Scenario: The Rise of AIO and the Snake Oil Salesmen (2026.2)

Every year, I publish a set of scenarios to anticipate the biggest shifts in tech and business for the coming year—a tradition I started as Tech Director at Elisa¹. These posts aren’t just predictions, but frameworks for thinking about the disruptive changes ahead. If you’re new here, this series is my way of mapping out what might matter most in the next twelve months. Here’s the second scenario for this year².

I’m writing this post in Grammarly’s new Coda editor. It helps me craft language that sounds more natural than the text I’d normally produce in English—a second language I’ve historically reserved for professional, jargon-heavy contexts.

While Grammarly existed a bit before Large Language Models, it is a prime example of how those models and their predecessors³ have changed communication. Everyone now has a personal editor⁴ in their pocket. If you want to post a tweet, write a press release, or draft a manifesto, an AI can make it “good” in seconds. This means “good” doesn’t cut it anymore. “Good” is the baseline. To be successful, you now have to be exceptional because the barrier to entry for competence has effectively vanished.

That’s just context, not the real prediction.

The second scenario for this year is about the concequences of new industry forming beyond traditional SEO, PR, and content management. As AI fills the web with polished content, the critical shift is from optimizing for Search Engines to optimizing for Stochastic Prominence—ensuring your brand stands out in the unpredictable outputs of AI models.

The Jaconia Anomaly

I realized this shift wasn’t theoretical when we ran a small experiment with my own side business, Myrrys⁵.

Myrrys is a niche business. It’s a labor of love in a small genre within a small country⁶. But when we asked various LLMs—ChatGPT, Gemini, Claude—to analyze the Finnish RPG landscape, they all reliably linked the three most prominent figures in Finnish RPG history, and claimed that the flagship of Myrrys is Praedor by Mike Pohjola.

Here is the problem: None of those things is related to the other.⁷

1. Myrrys is the publisher we run.
2. Praedor (set in the world of Jaconia) is a major Finnish RPG, but we do not publish it.
3. Mike Pohjola is a prominent Finnish RPG author, but he has never contributed to Praedor, nor do we publish him.

To the LLM, these were just three high-weight nodes in the vector space of “Finnish RPGs.” It saw “Prominent Publisher,” “Prominent Game,” and “Prominent Author” and simply synthesized them into a single, cohesive (and completely false) narrative. It created a “Super Fact”⁸ because it lacked the structured data to understand the actual relationships.

This is the inverse of the “Invisible Brand” problem. If an LLM cannot “read” your authority via clear, structured data, it doesn’t just ignore you—it might rewrite your reality to fit a more probable pattern. This glitch is not a random edge case; it is a preview of the new market reality.

The Scenario

By 2026, a key scenario is the rise of Answer Engine Optimization (AIO) —strategies for having your content selected, synthesized, and cited as a direct answer by AI systems, replacing traditional click-focused strategies.

This will happen in two ways, one annoying and one critical.

1. The Snake Oil

We have already seen the deluge of agencies claiming they can “hack” the LLMs. The rise in “AI Visibility Experts” selling magical thinking—promising to insert your brand into ChatGPT’s latent space through prompt engineering tricks or “secret sauce” keywords. Most of this will be nonsense. No one fully understands the black box of these models, and anyone claiming they can guarantee you a citation in a GPT-5 response is lying.

And while there are people who genuinely seem to understad hacking the probability space⁹, those methods are merely a hint of the real change.

2. The Return of “Real” PR

The actual, legitimate shift will be a return to very traditional Public Relations, but with a twist.

In the world of Google Search, you want backlinks. In the world of transformers, you want Training Data.

LLMs rely heavily on “high-quality” sources to ground their answers—think The New York Times, Reddit, Wikipedia, and established niche forums. If you want your brand recommended by an AI, you need to be discussed in places the AI trusts.

This means “Digital PR”—getting mentions in reputable newspapers, reputable sites, and yes, even ensuring your Reddit reputation is solid—is no longer just about brand awareness. It is about existence. We are moving toward a world where marketing is less about “influencing people” directly and more about “influencing the training data” that the people consult.

The Playbook for 2026

Acting on the scenario requires recognizing that “optimizing for clicks” is fading, and “optimizing for truth” is emerging. Success will come from ensuring your brand is the answer AI systems recite as “truth.”

Now, I am not an expert in PR, marketing, or training data poisoning, but I have spoken to a handful of people who are¹⁰. Here are some ideas from those discussions on how to actually do this:

Elevate Your Experts: The Jaconia anomaly happened because the AI recognized a prominent person and desperately wanted to connect him to something. Use this. Faceless corporate brands are hard for these models to “connect” with, but they love quoting recognized experts. Encourage your team to publish and speak. If the machine trusts your people, it trusts your company¹¹.

PR as “Data Injection”: Stop measuring PR by referral traffic. Measure it by Semantic Association. Get mentioned in the training grounds: Wikipedia, high-authority media, and niche communities like Reddit subreddits.

Own Your Narrative: Your own website is losing relevance as the primary source of truth. To an AI, reality is defined by what established players—reputable news outlets, industry journals, and curated knowledge bases—publish about you. Shift your focus to traditional PR that feeds these trusted sources with accurate, structured stories. If a major news outlet or a verified knowledge base prints it, the algorithm accepts it as fact.

Support the Community: You don’t always have to be the direct source of content to stay relevant. By actively supporting professional communities and industry events—even as a sponsor, organizer, or participant—you position your brand at the heart of conversations that matter. The organic discussions and connections you nurture become the very training data that AI systems use to recognize your influence and cement your brand’s place in the ecosystem.

Feed the Commons: Today’s web is increasingly a “dark forest”—valuable insights locked away behind paywalls or proprietary barriers. Resist this trend. Proactively release your industry research, white papers, and datasets under Creative Commons or similarly open licenses. In a world hungry for reliable data, AI models—and by extension, the market—will gravitate toward transparent, freely available information. If your data becomes the backbone of authoritative answers, you will shape the narrative, set the standard, and own the truth.

Don’t just feed the machine. Be the truth it wants to tell.