A Guide to Writing Blogs That Actually Rank on Google, and Google AI Overview
What kind of content wins in the age of AI search?
First, know the goal of making content: grow your site.
You’re not making content (purely) for self-expression. You’re not making content to praise your product (even though it’s awesome and we all wanna hype it up.)
You’re making content to provide value to the people you want landing on this page.
This is mostly for blogs, but the traits apply across page types too, like landing pages, resource articles, and comparison pages.
1. The never-changing goal: quality
You recognize the importance of building pages for your site. You think, let’s spin up a ton with AI.
Remember this hard truth: engines are not stupid. As your game improves, their game is constantly evolving too. But there’s one thing that doesn’t change: Quality wins.
Quality >> quantity.
1 actually valuable blog a week >> 100 AI slop posts.
The goal of search engines is to surface the most valuable information for their users. So they’ll keep evolving their mechanisms to optimize for that goal. Which means if you manage to provide real value to someone, you’ll always have an advantage. It’s so simple. And honestly, it’s great, because it makes the world a better place.
2. Human
This is directly related to the first point: quality. The AI age is great. But unfortunately, what current LLMs spit out just isn’t useful. To provide actual value, you need a human in the loop.
I’m not banning the use of AI. But you cannot remove the human part. Where the human sits is different for everyone, and that depends on how you wanna design your playbook. A few examples:
- Human: do a quick, unpolished writeup of the whole piece of content AI: clean it up and improve the flow Human: pass it again to remove hallucinations and add more valuable info
- Human: feed AI a good title AI: draft the structure from the title Human: clean it up and make it actually useful
The first and last step should always be human.
Often, multiple rounds happen. But you need a human to come up with an actually useful idea, and you need a human to do the final cleanup run.
3. Expert
Engines favor expert POVs because they are deemed more authoritative and trustworthy. They don’t just care about what is being said. They also care about who’s saying it.
When you post a blog, you need to list the author and the author’s title. But the author’s role and the blog content need to make sense together. So you actually need to know who the author is, then write the blog from their POV.
You can create multiple personas and write from each persona’s POV. But each piece of content needs to sound like it came from that specific person.
If your title is:
“Best options to manage API keys in 2026”
You want the author to be a CSO, cybersecurity researcher, CTO, or senior engineer at XXX.
Natural hints start showing when you write from a real author POV, like:
- “I audited 5 tools thoroughly before onboarding my company to this one. Here’s why.”
- “We once had an API key leak that caused something bad because we didn’t know XXX was important. This tool solves that.”
Things like this make the content feel real, authentic, and authoritative.
Occasionally some content types let you act as a news reporter, to be as “irrelevant” and objective as possible. That’s something else you can play with. But whoever you are, lean into the role.
4. Objective
This follows the expert point. To be an expert, you oughta be as objective and honest as possible.
You won’t call any tool “industry-leading” or “number one.” People generally don’t care if you yelled “we’re the best.”
You’ll say: “This tool has these pros and these cons.”
I get that you want to praise your shit. I do too. But let the real voices do that. Keep the goal of content in mind: grow your site.
A level up from being objective is sounding really objective while subtly praising your thing:
“This tool has these pros and these cons. The cons are probably fine if you don’t need these specific things. The pros are especially strong if you’re looking for this use case.”
5. Research
Some level of research is always required. Same as writing any school paper. You need to show you’re not just rambling from inside your own world. Quote a few things. Add some external links. Keep it grounded. You can keep this light for most topics, but don’t skip it entirely.
Get AI to help with this.
6. Audience
Know who you’re talking to, so you can keep saying something valuable to them.
If the title is: “Best banking options for small LLCs”
Your audience probably cares about financing options, tax perks, ease of managing company and personal finances, fees, operational simplicity. So write about those things.
How it all compounds
All of these work together to do 2 things.
-
Let engines believe you’re for real.
After parsing your content, engines make an initial judgement you’re not AI slop, you’re trustworthy, and actually providing value for a certain group of people.
-
Let humans believe you’re for real.
Engines start surfacing you. Some humans click. They read longer because your writing actually makes sense. They click around a bit more. They may even view another page on your cite because some CTA appealed to them.
And the human reaction feeds back into engines. The compounding’s done.

Helena is a growth engineering expert with over 10 years of experience scaling AI products across product, engineering, SEO, content, and distribution. She has built 4 AI startups, including OpenArt from 0 to $60M ARR in a year and another to $10M ARR in 8 months. She studied Electrical and Computer Engineering at UCLA and built a 300K+ subscriber YouTube channel covering AI tools, creative AI, and growth workflows.
More from Helena →

