LLM SEO

LLM SEO is not a single technique. It is two distinct channels that serve different purposes. The first channel is answer engine tracking. You write the questions your buyers ask, and a tool runs those questions against the AI models on a schedule. It records the answer text, the websites the model cited, the brands it named, and the search queries it used internally. From these records, you compute your visibility, share of voice, and average position relative to competitors for each question and each engine.

The second channel is AI answer visibility, sometimes called source pitching. This reads the pages Google's AI Overviews cite when they answer a commercial question. When you search for 'best accounting software for small business' and see an AI answer, the links under that answer are the sources the model read. This channel finds those source pages for your niche, so you can pitch to be included on them. It is a subset of traditional digital PR, focused specifically on the pages that feed AI answers.

These are separate tasks. A tool that tracks answer engines records what ChatGPT and the others say about you. A tool that finds opportunities shows you where your brand could appear next. Some vendors offer one, some offer the other. Mentionry is an AI marketing platform that does both. It records what the answer engines say about your brand every day, and its Opportunities feature finds places like articles, podcasts, and journalist requests where you could earn a citation.

The market for LLM SEO tools is new and pricing models vary widely. Many tools sell access to a dataset of source pages or AI search results. You typically pay a monthly subscription for a number of credits or tracked queries. Prices range from about fifty dollars to several hundred dollars per month. Some vendors also sell managed services, where they commit to a number of pitches or placements per month for a fixed fee, often in the thousands of dollars.

A common model is the per-placement fee. A vendor might charge a flat rate, for example two hundred dollars, for each link they secure on a page that is a known source for AI answers. This turns link acquisition into a direct cost of sale. Another model is the success fee, where payment is due only if a link is obtained, but the fee is often a multiple of a standard outreach cost. These models treat links as a commodity to be delivered.

You should expect to connect API keys for some tools, particularly those that track multiple AI models. They need your OpenAI key for ChatGPT, your Google AI Studio key for Gemini, and so on. This delegates the cost of queries to your accounts. Other tools, especially those focused on source page discovery, may not require any keys. Mentionry requires no API keys for its core discovery and pitching engines. The only connection is your own mailbox, used to send the drafted outreach.

The primary risk is paying for activity instead of a measurable outcome. A tool that sells credits for scanning source pages is selling scans, not links. A service that commits to sending fifty pitches a month is selling email drafting, not placements. The outcome you want, a link on a relevant source page, is not under the vendor's control. It depends on the site owner's decision. Nothing in this field is promised, and any vendor that claims otherwise is selling on faith.

A technical risk involves data freshness. AI models change their answers, and the set of pages they cite evolves. A dataset of source pages compiled last week may already be stale. A tool that runs your tracking queries only once a day may miss fluctuations. There is also the risk of misattribution. A page might be cited by Google AI Overviews for one query but not for the commercial query you care about. Not all visibility in an AI answer is equally valuable.

The trade-off is between volume and precision. A broad scan of the web for any page mentioning a topic will yield many low-quality opportunities. A precise scan for pages that actually appear under AI answers for commercial queries will yield fewer, but more relevant, leads. The filtering is the product. In a measured run for one domain, Mentionry read 110 openings, judged 40 as plausible, and kept 14 as high-quality prospects ready for a pitch. The value is in the severe filtering, not the raw count.

First, ask what is actually being measured. If they talk about tracking rankings, are they tracking your position in AI answers for your specific prompts, or are they estimating search volume for keywords? Estimation requires a licensed opt-in panel, which most vendors do not have. Mentionry measures the answers to your own tracked prompts, run daily. It does not estimate what the world is asking.

Second, examine the connection requirements. A tool that needs your OpenAI, Anthropic, and Google API keys is passing the query costs to you. Check if those costs are capped or unpredictable. A tool with no keyless connectors may have limited reach. Mentionry uses nine keyless connectors for discovery, so you connect nothing to find opportunities. You only connect a mailbox to send the pitch.

Third, look for verification. After a pitch is sent, does the tool check back to see if a link appeared? A link can be removed or turned nofollow later. Mentionry verifies every pitched opportunity. It reports the page the link is on, the anchor text used, and whether the link is still followed. The scoreboard shows earned links, not sent emails. This is a critical difference.

Fourth, understand what is excluded. A reputable provider will exclude paid link placements, because buying links that pass ranking signals violates Google's policy. Mentionry's drafting model will refuse a reply that asks for payment for a link. It will also reject opportunities where it judges your business cannot honestly answer the query. This judgment happens for each opportunity individually.

Mentionry combines both channels of LLM SEO into one platform, but keeps their functions separate. The Track Answer Engines feature runs your custom prompts against the models on your plan, which can include ChatGPT, Google AI Overviews, Google AI Mode, Gemini, Claude, and Perplexity. It records the full answer text, citations, named brands, and internal queries every day. There is no backfill. Recording starts the day you add a prompt.

Opportunities is one part of the platform. It finds places your brand could appear, like pages that cite competitors, journalist requests, podcasts, and awards. It reads the pages Google's AI Overviews and other answer engines cite when answering questions. These are potential places to pitch your expertise. For tracking what the engines themselves say, you use the Monitor feature. This separation is important. One part finds opportunities. The other records answers from all the engines on your plan.

When an opportunity is worth pursuing, the AI Marketer drafts the email, form submission, or reply needed. The draft is sent from your own connected mailbox, using your address and domain. Replies land in your email thread. You approve every send. Mentionry sends the pitch, but it does not place, publish, or pay for a link. Whether a page then cites your brand is the publisher's decision. The product is built around this distinction.

There is no database of opportunities held on our servers. Drafts live in your browser. The state of a sent pitch is read live from your mailbox. This design means you own the data and the sending reputation. Pricing is a free seven-day trial, then ninety nine dollars a month for the Starter plan, which includes tracking and drafting. The Starter plan carries the full pitching engines; the trial writes every pitch and sends none. Enterprise pricing is quoted. An annual term is available at a lower effective rate.