Phasewheel In Brief

Most web pages are written for humans. But humans aren't the only ones reading anymore, and bot traffic keeps climbing. In this episode of Phasewheel in Brief, Kiley Hylton and Caitlin Morin explain how to structure content so AI models can pull it and cite it, without losing the people who read it too. You'll learn why bots reward short, self-contained answers, what a concise answer block is and how to write one, whether FAQs still matter for AI visibility, how long content should run for the best citation odds, and the fastest place to start if you have no idea where to begin. Phasewheel helps brands get found, cited and trusted by AI engines. Want help making your site and blog citable? Email [email protected]. Q: Why doesn't well-written content get cited by AI models? Most pages are written for humans, but humans are no longer the only ones reading the web, and bot traffic climbs day over day. Human-friendly prose is full of filler and long phrasing that bots skip. AI models pull information the way they talk to themselves: short, clear and direct. To get cited, content has to be built for extraction first while still making sense to people. Q: What is a concise answer block? A concise answer block is a self-contained chunk of 40 to 100 words that states the answer to a single question up front, like the answer to an FAQ. If a model only read that one block, it would have everything it needs. Think of it as your thesis statement: put the summary and the answer first, then explain below it. Q: How do I write a concise answer block? Repeat the question and give the answer in the same chunk, the way a reality TV interviewee restates the question before answering. Keep it self-contained, skip the jargon and filler, lead with the most important point and place it as high on the page as you can. Completeness matters more than exact word count: one clean block that wraps the whole idea beats content split across paragraph breaks. Q: Do FAQs still work for AI visibility? Yes. FAQs work because of their format, question then answer then question then answer, which is clean structured data. It's the structure that matters, not the "FAQ" label. A concise answer block is basically an FAQ answer. You can build the same structure without the FAQ label, but FAQs are still the easiest way to hold the information clearly for models and humans alike. Add FAQ schema to your pages. Q: How long should content be to maximize citation odds? Roughly 40 to 100 words per idea, and try not to run past 100. What matters more than length is completeness: each chunk should stand on its own. Models like to return a whole chunk, so a page break or paragraph break in the middle of an idea weakens it. Q: What is EEAT and why does it matter? EEAT stands for expertise, experience, authoritativeness and trustworthiness. To make content citable, build credibility: use the expertise of people on your team, add credentials and back claims with as many facts as you can. Give an AI model a reason to trust and cite the source. Q: Where should a brand start if they don't know where to begin? Start with your top five (or top 10) blog posts. Use Google Search Console or your citation tool to find which ones surface most often. Add a concise answer block to the top of each one and add your schema, whether that's blog schema or org schema. You don't need to rewrite everything. Add the chunk to the top and watch what it does. Q: How can I check if my content is citable? Run your blog post through an LLM and ask what it thinks. It will give you feedback. And if you don't have the capacity to do this yourself, it's one of the things Phasewheel does. Reach out at [email protected] to talk through your website or blog.

Schemas, robots.txt and structured data determine whether AI engines can find, understand and cite your brand. In this episode of Phasewheel In Brief, Kiley Hylton sits down with Caitlin Morin to break down what schemas are, which ones you actually need and how to make sure AI crawlers can access your site. What you'll learn: • What schemas are and why they matter for GEO (Generative Engine Optimization) • The difference between a robots.txt file and a schema • Which schema types every website needs, regardless of industry • Whether you need FAQs on a page to add an FAQ schema • How to test if AI bots are actually crawling your site • Why llms.txt is a nice-to-have, not a must • When to hire a GEO consultant vs. doing it yourself Resources Mentioned • Schema validator: validator.schema.org • Google Rich Results Test • Screaming Frog (SEO crawler) • Cloudflare (for bot traffic monitoring and site speed) • Google Analytics (GA4): check for LLM referral traffic • Claude.ai: for generating optimized robots.txt files and schema code Q&A Q: What is a schema? A schema is a hidden piece of JSON code that lives in the backend of your website. It's invisible to visitors but tells AI crawlers and LLMs what is on your site, where to find it and why it matters. Think of it like an aisle sign in a grocery store. It doesn't hold the product; it just tells you where to look. Schemas are safe to add and carry near-zero risk of breaking your site. Q: What's the difference between a robots.txt file and a schema? A robots.txt file is the gatekeeper. It controls which crawlers and bots are allowed onto your site at all. A schema guides those bots once they're already inside. If robots.txt is the front door passcode, schemas are the directory once you're in the building. Q: Which schemas does every website need? Three schemas should be on every single page, regardless of industry: Organization (or LocalBusiness), Website and FAQ. Beyond those, prioritize based on what you sell: Product/Offer schemas for DTC brands, Service schemas for service businesses, Article/BlogPosting schemas for content and AggregateRating schemas if you have reviews. If you have a credentialed team member, a Person schema adds authority. Q: Do I need FAQs on a page to add an FAQ schema? No. You can add an FAQ schema to any page without having visible FAQ content on it. Pull your highest-traffic queries from Google Search Console and generate FAQ schemas based on those. Adding FAQ schemas across all pages is recommended for GEO purposes. This differs from traditional SEO guidance, and that's intentional. Q: Should I allow all AI crawlers on my site? Yes, with a few exceptions. Block crawlers from any page you don't want surfacing in AI results: checkout pages, patient portals, login pages or anywhere users enter PII. For the rest of your site, allow all major AI bots (GPTBot, ClaudeBot, PerplexityBot, etc.). Explicitly name the bots you're allowing in your robots.txt rather than using a blanket allow. You can generate an optimized robots.txt in Claude in under a minute. Q: What is an llms.txt file and do I need one? An llms.txt file consolidates your key schemas into one place for LLMs to reference. A recent Ahrefs study found only about 10% of llms.txt files are ever called by AI engines. It's a nice-to-have, not a priority. Invest in schemas, robots.txt and content quality first. Q: How can I tell if AI bots are actually crawling my site? Check your server logs through Cloudflare or your hosting provider. It's the most direct method. If you don't have server log access, use an SEO tool like Screaming Frog or check Google Analytics for LLM referral traffic. Your robots.txt tells you whether bots are allowed in; server logs tell you whether they're actually showing up. Q: Should my web developer handle schemas during a site build? Ask. But verify. Shopify automatically generates many schemas, and some WordPress themes (like WP Engine) include some schema support. Always check with the validator at validator.schema.org after a new build. For anything beyond the basics, working with a GEO consultant is worth it. Schema setup for most sites takes about two hours of consultation time and has an outsized impact on AI visibility.

Most brands focus GEO efforts on their website. The bigger risk is inside the building. In this episode of Phasewheel in Brief, Kiley Hylton and Caitlin Morin break down the organizational side of Generative Engine Optimization, why inconsistent messaging across sales, marketing, product, and customer support teams creates signal noise for LLMs, and what to actually do about it. FAQs answered in this episode: Why is my brand being described inaccurately by AI? LLMs like ChatGPT, Claude, and Perplexity pull from every available source — your website, social posts, help articles, case studies, and even old pages you forgot to remove. When those sources say different things, AI averages them into one description. Inconsistent internal messaging is often the root cause. What is a tiger team and how does it help with GEO? A tiger team is a small, cross-functional group built to solve a specific problem without a full reorg. For GEO alignment, the recommended structure is seven people: a program lead with decision-making authority, an SEO and AI search strategy lead, a content lead, a web engineering lead, and an analytics lead as core members — with product marketing, legal, and a sales or customer success rep rotating in as needed. The team runs on a fixed six-month charter with a weekly sprint cadence (Monday planning, midweek async unblocking, Friday demo and decisions) and defined decision rights so changes get published without committee delays. Read the full blueprint → phasewheel.com/bloglink/building-an-ai-search-tiger-team-in-a-mid-market-organization How do I audit how AI describes my brand? Open 5–6 LLMs (ChatGPT, Claude, Perplexity, Gemini, etc.) in incognito mode and search for your company. Compare how each one describes you. Note any inaccuracies, missing services, or competitor mentions. The whole process takes under 30 minutes and requires no tools or budget. Does GEO work differently for e-commerce vs. service-based companies? The audit process is the same, but e-commerce brands need to go deeper, down to the product level. If your product feed descriptions are thin, LLMs will hallucinate specs or omit products entirely. Service companies tend to have brand-level messaging gaps. What should I do if AI is describing my company inaccurately? Start by gathering all your external-facing materials — website copy, social posts, sales collateral, support articles — and feed them into an LLM. Ask it to identify inconsistencies and gaps. That gives you a starting point before making any content changes. 🔗 Tiger team charter, RACI + 6-month plan → phasewheel.com/bloglink/building-an-ai-search-tiger-team-in-a-mid-market-organization 🔗 Learn more about GEO → phasewheel.com

What is GEO (Generative Engine Optimization), and how is it different from SEO? In this episode of the Phasewheel in Brief, Phasewheel’s Kiley Hylton (Director) and Caitlin Morin (Co-Founder & CTO) break down what it actually takes for a brand to show up accurately in AI answers, not just rank on Google. What you'll learn: Who is Phasewheel What GEO is The AI Discovery Infrastructure framework: Signal, Align, Illuminate, Sustain Why your website might be written in "French" while LLMs speak "Spanish” FAQ’s Question: Is AI Discovery infrastructure just SEO with a new label? Answer: No. However, there's a lot of overlap, things like structured content, and authority signals are important. Technical like hygiene clarity is really important. So those things are definitely overlapping and I do think SEO is very important and that's a foundation that that GEO is built on. You want to have a good base. The difference is the decision is happening at very different points. So SEO, you're optimizing for a click. GEO, you're optimizing to for your site for a citation basically. So you're optimizing to show up when someone's looking for you in an LLM. So the difference is that the actual automation point click versus citation. The difference is to you're going before you wanted to be found and ranked, now we wanted to be trusted and recommended. So it's more intent driven, it's higher trust, it's a lot more authority. The ranking signal changes from back links and page rank to whether your brand is being cited in a way that's accurate, consistent and trustworthy. Question: Which brands in the space today are getting it right and what are they specifically doing when we say getting it right? Answer: There's two different ways people are going to get right right now. The first one is infrastructure and then second one is content. Infrastructure, Glossier, Skims, Vouri and Spanks are like the four that are crushing it. They're already ready and hygienically ready for agentic check out. So that's what I would say is their infrastructure is great. Second thing is content. The strategist, Wire Cutter and Hubspot are all doing a good job. All of them have the answer first question print content that's very very clear and very very well structured and consistent. So the first 200 words of the content answer the question. The LLMs. want the answer as quickly as possible with as little resistance as possible because that's how they keep the users on their site. Question: What is Entity TrustAnswer: LLMs using your website as a source because you are defining what you offer in a very clear consistent way. Question:If you were a brand and you had 90 days, where do you start? Answer: First, do an audit. Go to different LLMs in incognito mode and search for your brand or questions about your vertical. Second step is look at your website and ensure its clear and technically sound (does it have schemas, do your meta titles and descriptions make sense, phone numbers align, names are spelled correctly and you have no errors like 301s). Lastly start putting out structured content. If you have the resources then engaged in PR is also great Work with us: [email protected]• Learn more: phasewheel.com