7 Steps to Get a New Brand Into AI Search Results From Scratch

If you are launching a new brand today, the online visibility landscape looks nothing like it did even two years ago.

For now, Google is still the dominant discovery engine. But AI-powered search, such as ChatGPT, Perplexity, Google’s AI Overviews, Gemini, Claude, and others, now sit alongside it as a parallel recommendation layer that increasingly shapes which brands people consider and which ones they skip entirely.

For completely new brands, this poses a significant hurdle. If you have no history, no backlinks, no reviews, no editorial mentions, then you are invisible to both systems.

Another problem is that AI search does not simply mirror Google’s top ten results. As we’ve explained in our article about how ChatGPT and similar tools recommend businesses, while AIs do look at search rankings, they appear to prefer external validation by third party sources as well as clear messaging by brands.

Continue reading: How ChatGPT Decides Which Businesses to Recommend

Additionally, as Perplexity’s team explained in a Search Engine Journal interview, AI answers are context-dependent and user-dependent – meaning that the same query can produce different results depending on who asks and in what context.

Consequently, a new brand that builds the right foundation deliberately can show up in AI answers faster than you might expect – sometimes before it cracks page one on Google.

That said, there is no shortcut. You cannot prompt-hack your way into recommendations. What follows are seven steps to get a brand new company indexed by LLMs and other AI tools – all of which build upon each other.

1. Define your positioning

Before you publish a single page, you need to answer one question with painful specificity: what category should AI systems place you in when someone asks for recommendations?

In my experience, one of the most common mistakes new brands make is trying to occupy multiple categories at once. I have seen an agency that positioned itself as both a services firm and a learning platform. The result was that AI systems could not classify it cleanly – and it showed up in neither category consistently.

Your positioning strategy should start with the prompts your likely customers would actually type into ChatGPT or Perplexity. The actual words a potential buyer would use.

If someone asks “best sustainable packaging supplier for DTC brands in Europe,” the AI system needs to find consistent evidence that you are exactly that – a sustainable packaging supplier, serving DTC brands, operating in Europe. Every ambiguity you introduce dilutes that signal.

The GEO research paper introduced a benchmark called GEO-bench (a smart word for AI SEO), specifically for measuring visibility in generative engine responses. Their experiments showed that deliberate optimization for generative retrieval environments improved source visibility by up to 40%, though the effect size varied significantly by domain.

In practical terms, it means you need to decide your category before you build anything else, because every subsequent step depends on this clarity.

Pick one category. Own it. You can expand later, but only with the right clarity to avoid confusing AI tools.

2. Build your identity seed set

Once your positioning is defined, your next step is to make your brand easy for AI systems to identify consistently. They need to find, analyze, and verify your identity across multiple surfaces before you start producing content at scale.

Think of this as your identity seed set: the minimum footprint that tells both Google and AI crawlers “this brand exists, this is what it does, and here is where to verify it.”

The minimum set includes your homepage with clear, structured information about what you do and who you serve. An about page that names your team, your location, and your specific expertise.

Also make sure to add directory listings relevant to your niche. I’m not talking about fifty generic directories, but the ones that matter in your vertical.

You will also need review site profiles on the platforms your buyers actually check (e.g. Trustpilot for consumer brands, G2 for agencies).

And last but not least, clear USPs and trust signals that a machine can extract without interpretation.

According to the Perplexity interview, AI systems don’t only evaluate whole pages. They can also pull specific snippets from pages, which means small, clear text blocks often travel better than long, dense page narratives. If your brand information is scattered, inconsistent, or buried in hard-to-analyze layouts, the retrieval system has less usable material to retrieve.

You should put the emphasis on consistency, not volume. Your brand name, positioning, core offering, trust signals, and geographic scope should match exactly across every surface in your seed set.

Get this wrong, and you will spend months wondering why AI systems recommend your competitors instead.

3. Create your proof signals

AI systems should be able to find your brand after step two. But finding it and recommending it are different things. To move from discoverable to recommendable, you need proof – evidence that your claims are real.

This is where many new brands stall. They build a polished website, list themselves in directories, and then wait. But AI systems – much like a cautious human – look for corroboration before making a recommendation.

If you claim to be the best at something, the model wants to see that claim validated somewhere outside your own domain.

For a new brand, the minimum proof stack depends heavily on your niche, geography, and competitive landscape. In high-competition verticals where your competitors have years of PR coverage and editorial mentions, launching without any press at all can make recommendation extremely difficult.

That said, proof does not always mean PR. The priority order for most new brands should be

Continue reading: 5 Steps to Build External Validation for AI Recommendations

• genuine testimonials and reviews first
• third-party directory and list inclusions second
• team credentials and background information third
• Organic and paid PR fourth

The exact weighting shifts by industry – a B2B SaaS company might need G2 reviews and analyst mentions, while a local service business needs Google Business Profile or Trustpilot reviews.

The key is that every claim on your site should have at least one external source that corroborates it. Not because a human will check. Because the model will.

Fortunately, these steps will not only help with getting your new business recommended by AI, but also increase how many users will act on their recommendations.

After analyzing over 3,000 informational queries across 42 organizations from June 2024 to September 2025, research by Seer Interactive found that brands cited as sources within AI Overviews received 35% more organic clicks and 91% more paid clicks compared to brands that were not cited.

4. Create an AI-friendly content strategy

With your identity established and your proof stack in place, you can now build the content assets that AI systems will actually retrieve and reference.

But the content strategy for AI visibility is not the same as a traditional SEO content calendar.

You are optimizing for the specific types of prompts users enter into AI systems – and each prompt type requires a different kind of asset.

Moreover, you need to consider how AIs execute prompts. For example, if a user asks “which provider is best for me?”, the AI will go through several sub questions before presenting an answer, which could be:

• Discover (“what are the best X?”)
• Compare (“X vs Y”),
• Validate (“is X any good?”)
• Choose (“which X should I pick for Y use case?”).

Each of these requires content that AI can extract and reuse in a generated answer. The Perplexity team made this point explicitly: their system performs sub-document processing, pulling specific snippets and tokens from pages rather than retrieving whole documents wholesale.

The system is not sending users to your page – it is pulling information from your page into its own answer. Your content needs to be structured for extraction, not just for reading.

In practice, this isn’t fundamentally different from SEO. You will need clear headings that match likely prompt language. Concise, self-contained paragraphs that can stand alone when pulled out of context. Specific data points rather than vague claims. Comparison tables where appropriate. And factual density over word count.

Since AIs think in more broad terms, specific keywords do not appear to be as important anymore. It’s more about the actual information that is being presented.

5. Identify what works in your specific industry

Here is where most generic AI SEO advice falls apart: what works in one vertical does not necessarily work in another. For some niches there are very specific things to look out for.

Malte Landwehr, CMO at Peec AI, pointed out a striking example in ecommerce: the shopping widget inside ChatGPT is powered by Google Shopping data. If your products do not appear in Google Shopping results or your product feed attribution is weak, this means you are invisible in one of the highest-intent AI surfaces that exists.

Did you know the shopping widget in ChatGPT is powered by Google Shopping?

You heard that right. No ACP, UCP, Agentic Checkout, or direct feed integrations.

OpenAI is simply scraping Google Shopping.

My Peec AI colleague Tom Wells even reverse engineered which scraping… pic.twitter.com/jIQaO5z2Om

— Malte Landwehr (@MalteLandwehr) March 5, 2026

The fix is not more blog content – it is checking your Google Merchant Center setup and optimizing your product data feeds.

In our own research on how ChatGPT decides which businesses to recommend, we also noticed that it relies heavily on external listicles in some sectors (e.g. agencies), while focusing on established players with a long history in others (e.g. law firms).

SEO consultant Aleyda Solis noticed a few more interesting differences based on Similarweb’s Generative AI Brand Visibility Index. For example, finance brands should look into publishing unique research, while electronics brands should focus on technical specification depth.

🚨 Comprehensive Generative AI Brand Visibility Index in 2026 by @Similarweb with actionable insights across sectors and recommendations from specialists based on the findings – Here are some patterns based on outperformers per sector:

* Finance: Research sites punch above their… pic.twitter.com/tTnsIvIGDB

— Aleyda Solis 🕊️ (@aleyda) March 5, 2026

While all brands need to fulfill certain basics, these examples can translate into fundamentally different playbooks.

Before you invest in content production, study how AI systems handle recommendations in your specific vertical. Run the prompts your customers would use. Note which brands appear, what format the answers take, and what sources are cited.

Your strategy should be shaped by what you observe, not by a cookie cutter framework. The industry-specific mechanics will dictate whether you need more reviews, more editorial coverage, more product data, or more technical documentation.

6. Build external validation loops (aka PR)

Steps one through five give you a foundation. Step six is what makes AI systems confident enough to actually recommend you repeatedly.

Corroboration is the recurring theme across every study on AI visibility, including the ones mentioned in this article.

For a new brand with limited budget, external validation does not require a six-figure PR retainer – although, so far, it appears to be the more the better.

Inclusion in curated lists and directories can be relatively affordable, if the authors agree to include you for free. Product reviews in niche media can often be achieved by sending samples to reporters or users.

Community contributions that get referenced can be stimulated by incentivizing users (in ways that don’t go against platform TOS of certain platforms, of course). Even guest contributions on relevant industry publications can be an affordable way to get started.

The options split into two paths: in-house outreach, which is slower and cheaper, or working with a PR agency, which is faster but more expensive. Both can work. The choice depends on your timeline and resources.

What matters is the loop – not a one-time push, but ongoing reinforcement. Every new third-party mention adds another data point that AI systems can use to validate your brand. Over time, these mentions compound. The model encounters your brand in enough independent contexts that recommending you becomes a low-risk output.

7. Track mention quality and improve

The last step is the one most brands skip – and it is the one that separates brands that maintain AI visibility from those that appear briefly and then fade.

Even if your business is being recommended now, that could change at any moment.

For example, our own research showed us that AIs appear to value recency. If your organic PR and reviews are from several years ago and there’s no fresh validation, that could be a problem.

You also shouldn’t count out your competitors – they will likely be optimizing for GEO too, which means you could easily get pushed out of recommendation sets.

And finally, if AIs encounter negative information about your brand, which could be anything from a large number of negative reviews, an online shitstorm, or negative press articles – that could easily cost you the trust of AIs.

Therefore, it’s absolutely essential to stay on top of your AI rankings.

There are software solutions, like the DataForSEO LLM mention API, which allow you to track any mentions of your brand across AI queries.

You can also simply open Gemini, ChatGPT, Claude, or Grok and run a fixed prompt set once a month.

The prompts that matter most fall into three categories:

• “best X for Y” prompts that test whether you appear in recommendation sets
• “X vs Y” prompts that test whether AI systems consider you a viable alternative
• “why not X” prompts that reveal what evidence gaps are keeping you out of answers.

When you are not appearing, the diagnostic question is always the same: what evidence is missing? Do other brands have more proof? Does the AI understand who you are? Did it find anything negative about your brand?

The good news is that you can often take out the guesswork by asking the AI why it doesn’t rank your business. Although this information should always be taken with a grain of salt, it can certainly help to figure out the problem quickly.

Quick execution checklist

The bottom line

Getting a new brand into ChatGPT, Claude, and Gemini search results from scratch is not a mystery.

The brands that will win in this environment are not necessarily the ones with the biggest content budgets or the most aggressive SEO tactics. They are the ones that understand what both Google’s algorithms and AI retrieval systems need to see before they are willing to put a name forward as a recommendation.

That bar is not impossibly high. But it is specific. And for new brands willing to build methodically, the window is wide open – because most competitors are still optimising exclusively for a search paradigm that is already shifting underneath them.