- Answer engine
- A system that returns a written answer, often with citations, instead of only a list of links. ChatGPT, Perplexity, and Google AI Overviews are examples.
- AEO
- Answer Engine Optimization. The work of making a brand understandable and citable when people ask assistants for a recommendation, a definition, or a comparison.
- GEO
- Generative Engine Optimization. The companion discipline that shapes what generative systems repeat, including sources off your own domain.
- LLM
- Large language model. The model family behind many assistants. AI Solo does not claim to control model weights or training runs.
- Citation
- A visible reference from an assistant answer to a URL or named source. A mention of your brand without a link is not a citation.
- Mention
- Your brand or product name appearing in an answer. It may be positive, neutral, or absent of any link.
- Share of answer
- How often your brand appears across a defined prompt set, compared with the competitors you named. It is a report metric, not a guarantee.
- Prompt set
- The list of buyer questions we baseline and re-check. Growth and Scale plans cap how many prompts are in the active set.
- AI Overview
- Google’s AI summary on some search results. It is one surface among several, and Google decides when it appears.
- Entity
- A distinct thing a model should not confuse: a company, product, person, location, or offer. Entity work is disambiguation, not a trick.
- Entity optimization
- Making those identities consistent across your site and the profiles assistants already read.
- Schema markup
- Structured data on a page that describes the content in a machine-readable way. It supports extraction. It does not force a citation.
- Knowledge panel
- A boxed summary on a search engine for a well-defined entity. Helpful context, not the same thing as an assistant citation.
- llms.txt
- A plain-text guide some sites publish for agents. AI Solo publishes one, plus a longer llms-full.txt. Crawlers may ignore both.
- Agentic SEO
- Using software agents for research, drafts, and checks while a person owns the plan and approves customer-facing changes.
- Parasite SEO
- A tactic that places content on someone else’s strong domain mainly to borrow their authority. AI Solo does not sell this.
- Digital PR
- Earned coverage and expert commentary on relevant publications. For us it is a source strategy, not a promise of links.
- Unlinked mention
- A brand name in an article or answer without a hyperlink. Still useful to track, because assistants quote names.
- Query fan-out
- An assistant expanding one user question into several searches. Your page may be retrieved for a subquery you did not write.
- Retrieval
- The step where a system fetches documents before it writes. Being retrieved is earlier than being cited.
- Grounding
- Tying an answer to sources. Products implement grounding differently. We report what the product showed, not how the lab trained it.
- Hallucination
- A confident answer that is not supported. Our pages try to be easy to quote so the model has less reason to invent your offer.
- Canonical
- The URL you want treated as the primary version of a page. Important when printers, tags, and parameters create copies.
- Indexation
- A search engine choosing to keep a URL. Requesting indexation is a step we log. Compliance is the engine’s decision.
- Internal link
- A link from one of your pages to another. Used to point assistants and people at the passage that answers the question.
- Answer block
- A short, quotable passage near the top of a page that answers one question without the surrounding sales copy.
- Prompt baseline
- The first recorded run of the prompt set, used later to describe change. It is a starting photograph, not a target rank.
- AI-referred session
- A visit your analytics attributes to an assistant or AI surface. We only discuss it when the source data supports the label.
- White-label
- Delivery under the partner’s brand. The work and the approval rules stay the same.
- Founding offer
- The December 2026 onboarding terms: limited slots, a 12-month rate lock on Growth and Scale, and a 72-hour kickoff subject to capacity.
- How we measure
- We report mention, citation, and share of answer on an agreed prompt set, plus your analytics when you grant access. We do not guarantee placement.
- Early access
- A platform feature that exists only in a limited or internal form. It is not something we will pretend you can log into today.
- Rolling out
- A platform feature we are putting into client delivery but do not yet treat as generally available self-serve software.