The invisible AI search campaign
AI search removed the window from window shopping. Political campaigns noticed.
👋 Hey, I’m George Chasiotis. Welcome to GrowthWaves, your weekly dose of B2B growth insights—featuring powerful case studies, emerging trends, and unconventional strategies you won’t find anywhere else.
This note is brought to you by Minuttia.
According to our survey of 599 marketers:
Over 80% of organizations are actively addressing AI search.
And with over half of respondents expecting their AEO budget to rise in the next 12 months, if you’re not investing in AI search, you risk falling behind.
That’s where Minuttia comes in.
Minuttia helps B2B companies dominate search through a suite of human-first, high-touch, value-driven services:
Google and AI Search Strategy
Content Audits
Human Content Creation
Digital PR
If you want to win in AI search before it becomes table stakes, request an intro call with us.
We’ve already helped 50+ B2B SaaS companies. You could be next.
I had a conversation recently that I have not been able to stop thinking about.
A company I’m associated with in the AI search tracking and optimization space told me something I was not ready to hear.
One of their emerging use cases, one they are actively piloting, is helping political parties influence how AI search engines respond to election-related queries. In favor of one candidate over another.
I am not going to name the company or the parties. What I will say is that this is happening now, in multiple countries, and it is not hypothetical.
I have stayed away from politics for my entire career… Deliberately because I think I have more value to add on topics that sit outside that arena.
I have no interest in turning GrowthWaves or any other of my platforms into political commentary.
And as a disclaimer, I am not a political advisor or a campaign strategist.
But this topic sits squarely in my lane. AI search optimization is what I spend my days thinking about.
So… the moment someone tells me the same tools and techniques we use to get a SaaS company recommended by ChatGPT are being used to get a political candidate recommended by ChatGPT, I pay attention.
The Shift That Changes Everything
Here is the thing about traditional search:
Google was never neutral. We all know that.
Rankings could be gamed
Ads were everywhere
SEO itself is a form of influence (read: manipulation)
But the user still had options.
You typed a query, you got ten links, you clicked the ones that looked right.
You could window-shop. You could cross-reference. You could form your own opinion by reading three articles from three different sources.
AI search does not work that way.
When someone asks ChatGPT or Google’s AI Overview/Mode “who should I vote for in the midterms” or “what does Candidate X stand for,” they get one synthesized answer.
One voice. One framing.
The user does not see the ten sources that informed that answer. They do not see what was excluded. They do not see whose content was weighted more heavily.
That is a very different power structure. And when I say different, I mean unprecedented.
This Is Already Happening
Stanford researchers ran an experiment during Japan’s February 2026 general election.
They queried five major AI models from OpenAI, Google and xAI, asking each one to recommend a political party based on 36,300 synthetic voter profiles.
The result: models overwhelmingly directed voter profiles with left-leaning policy positions toward a single party. The recommendations were not balanced or randomized.
The models had biases. And those biases shaped the output.
Now layer on what I learned from the company I mentioned in the intro.
If models already have biases, and if those biases can be further shaped by the same content optimization techniques we use in our industry, you can see where this goes.
Campaigns & Elections, the trade publication for the political campaign industry, published a guide on how campaigns can win AI search.
The recommendations read like an AEO playbook: publish more content, make your site AI-friendly, diversify media outreach, repurpose positive coverage.
This is not a fringe idea. Campaign professionals are already treating AI search as a channel to optimize.
The Window Shopping Problem
I keep coming back to this phrase: window shopping.
In traditional search, voters could window-shop information. Ten blue links. Different sources. Different angles.
Different biases, yes. But at least the user was exposed to variety and could triangulate.
AI search removes the window. There is only one display. One answer. And the voter has to trust that the answer was assembled fairly.
AI companies are aware of this risk. OpenAI published their election safeguards for the 2026 midterms. They prohibit using their tools for voter demobilization or deceptive campaigning.
They are working with Democracy Works on voter registration and investing in deepfake detection using SynthID watermarks. Anthropic has published similar safeguards.
Those are good steps.
But here is my question:
If a political party hires an AEO agency or tool to optimize their candidate’s web presence so that AI models cite that candidate more favorably, is that a violation of any policy? Or is that just marketing?
That gray area is where I think the real risk lives.
The Gray Area
Let me be clear about something:
I am not saying that using AI search optimization for political campaigns is evil.
Every new communication channel gets used for political messaging eventually.
Television
Radio
Digital Ads
It would be strange if AI search were somehow exempt!
The problem is transparency.
When you see a political ad on Facebook, it is labeled as a political ad. When you see a sponsored result on Google, it says “Sponsored.”
Regulations vary by country, but in most democracies there are rules about disclosure.
AI search has none of that.
If a campaign optimizes their content so aggressively that ChatGPT cites their candidate’s policy page as the primary source for “where does Candidate X stand on healthcare,” the user has no way of knowing that result was engineered.
It looks organic. It feels organic.
But... it is not organic.
The New York Attorney General’s office tested multiple AI chatbots and found they frequently provided inaccurate information about voting procedures.
If models can get basic logistics wrong, imagine the room for manipulation on subjective questions about candidate positions and policy trade-offs.
What Makes This Different From Regular SEO/AEO
Someone reading this might think:
George, this is the same argument people made about Google SEO and elections. What is new?
Two things are new.
First, the answer consolidation. Google SEO could push a page higher in rankings, but the user still saw multiple results.
Author’s Note: I know there were features such as featured snippets, but you can’t equate featured snippets to something such as a synthesized AI search answer.
AI search collapses those results into one answer.
The user who gets a synthesized response about a candidate is getting a single editorial judgment that they did not ask a human editor to make.
Second, the speed of influence. Our AEO survey data (from the survey Kevin Indig and I just ran with 599 respondents) shows that 52.8% of organizations have their SEO team handling AI search optimization.
Which means that the infrastructure to do this work already exists.
Campaign teams could plug into established AEO workflows tomorrow. Given that Campaigns & Elections is already publishing playbooks for this, some of them already have.
I Am Not Being Alarmist
I want to be careful here. I believe AI search is a good thing.
It expands the search pie and makes information for more complex topics easier to access.
And I believe that political campaigns using new channels to reach voters is a normal part of democracy. Campaigns adapt. They always have.
What concerns me is the absence of disclosure.
When a voter asks an AI model about a candidate and gets a confident, well-sourced answer, they deserve to know whether the content that shaped that answer was produced by a neutral source or by the candidate’s campaign team.
(Which is almost impossible in the context of AI search and LLMs, I know.)
Right now, they cannot know that. And that gap, between what feels organic and what is actually engineered, is where trust erodes.
Final Thoughts
I wrote this note because I could not stop thinking about that conversation.
Someone in my professional network told me, casually, that they are helping political parties shape how AI talks about their candidates. As if it were any other client engagement.
Maybe it is. Maybe this is just the next evolution of political marketing. I genuinely do not know.
What I do know is that the midterms are coming up in the US. Elections are happening across the globe constantly.
The tools to shape AI search results are real. They are commercially available. They are being used as you’re reading this note.
The next time you ask ChatGPT, Gemini, or Perplexity about a candidate, consider the possibility that someone spent good money making sure you got the answer you got.
I am not telling you what to think. I am asking you to think about where the answer came from.
Thank you for reading today’s note, and see you again next week.
Research Disclaimers and Limitations
GrowthWaves and its author are not sponsored by or compensated by any company mentioned in this note. This is independent editorial analysis and does not constitute investment, financial, or legal advice. The author may have relationships with, work with, or hold equity in companies referenced; however, no content in this piece was influenced, commissioned, or incentivized by any such relationship. AI tools were used as a research assistant in the preparation of this piece. All claims are sourced and linked throughout.
Sources
Stanford Freeman Spogli Institute, “Voters Increasingly Use AI as Political Advisor. A New Study Shows the Risks.”
Anthropic, “An update on our election safeguards”
Campaigns & Elections, “Five Ways Campaigns Can Win AI Search”
New York State Attorney General, “Protecting New York Voters from AI-Generated Election Misinformation”
GrowthWaves x Growth Memo, “The State of AEO: What 599 Marketers Told Us About AI Search”