Stop Buying Blog Posts by the Month – Build the Knowledge Base AI Will Read, With Olesia Korobka

How many blog posts a month can you deliver? If you’ve ever sold SEO to a large company, you’ve heard that question before you’ve heard anything about their customers. Olesia Korobka hears it too. When I interviewed her on the Unscripted SEO Interview Podcast in February 2025, she was still being approached by big companies who wanted “lots of blog posts about very generic topics with their corporate blogs, which has not been ranking for five years and will not be ranking for five more, but they still somehow are used to spending money on that.”

Olesia is an SEO entrepreneur with fractional roles in SaaS companies, an agency, and a newer venture she was pouring all her time into. She’s also the founder of WhiteLobby, which does AI visibility work for competitive industries like iGaming. Her alternative to the blog retainer is a knowledge base: one place that holds what your company actually knows, which you can then turn into a website, social posts, a podcast or whatever the AI engines want to read. If you’ve been paying for content that never ranks, this one’s for you.

The retainer that counts blog posts

Here’s how Olesia describes those engagements: “It’s always in the retainer. And the only thing that is interesting for them is how many blog posts a month can you deliver.” Nobody offers a subject matter expert to review the posts. And in 2025, people were “still looking at the blog posts as like traffic producing machine. I don’t know why.”

I’ve watched the same thing from the inside. It took years to convince large organizations that they needed a lot of content, and somehow that got translated into we need a lot of blog posts. The blog is the only part of the site anyone touches. Nobody says let’s build a case study section, a master class, a tutorial series. It’s always “what’s our blog post topic, and can you optimize it after it’s done?” (I wrote about the economics of that in four blogs, two links.)

If you’re the one signing the retainer, ask yourself a simple question. Which of last year’s posts brought you a customer? If you can’t name one, you’re paying for volume, not for knowledge.

Your support docs are content too

Here’s a place most companies forget to look. I’ve set up support knowledge bases for several organizations, and every time the company fights to hide them. Support can’t be front and center, because that implies the tool is hard to use. So it goes in the footer, then a support page, then a knowledge base page, then a search box. That’s three or four hops before your customer can even type the question.

Meanwhile, people are Googling “how do I do this with” every piece of software they own. Your support docs are the answer to those searches, and you’ve buried them.

Olesia had a darker version of this story. One tool owner told her he sometimes answers support tickets himself, and he noticed people cursing at the support system right from the first message. Why? Because you “need to incorporate some of these words to scare off the AI,” and then the bot hands you to a human faster. Your customers have learned to swear at your chatbot to get help. That’s a content problem as much as a support problem.

Buckets of information: build the knowledge base first

This was the part of the conversation I kept thinking about afterward. Olesia doesn’t think the web needs as many websites as it has:

“I’m not sure if we have that many websites in the future. I’m not sure if we need them at all.”

Her reasoning is that a lot of informational content is generic and duplicated, and if people move to chatbots and personal assistants, “they will not be visiting those sites ever again.” So she helps companies get ready for that:

“I want the companies to have their buckets of information, valuable information, specific to their business and specific to their business goals.”

Once you have that bucket, the output format stops mattering. You “can use this knowledge base to generate website, social profile, even blog posts, if you like, podcast, a video, anything, a tutorial, whatever else.” The blog post becomes one possible output rather than the whole program.

Here’s what a knowledge base for AI search can hold for your business:

• What you do, for whom, in specific words (the about page you should already have)
• How your product or service actually works, step by step
• The questions customers ask support, with the real answers
• What your people know that competitors don’t: interviews, talks, case notes
• The facts about your company that third-party sites should be repeating correctly

A web where you don’t need a website

Olesia took it further than I expected. Language models could pull straight from your knowledge base through an API, so “you don’t need to create the website and all these interfaces.” You’d have “this pod or I don’t know, a bucket with your information and various systems just connect to it to generate the content from it.”

My first reaction was that it’s mind blowing — and that it runs straight into money. Google built the deal we all live with: let them show our stuff, sell ads around it, and send us the click eventually. Zero-click search already took over travel, tickets and hotels. I’d love for an alternative to work. Olesia’s answer was more realistic than mine. Only big tech has the money to build that alternative, which is why they’re in such a hurry, because “who controls this access to private information will win actually in the long run.”

You don’t have to believe websites are going away to act on this. A clean knowledge base helps your site today and is ready for whatever replaces it.

What I’m doing with 100+ interviews

I’m building the same thing for myself. I’ve recorded a lot of podcast interviews, and I’m pulling them into a single database I can query: give me seven quotes from different people about citations, then use them in an article. That kind of pooled knowledge wasn’t available to a small business five years ago. Olesia is doing the same with her videos, using tools like NotebookLM and LangChain to connect everything into one base.

I also pitched her an idea. Olesia helps run SEO for Paws, the online charity SEO conference started by Anton Shulke that raises money for small shelters caring for cats and dogs hurt by the war in Ukraine. Their third conference ran on 25 March 2025. I learned SEO from Aaron Wall’s SEO Book, and Aleyda Solis built learningseo.io into one of the best-structured resources in the industry. Why not turn the conference talks into a book, with the sales going to the shelters? Olesia said that’s what she planned to do once all three conferences were done. That’s the knowledge base idea again: the talks already exist, they just need a bucket.

Knowledge base for AI search: where to start this quarter

You don’t need a new platform to start. You need answers to a few questions:

1. What does your company know that nobody else has written down?
2. Where does that knowledge live right now: in people’s heads, in support tickets, in sales calls, in old webinars?
3. How many hops does it take a customer to find your support docs?
4. Does your about page say anything a competitor couldn’t copy word for word?
5. If you cancelled the blog retainer tomorrow, what would you build with the budget?

Olesia’s advice on how to find her is a good test of whether this works: “In Google, whatever pops up there for you, just use it. I have optimized myself a bit, not too much, but a bit.” You can see what she’s built at olesiakorobka.com and WhiteLobby. Listen to the full episode on Castos, read the episode recap on Unscripted SEO, and if you run an agency, the SEO Arcade piece on brand mentions for AI visibility covers the off-site half of her playbook. For more on why your reputation lives on other people’s sites, see authority is what others say about you.