We use necessary storage to keep you signed in and save your preferences. Optional analytics helps us understand product use. It stays off unless you allow it.
Shared language for AI visibility, generative engine optimization, and how DoneHat monitors and improves your brand's presence in AI answers.
Whether your brand appears in answers when buyers ask AI and search systems about your category, competitors, or products. Unlike traditional search rankings, AI visibility depends on how systems interpret your content, structured data, brand facts, and third-party citations rather than only keyword position on a results page.
A conversational or answer-style system that generates natural-language responses to user queries. DoneHat launch scans start with ChatGPT observations and can add extra answer sources such as Google, Claude, Gemini, and Perplexity when configured. Each uses different models, indexes, and retrieval behavior, so visibility can vary across them.
The share of sampled AI answers in which a brand appears compared with named competitors for the same buyer questions. It should be interpreted with sample size, answer source, prompt mix, and confidence range rather than treated as a permanent market-share number.
The specific AI or search surface where an answer check is run, such as ChatGPT or another configured provider. Different answer sources can cite different pages, update at different speeds, and produce different recommendations for the same buyer question.
A crawler or user agent operated by an AI or search provider to retrieve public web content. AI crawler access can affect whether public pages are available for retrieval, but allowing a crawler does not guarantee that an answer system will cite or recommend a page.
A single data point: one buyer question checked on one AI or search answer source. Each answer check records platform, answer text or failure reason, mention status, competitor mentions, sources when available, extraction confidence, and timestamp.
A fix task created from an issue. An action contains root-cause analysis, a fix strategy, and drafted deliverables grounded in the evidence ledger. Actions move through a pipeline: created, in progress, review, approved, published, proven.
Optimizing content to appear in direct-answer systems like ChatGPT, Google Featured Snippets, and other AI answer surfaces. AEO prioritizes clear, structured answers to specific questions, FAQ schema markup, and concise product descriptions that answer models can extract and cite directly.
A buyer-intent question that DoneHat checks with AI engines on your behalf, for example "best project management tools for startups" or "alternatives to Slack." DoneHat auto-generates questions based on your category and competitors, and you can add, edit, or remove them at any time.
The verified attributes DoneHat extracts from your website during setup: product names, pricing tiers, founding date, target audience, key features, and competitive positioning. Brand facts serve as ground truth for detecting inaccurate claims in AI responses.
A per-answer metric indicating how strongly the AI engine recommended your brand in its answer. High confidence means you were the primary recommendation with positive framing. Low confidence means you were mentioned briefly, in a list, or with qualifications.
A range that describes uncertainty around a sampled measurement such as mention rate. In AI visibility reporting, a confidence interval helps teams avoid overreading small samples or treating one answer as a stable market truth.
How easy it is for AI and search answer systems to extract, understand, and cite a page. Citeable pages use clear headings, short answer blocks, visible facts, structured data where appropriate, internal links, and source-backed claims.
An issue type that appears when a competitor is named as the primary recommendation across two or more buyer questions while your brand is absent or downplayed. It indicates the competitor has stronger public signals in the model's training data or retrieval sources.
A concrete output produced from a supported finding: FAQ page drafts, JSON-LD structured data snippets, page rewrite suggestions, entity alignment briefs, or comparison page outlines. Every deliverable passes through governance review and requires human approval before use.
How well your content maps your brand's attributes (features, pricing, use cases, differentiators) to the concepts and categories that AI models associate with buyer-intent queries. Strong entity alignment means the model can confidently match your brand to relevant buyer questions.
DoneHat's assisted workflow for turning supported findings into reviewed deliverables. It diagnoses likely root causes, plans a fix strategy, drafts deliverables such as FAQ content, structured data, and page rewrites, then submits them for governance review.
The process of fixing AI visibility issues, from diagnosis through content changes to proof. Fix strategies include adding structured data, creating FAQ content, building comparison pages, improving entity alignment, and increasing third-party mentions.
The review layer that keeps fixes aligned with configured brand facts, proof points, differentiators, and forbidden claims. DoneHat uses these fields to detect inaccurate claims and to keep drafts from introducing unsupported positioning.
The practice of optimizing your online presence for AI-powered answer engines and AI search surfaces. Unlike traditional SEO which targets search result rankings, GEO focuses on structured data, entity relationships, citation density, and content depth that AI models use to construct answers.
A visibility problem detected during a scan. Issue types include missing mention (brand absent from the answer), weak mention (brand mentioned in a subordinate position), competitor dominance (competitor recommended over you), inaccurate claim (factually wrong information), and missing context (key differentiators omitted).
A source link or source-derived reference used by an AI answer system when constructing an answer. A useful LLM citation should point to a page that visibly supports the answer rather than to unrelated or stale content.
Your brand name or product identifier appears in the AI engine's natural-language answer — not only in a list of source links below the answer. DoneHat distinguishes between strong mentions (primary recommendation) and weak mentions (passing reference).
The AI engine answered a buyer-intent question relevant to your category but did not reference your brand at all. This is the most common issue type for brands that are new to AI visibility monitoring. It signals that the model lacks sufficient evidence to include you in its answer.
Content you control — your website, blog, documentation, help center, and social media profiles. Owned sources are the foundation of your AI visibility because you can directly improve them with better structured data, deeper content, and entity alignment.
A targeted repeat check of the questions that originally surfaced an issue, executed after a fix is published. A proof check compares the new AI responses against the original scan and reports whether visibility improved, regressed, stayed unchanged, or was inconclusive.
A group of related buyer-intent questions that target the same topic or purchase intent. For example, "best CRM for small business," "top CRM tools 2025," and "CRM software comparison" form a group. DoneHat groups questions to identify systematic visibility gaps rather than one-off misses.
A composite metric (0-100) that summarizes supported AI visibility evidence. It combines supported mentions, source evidence, answer consistency, and known failure coverage. Unsupported model-judge claims are rejected instead of counted as wins.
The process by which an AI or search answer system fetches external information to answer a question. Retrieval can use live web results, indexes, partner sources, or other systems, which is why answer freshness and citations vary by platform.
The established practice of optimizing web pages to rank in traditional search engine results (Google, Bing). While SEO and GEO share some foundations (quality content, technical markup), AI engines weight signals differently — third-party citations and entity relationships matter more than backlink profiles and keyword density.
Machine-readable markup (typically JSON-LD) added to your web pages that explicitly tells AI engines what your content means. Relevant schema types include FAQ, HowTo, Product, Organization, and SoftwareApplication. Structured data helps AI retrieval systems extract accurate, structured answers from your site.
Content about your brand that you don't control — review sites (G2, Capterra), industry publications, directories, forum discussions, and Wikipedia entries. Third-party sources act as independent training signals for AI models and are critical for building authority.
A complete monitoring scan that checks tracked buyer questions across enabled AI and search answer sources. A visibility scan captures a snapshot of your AI visibility at a specific point in time. Subsequent scans are compared to detect improvements, regressions, or new issues.
Your brand appears in the AI answer but in a subordinate position: listed as an alternative rather than a primary recommendation, mentioned with qualifications ("You might also consider..."), or included without key differentiators that would help the buyer choose you.