Human behavior and algorithms are evolving, and artificial intelligence is at the center of many of the search options we have today. ChatGPT is continuing to gain ground on Google as people’s preferred place to do informational-based searches.
Increasingly, our clients are reporting that prospects are discovering them via Large Language Model search as AI platforms become the new way many people find brands and information. It’s no surprise because LLMs act as answer engines that conveniently aggregate and serve up plain-language info. And in most aspects of life, convenience wins.
As such, many companies feel a sense of urgency to evolve their digital content and comms strategies to influence LLM-powered search. You’re not alone if you feel like it’s confusing to unpack a fluid situation and make confident moves.
Before tossing out your existing content marketing practices, it’s worth noting that web search is still an important inbound channel for plenty of brands, and SEO best practices matter. But showing up in LLM searches is a whole new game. It’s about creating repeatable impact in the places and on the conversation topics the model has seen and learned from.
Here are six tips to support success with AI search visibility:
You need a clear narrative. The fundamentals of strong storytelling are even more important in the age of AI. Clear language is easier for the model to learn. A strong, crisp narrative helps the LLM understand your content more accurately, connect it to relevant queries, and prioritize it as a high-signal answer.
Another way to think about it is that the pattern of a story—the arc—gives AI the structure it needs to think and the meaning it needs to match words with someone’s intent. If we think about prospects using ChatGPT to solve a pain point they’re facing or answer a need they have, then we want to map our storytelling to that conversation. We want our narrative arc to clearly lay out the challenge and solution from the standpoint of understanding the customer deeply.
From here, look for opportunities to write or speak in terms of specific patterns that ChatGPT and others favor, including frameworks like “3 steps to X” or “top tips for X” or “The ABC Model,” as well as analogies, FAQs, and step-by-step guides. We’ll be talking more in a future blog post about other LLM-optimization strategies for content, also known as Generative Engine Optimization (GEO).
Know your audience to map to key conversations. Use sales team insights to inform strategy. Consult with sales to understand what questions prospects and clients are asking—and if they’re asking ChatGPT or another LLM.
Understand your closed leads. What content did they interact with? What pain points does it speak to? What questions does it answer? Then compare sales team intel and customer content journeys with your SEO keyword list and AI chatbot logs.
Pair this with keyword research into long-tail terms (longer and more specific search phrases) and, together, you’ll have a cross-section of data to help you create a target list of LLM prompts that will inform storytelling and content strategies around the conversations you want to be known for.
Benchmark and learn from LLM search performance. Once you have your list of target prompts ready, run a search for those prompts across the LLMs (ChatGPT, Gemini, Perplexity, etc.). Understand where your brand shows up in answer outputs and which queries trigger mentions (or not) of your company, product, or content. Run competitor comparisons for the same prompts. Learn from the data, implement strategies, and repeat the searches frequently to get a usable sample.
Of note, the market is growing for tools to automate the process of tracking and improving brand mentions in AI search, with options like Cognizo, LLMrefs, Share of Model, and Trackerly.ai to assist in this process.
In terms of strategies, it’s important to pay attention to what kind of content is consistently showing up in LLM responses and from what places. If, for example, frameworks and structured checklists are cited more frequently and YouTube transcripts and LinkedIn posts get referenced, then you can lean into these patterns in your own content.
PR and awards feed the model. ChatGPT and LLMs pull from everywhere, from website content and articles on news sites to blogs, forums, GitHub, Wikipedia, and social platforms. Investing in third-party mentions matters. The more places your story lives online, from LinkedIn and Medium to awards lists, other listicles, podcast transcripts, and interviews, the more data points there will be to feed the model, which means the more chances to be surfaced.
AI models are trained on frequency as a signal of importance, so if your name and story appear in multiple sources, then AI is more likely to associate you with that topic and the phrases that show up around those appearances, which is why a consistent, pre-determined narrative is so important. Pro tip: Even YouTube captions are a good place to reinforce a topic or conversation association.
Think omnipresence, not authority. While Google prioritizes authority signals like backlinks and traffic, LLMs don’t care about that. They’re pattern recognizers that predict the most likely next word or phrase based on the text they’re trained on, not web rankings. In fact, we’re seeing firsthand some brands having success with AI search visibility despite having lower website authority rankings.
When ChatGPT responds to a question, it’s not saying, “This source is more credible because it’s an expert.” It’s saying, “This kind of answer usually follows a question like this in the data I’ve seen.” So our job is to teach the LLM the right pattern by being clear, consistent, and repeatable online. This is again why having a clear narrative and honing in on specific conversations is so important to support consistent storytelling across all channels, including your earned media strategy.
From a PR standpoint, it makes sense to cast a wider net with earned and paid opportunities to increase the volume of mentions that map to the conversations (and associated prompts) you want to own. Also, it warrants re-thinking the definition of Tier 1 results, in the context of business goals and also now in consideration of the fact that on-message coverage in a niche blog could be the thing that delivers a lead from an LLM search result.
Add structured data to your website. Recently, Google’s Search Relations Team confirmed that structured data is important for AI models. This is information that is organized in a labeled format, for example, tables, FAQs, headings, and bullet points, as opposed to unstructured data like paragraphs of free-flowing text. Structured data makes it easier for AI to read, understand, and reuse your content. It’s not just about formatting, it’s about being recognizable and repeatable in the language of machines.
What other tips do you have? I’d love to keep the conversation going. Contact: [email protected].
Want to read more about how to use AI to be successful in SEO? Check out this blog from our Orchestra sibling agency, Message Lab.
