phase from days to hours, not to outsource the strategy. A 40 to 60 word direct answer looks like this. Feed your seed terms into Semrush or Ahrefs to pull search volume and difficulty. Then feed the top-ranking pages into ChatGPT or Claude and ask it to cluster queries by intent. You get a map in an afternoon that used to take a week.

Here is what the workflow actually looks like in practice.

Start with a seed list of 20 to 30 terms from a sales call transcript or a support ticket export. Real customer language beats a keyword tool every time. Run them through Semrush or Ahrefs to pull volume, difficulty, and SERP features. Export the top 10 ranking URLs per query.

Then prompt Claude or ChatGPT with something concrete:

“Here are 30 queries my customers use and the top 10 ranking URLs for each. Cluster them by search intent (informational, commercial, navigational). For each cluster, suggest a content format (pillar post, comparison, how-to, data study). Flag any clusters where AI Overviews are showing up and note what kind of answer they reward.”

That last part matters in 2026. If an AI Overview is showing for your target keyword, you need to write for extraction, not for a ten-blue-links world. Research from Search Engine Journal shows AI Overviews can cut click-through rates by 35 to 65% on informational queries. That is not a reason to stop. It is a reason to restructure how you answer.

Before: three days of manual keyword pulling and competitor tab hopping. After: one afternoon of structured AI-assisted clustering with a clear content plan tied to real search intent.

Which Routine SEO Tasks Should You Automate First?

Automate the tasks that consume the most time and produce the least strategic value. Meta descriptions, alt text, internal link suggestions, and technical audits. A 40 to 60 word direct answer: these four categories eat 10 to 15 hours a week on a mid-sized content team and almost never require strategic judgment. AI handles them well with light human review. You redirect the saved hours to strategy and original reporting, which is where ranking actually comes from.

Here is the comparison that changed how I run content teams.

Traditional SEO Tactic	AI-Enhanced Approach	Time Savings
Writing meta descriptions one by one	Bulk-generate with ChatGPT using a brand voice prompt, human spot-checks top 20% of pages	From 8 hours to 45 minutes per 100 pages
Manually writing alt text for every image	Claude or GPT-4V describes images in bulk with brand and accessibility guidelines	From 5 hours to 30 minutes per 200 images
Internal link mapping by spreadsheet	AI crawls the sitemap and suggests contextual links per new post	From 3 hours to 15 minutes per post
Quarterly technical SEO audit	Continuous audits via Semrush or Screaming Frog with AI-prioritized fix lists	From 20 hours per quarter to rolling 2 hours per week
Keyword research from scratch per brief	AI-clustered opportunity maps refreshed monthly, briefs pull from the map	From 4 hours per brief to 30 minutes

That table is the whole shift. Every row is a task that used to be gated by someone’s bandwidth and is now gated by the quality of your prompts and review process.

Tools worth naming: SurferSEO and Clearscope for content optimization. Frase for brief generation. Screaming Frog for technical crawls. ContentKing and SEOmonitor for real-time monitoring. Any of these combined with a good LLM will collapse the routine work.

How Do You Blend AI Drafts With Human Voice Without Sounding Generic?

Write the brief like a senior editor, not a search engine. A 40 to 60 word direct answer: most teams get this wrong by treating AI as a vending machine. They type a sentence, accept the draft, publish it, and wonder why it reads like every other page on the internet. The fix is to brief AI the way you would brief a senior staff writer. Concept first. Voice second. Outline third. Draft last.

Step one is the concept. What is the single insight this piece delivers that the top ten ranking pages do not? If you cannot answer that in one sentence, do not write the piece yet.

Step two is the voice. I keep a brand voice document per client with 200-word samples, banned phrases, and structural preferences. I paste that into every prompt. Without it, AI defaults to the flat, hedging, faintly American-corporate voice that every reader now recognizes as machine output.

Step three is the outline. I ask Claude or ChatGPT to propose an H2 structure based on the top-ranking pages plus the concept from step one. Then I rewrite the outline myself. This is where strategy lives.

Step four is the draft. AI writes section by section, fed the outline and the voice doc. I edit in place. I kill filler sentences, add specific examples from real engagements, and swap every vague claim for a statistic.

Before and after example. AI draft: “Content marketing is important for brand visibility and can help businesses grow.” Human-refined: “At TubeBuddy we published daily for two years and grew traffic 500%. The driver was not volume. It was that every piece answered a specific creator question no one else was answering clearly.”

The second version cannot be written without a human who lived the story. That is the part AI cannot replace.

How Should You Optimize Content for AI Answer Engines in 2026?

Structure every section for extraction. A 40 to 60 word direct answer: AI engines like ChatGPT, Perplexity, and Google AI Overviews do not read your page top to bottom. They scan for passages they can pull into an answer. Question-style H2s, 40 to 60 word direct answers, HTML tables, and named sources get extracted. Flowing narrative prose does not. If you want AI citations, you write for how AI reads.

This is where SEO in 2026 diverges from SEO in 2020. The old job was rank on page one. The new job is rank on page one and get cited by the AI engines sitting above page one.

The formats that get cited most, according to a 2025 extraction study from Princeton and Georgia Tech:

• Tables get cited 34% of the time against plain paragraphs at 3%
• Question-style H2s with short direct answers get cited 29% of the time
• Structured lists get cited 21% of the time
• Content with real statistics and named source citations gets cited 30 to 40% more often
• Content with a named author and visible credentials gets cited 41% more often

This is not guesswork. It is a measurable extraction bias. Research from SparkToro and the team at Ahrefs both back the pattern. AI answer engines reward structure and specificity. They punish vagueness.

We wrote a full teardown of the five factors that drive AI citations and how to build a tracker for your own site. If you want the deep version, read our post on building an AEO/GEO visibility tracker. That post is the operating manual for this strategy.

The practical change is small. Every new page gets a question-style H2. Every H2 opens with a 40 to 60 word answer. Every comparison becomes a table. Every statistic gets a source link. Every author byline links to a real person with real credentials. Do that consistently and AI citation rates climb.

Why Do Real-Time Analytics Matter More Than Monthly Reports?

Because SEO in 2026 moves faster than monthly. A 40 to 60 word direct answer: a Google algorithm update, an AI Overview rollout, or a competitor shipping a better page can shift your rankings in days. Monthly reports tell you what happened six weeks ago. Real-time analytics tell you what is happening now, which is the only window where you can still intervene.

The stack I run in 2026:

• ContentKing for real-time change detection on your own site and key competitors
• SEOmonitor for rank tracking with forecasting and AI Overview presence flags
• Google Analytics 4 with custom dashboards for behavior pattern analysis tied to content clusters
• A weekly AI scan using a tool like our own AI Discovery Audit to track which pages are getting cited by AI engines and which are invisible

The behavior pattern piece is the one most teams skip. Rankings tell you if you are findable. Behavior data tells you if the page is doing its job once found. If your bounce rate on a ranking page is 85% and your scroll depth is 20%, Google will notice. AI engines will notice. You have a content problem, not a ranking problem.

Set alerts. Review weekly. Act inside the window where action still moves the needle.

Where Does AEO and GEO Fit Into Modern SEO?

Treat AEO and GEO as part of SEO, not separate disciplines. A 40 to 60 word direct answer: Answer Engine Optimization (AEO) and Generative Engine Optimization (GEO) are the practices of getting your content extracted and cited by AI engines. They are not replacing traditional SEO. They are extending it. The same structural, authority, and freshness signals that win in AI engines also help you rank in Google. Do both at once.

Here is what I tell clients. If you are starting SEO in 2026, you cannot build a program that ignores AI engines. Google’s own guidance is that helpful, reliable, people-first content wins regardless of whether it was written with AI. That same guidance is what AI engines use to decide who to cite.

The overlap is larger than most people think. Well-structured pages with named authors, cited sources, fresh updates, and real statistics win both in traditional search and in AI citations. The content rules I apply to every page on theremarkableagency.com come from this overlap. Question-style H2s, direct answers, HTML tables, source citations, named authors. Every one of those signals is doing double duty.

If you want to go deeper on the AI citation side specifically, our tracker post breaks down the five factors that account for about 80% of whether AI engines cite you.

What’s the Actual Bottom Line?

AI did not replace SEO. It collapsed the cost of doing SEO well. That is the thesis, and it is still the thesis two years after I first ran this playbook at TubeBuddy.

The teams winning in 2026 are using AI for the data-heavy parts of the job (keyword clustering, draft production, meta and alt text, technical audits) and spending the saved time on the parts that still require a human: original insight, brand voice, editorial judgment, and the strategic call on what to publish next. They are also writing every page for both Google and AI engines from the jump, because the overlap is real and the rewards compound.

Start with one workflow. Pick the task that eats the most of your team’s week. Replace it with an AI-assisted version. Measure the hours saved and the output quality. Then pick the next one. That is how TubeBuddy went from one post a week to daily. It is how we run the SEO side of engagements at The Remarkable Agency today.

If you are rethinking your SEO program for 2026 and want a second set of eyes, send me a note at alex@theremarkableagency.com. Happy to walk through it.