Generative artificial intelligence and large language models (LLMs) are transforming how people access information and changing which sources are viewed as trusted enough to be shown. Traditional SEO, built by page rankings and keyword optimization, has expanded into AIO (AI Optimization) and GEO (Generative Engine Optimization), where the goal isn’t just to rank but to be retrieved, understood and cited by AI systems in their chats with human users. Where people once searched their questions in Google and clicked on a handful of websites to find the best answer, they’re now asking AI chatbots to do the searching for them and deploying AI agents to complete increasingly complicated tasks. Beyond simply finding answers, people are using agents to build decks, write code, automate customer experiences and organize data, with little to no human involvement.
As amazing as that is, it’s a big change for communicators looking to maintain brand visibility online. Gartner forecasts up to a 25% drop in traditional web searches this year compared to two years ago, and up to a 50% hit in organic search traffic by 2028, as more people embrace chatbots over search engines.
The challenge for communicators is ensuring your organization’s expertise is referenced within AI-generated answers, rather than being buried behind them. This is less about manipulating algorithms than about earning trust and building retrievability. At 3BL, we’ve helped more than 1,500 companies and nonprofits tell their stories online — from the early days of Twitter chats and live Google Hangouts, to engineering for AIO and GEO today. While we do the work in the back-end to make sure our clients’ content is well-structured and readable to AI systems, here’s what we tell them about how to format their articles, press releases, and other content for AIO and GEO.
Inside the Machine: How LLMs actually find and use information
To optimize for AI, we need to understand how AI systems “think,” or how they find and process information. At the heart of AI visibility is RAG: Retrieval-Augmented Generation. Here’s what that means, in plain English:
- Retrieval: AI-powered search experiences pull relevant and recent information from trusted sources such as researchers, nonprofits, government websites, and news providers like 3BL.
- Augmentation (Grounding): The system blends what it retrieved from the first step with its model training data so the response is based on real source material rather than memory alone.
- Generation: The model writes an answer for the user based on a mix of retrieved content, pretrained knowledge and the user’s query.
In other words, visibility in AI comes down to:
Retrievability + Authority + Structure.
This is what’s often referred to as “RAG-ready content.” It’s legible to both machines and people. The good news for communicators is that a lot of the things we already do to tell great stories are also good for AIO and GEO, and a few intentional tweaks can make a big difference.
Writing for AIO and GEO: Designing content that works for both humans and machines
On today’s internet, you’re really writing for two audiences: the human user (i.e., your target audience) and the AI chatbot that is more and more likely to be assisting them. Think of your content as two-layered: It’s organized in a way that can be parsed and referenced by AI systems, but it still reads naturally to people.
Writing for people is second nature to communicators, but what about writing for robots? It may sound foreign, but it’s not as different as you may think. Checking these six boxes can help.
Structure your content with intention.
AI systems see the headline, description snippet and headers as key pieces of information to determine how your content can be used to answer people’s questions and perform tasks. Lead with the answer or most important point first, then add context, citations, examples and storytelling that connects with the reader while still being easy for AI systems to parse.
- Your headline should clearly communicate what the content is about in less than 60 characters.
- Use the description to add context about why this topic matters and why your organization is positioned to speak about it.
- Within the content, the first header (H1) should mirror your headline, using words that communicate authority or nod toward search intent where relevant (i.e., “how,” “why,” “answered,” “explained”).
- Secondary headers (H2, H3) help break up your content and transition from one topic to the next, which is more readable to robots than a long block of text.
- Keep each section modular, using short and self-contained sections with one clear idea to avoid long, blended paragraphs covering multiple topics.
- Add descriptive alt text to images to help visually impaired readers and AI systems interpret the visuals you use.
Formats like Q&As and lists are also more likely to rank in AI search because they speak directly to people’s questions and provide valuable information in response. For example, consider reformatting the next executive op/ed on your content calendar from a long thought piece into a Q&A that showcases their expertise in a digestible and discoverable way, or talk about your next big program launch in a list that’s oriented toward solving problems.
Add substance to improve visibility.
Including quotes, statistics and links to sources not only makes your content more valuable and trustworthy to human readers, but it can also increase visibility in AI queries by up to 40%, according to an analysis of 10,000 real searches. When citing a source or mentioning an announcement, be clear about time orientation (i.e., “as of 2025”) to inform readers and signal recency to AI systems.
Use a clear and authoritative tone.
An eye toward clarity in your writing helps engage human audiences and showcases answer density to AI systems.
- Clearly define the industry or topic-specific terms you use, and break down steps explicitly to help both humans and robots follow along.
- Avoid the passive voice (subject, past participle, object, i.e., “mistakes were made”) in favor of declarative statements (subject-verb-object, i.e., “cats are mammals”).
- Where your organization has authority to speak on a topic, use an authoritative tone to communicate this to your audience as well as their AI assistants.
Mention your brand and focus topic, but not too much.
Just like with SEO, keyword optimization is helpful, but keyword stuffing is not. Including explicit details like company names, platforms, frameworks, products, people, and locations helps give valuable context to both human and machine, but mentioning keywords too much can be a red flag to AI systems. Aim to keep your core keywords below 3% density (about five mentions in a 600-word article). This goes for your brand name, too. Mention it once in key areas like your headline, description, H1 header and first paragraph where applicable, but avoid overuse by incorporating synonyms elsewhere (i.e., “our,” “us,” “we,” “our team”).
Foster content hubs to build credibility and brand association.
Create “hubs” for the topics you’d most like to be known for, such as a page on your website, a category of your blog or a campaign in your 3BL newsroom. Update these pages as you publish relevant content, and link back to them often so crawlers associate your brand with these topics and begin to see you as a trusted source for information about them.
Publish to trusted domains.
As well as on your own website, publishing to trusted domains and syndicating to distribution networks like 3BL’s is shown to improve AI visibility. AI systems read laterally across the internet, meaning the more your content is out there and in trusted places, the more likely it is to be cited in chats.
The bottom line
Writing for AIO and GEO isn’t all that much different from writing good content for human audiences. But it does call for communicators to be intentional about clarity and structure in their writing and strategic in how and where they publish. You can’t buy placement in AI answers (yet), but you can earn it by providing valuable information often and proving your organization is a source AI systems can depend upon for answers. Still have questions about writing for AIO or how 3BL can help? Get in touch below.
(Image credit: Mariia Shalabaieva/Unsplash)
