Everything you wanted to ask about AI visibility.

AI search & answer engines: the basics

If you're new to AEO/GEO, start here. These answers explain what's changed about search and why every brand needs an AI visibility strategy.

• What is AI search optimization?

  AI search optimization is the practice of making your brand visible in the answers generated by AI assistants like ChatGPT, Claude, Gemini and Perplexity. Instead of optimizing for a ranked list of links, you optimize for inclusion in a synthesized answer. It's also called Answer Engine Optimization (AEO) or Generative Engine Optimization (GEO).
• What is the difference between AEO, GEO and SEO?

  SEO targets ranked link positions on a search engine results page. AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization) target the answer itself — what a generative model says when a buyer asks a question. The metric of success shifts from rankings to citations and brand mentions inside generated text.
• Are AEO and GEO the same thing?

  In practice, yes. The two acronyms cover the same discipline — making your brand more likely to be named, cited, and described favorably inside AI-generated answers. AEO emphasizes "answer engines" (ChatGPT, Claude, Gemini, Perplexity) while GEO emphasizes "generative" output. We use them interchangeably.
• Why does AI search matter for B2B and SaaS?

  B2B buying journeys now routinely start in a chat window. Buyers ask "best CRM for my industry," "is X better than Y," and "who's a credible alternative to Z" — and they trust the answer. If your brand isn't in that answer, you're not in the consideration set.
• Will Google search still matter?

  Yes — for now. But it's increasingly mediated. Google's AI Overviews, ChatGPT search, and Perplexity all summarize answers above traditional links. AI visibility is becoming the layer that decides whether a buyer ever clicks at all.
• Where do AI models pull their information from?

  Models combine training data with retrieval-augmented sources. In practice, that means Wikipedia, Reddit, high-authority trade press, structured data on the public web, and product schema. Many traditional "rank-the-blog" SEO investments are invisible to these systems.
• How is "AI visibility" measured?

  It boils down to four numbers per prompt: are you mentioned at all (mention rate), where in the answer are you (position), how favorably (sentiment), and what citations were used (sources). Intendity rolls these up into a single 0–100 visibility score and a per-competitor share of voice.
• Why are Wikipedia and Reddit so over-represented in AI answers?

  Both are continuously crawled, structured (Wikipedia infoboxes, Reddit upvote signal), and pre-filtered for quality by the models themselves. That makes them disproportionately "trusted" inputs. Most AEO programs include a Wikipedia track and a Reddit/forum-engagement track for that reason.
• Is AEO a fad or here to stay?

  AI assistants now route a meaningful share of category-defining questions away from Google. Even if the exact tools change, the discipline — measuring brand mentions inside generated answers — won't. The teams setting baselines today will spend the next two years compounding wins.

How Intendity works

• What does Intendity do, exactly?

  Intendity is a monitoring and optimization platform for AI search. You define your brand and your competitors, list the questions your buyers ask, and Intendity runs those prompts against ChatGPT, Claude, Gemini and Perplexity on a schedule. We capture each response, score visibility and sentiment, surface citations, and recommend specific content and structured-data changes to win the answer.
• Which AI models do you track?

  Today we track ChatGPT (OpenAI), Claude (Anthropic), Gemini (Google) and Perplexity. We add new models as they reach meaningful adoption with B2B buyers.
• How often are prompts re-run?

  On the Pro plan we run every prompt automatically every day, and you can trigger on-demand runs at any time. The Free plan is manual only — two runs per day across your prompts, on your schedule.
• How does Intendity decide which prompts to run?

  You define them. We seed a starter set based on your category and competitors, you accept or reject suggestions, and you add your own. Suggested prompts continue to surface weekly on Pro — generated from gaps the system detects in your coverage.
• Can I track competitors?

  Yes. You add competitors when you set up a brand, and Intendity tracks their share of voice across the same prompts. You'll see exactly which prompts they win and the citations driving that win.
• How many competitors can I track per brand?

  There's no hard cap. Most teams settle on five to ten named competitors plus the long-tail Intendity discovers automatically — a competitor that gets mentioned in your prompts is added to the watchlist whether or not you named them upfront.
• Do you support multiple languages and regions?

  Yes. AI answers vary by language and region. Intendity captures answers per locale so you can target localized PR, content and schema work where it actually moves visibility.
• What's the difference between a "run," a "mention," and a "snapshot"?

  A run is one (prompt × model) execution — the raw answer. A mention is the structured analysis of that run — was your brand named, where, and how. A snapshot is the daily roll-up across every run for a brand — your visibility score on a given day.
• Can I export the data?

  Yes. PDF reports are available on every plan. CSV export and full programmatic access via the REST API and MCP server are on Pro.
• Do you offer page audits?

  Yes. Intendity audits the public pages on your brand's domain for AI-readiness signals — schema.org coverage, llms.txt presence, AI-bot robots rules, FAQ schema, structured data — and produces a 0–100 score with prioritized fixes.

Models & coverage

• Which exact ChatGPT/Claude/Gemini/Perplexity versions do you query?

  We use the production model your buyers actually see — not a smaller or older variant. Specifically the latest GPT model on ChatGPT, the latest Claude Sonnet/Opus on Anthropic, Gemini Pro on Google, and the default Sonar model on Perplexity. We update versions as providers ship new ones.
• Do you use the same prompts on every model?

  Yes. The same prompt is sent to every model for a given run, with provider-native parameters. That gives you a like-for-like comparison of how each assistant answers the same buyer question.
• Why do answers vary so much between models?

  Different training data, different retrieval pipelines, different default behaviors around citations. ChatGPT and Perplexity tend to cite the public web aggressively; Claude leans on training data and is conservative with names; Gemini blends in Google's own search index. Intendity captures the differences instead of papering over them.
• Will my visibility score be the same in every region?

  No — and that's informative. AI assistants serve different answers by locale, especially for "best X near me" or country-specific listicles. Intendity captures runs per region so you can spot a market where you're strong in EN-US but invisible in DE-DE.
• How do you handle browsing/web-search modes inside the assistants?

  For ChatGPT and Gemini we use the web-search-enabled mode where available, since that's closest to what a buyer experiences. Perplexity is web-search by default. Claude is queried both with and without browsing depending on the prompt — we surface both signals so you see the difference.
• Will you add Meta AI / Grok / DeepSeek / Mistral?

  Yes, on a rolling basis as adoption with B2B buyers grows. If a model is showing up in your buyers' workflows and we don't cover it yet, tell us — we prioritize by demand.

Accuracy, methodology & limits

• How is the visibility score calculated?

  The daily visibility score is the share of your brand's mentions over the total runs that ran that day, expressed as 0–100. So if 31 of 48 (prompt × model) runs mentioned you, your visibility for that day is 64. Sentiment and position are tracked separately so a high mention rate with negative sentiment is visible, not hidden.
• How do you detect a "mention"?

  A mention is detected by an LLM-based parser that reads each answer and extracts: whether the brand was named (including known aliases and the official name), in what position, in what sentiment, and which competitors were named alongside. Each mention carries a confidence score so you can filter low-confidence parses.
• Can the parser get it wrong?

  Yes — extraction is probabilistic. We attach a confidence score to each mention and a hand-correction loop is on the roadmap. In practice, agreement with hand-coded baselines runs in the 90%+ range across the prompts we've tested.
• How do you handle ambiguous brand names?

  When you set up a brand we let you add aliases (e.g. "Acme", "Acme Corp", "ACME") and a domain. The parser uses the domain plus the alias list to disambiguate from other entities with the same name, and flags low-confidence matches for review.
• Are model answers reproducible?

  Not perfectly — language models are non-deterministic by design. We don't pretend otherwise. Instead, the daily run captures the distribution: across hundreds of (prompt × model) executions per week, real signal emerges and one-off variance averages out.
• What if a model "hallucinates" facts about my brand?

  You'll see it. Intendity surfaces the exact context the model produced — including incorrect pricing, deprecated features, or a confused founder name. The recommendation engine then suggests the source-level fix (Wikipedia edit, schema update, or trade-press correction) most likely to displace the hallucination.

Setup, results and ROI

• How long does setup take?

  Under five minutes. You enter a brand name, the domain, and a few competitors. Intendity pre-fills competitive data and suggests a starting prompt set.
• How quickly will I see results?

  Brand-level visibility scores are available right after your first run. Strategic improvements — net-new mentions, better positioning, citation wins — typically emerge in 2–6 weeks once recommended content is published and re-crawled.
• What does a "good" visibility score look like?

  It depends on your category and competitive set. Intendity benchmarks you against named competitors and the implicit set of brands the models surface for your prompts, so you can see whether you're winning, on par, or losing the conversation. Most B2B SaaS brands start somewhere between 10 and 30 and target a sustained 50+ in their core category.
• Will Intendity actually publish content for me?

  No. We tell you exactly what to publish, where, and why — but we don't ghostwrite or distribute. This keeps editorial integrity with your team and avoids the kind of generic content the models routinely ignore.
• What happens if a competitor is dominating my category?

  Intendity shows you the citations behind their dominance — which Wikipedia entries, Reddit threads, listicles or trade outlets the models are pulling from. From there you can target the same sources, build new ones, or correct factual errors that work against you.
• How do I know whether a recommendation actually worked?

  Each recommendation is tied to the prompts it should move. After you publish or correct the cited source, Intendity's daily runs pick up the change — usually within two to six weeks of the source being recrawled by the model. The recommendation card tracks the before/after delta so you can attribute the lift.
• What if I publish something and visibility doesn't move?

  Sometimes the model hasn't recrawled the source yet. Sometimes the source isn't the one models cite for that prompt. Intendity's recommendation engine learns from outcomes — a recommendation that consistently underperforms is downranked and replaced. Tell us when something didn't work and we feed it back into the model.
• Who in our team should own AEO?

  Most teams put it under Content/SEO with a thin slice of PR and product marketing. The day-to-day is content and source management. The strategic ownership is whoever is accountable for category narrative — usually a head of marketing or a head of growth.

Pricing & billing

Full pricing detail is on the dedicated pricing page. The most common questions we get on the FAQ are below.

• How much does Intendity cost?

  Free plan is €0 forever — one brand, two manual runs a day, three days of history. Pro is €99 per brand per month — unlimited brands per seat, daily automated runs, full history, recommendations, email digests and the API. Enterprise is custom for agencies and multi-brand teams. See the pricing page for the full feature matrix.
• Do I need a credit card to start?

  No. You can sign up and run your first brand on the Free plan without entering payment details.
• How does per-brand pricing work?

  Pro is €99 per brand, per month. One brand at €99/mo, two brands at €198/mo. You add or remove brand seats from inside the app and your invoice adjusts on the next billing cycle.
• Is there an annual or enterprise plan?

  Yes — Enterprise covers volume pricing for 10+ brands, annual invoicing, custom DPA/MSA, dedicated onboarding and SSO. Get in touch at [email protected].
• Are prices in EUR or USD? Do they include VAT?

  List price is EUR exclusive of VAT. Stripe handles conversion at checkout and collects VAT where required, with EU reverse-charge for businesses with a valid VAT ID.
• Can I cancel any time?

  Yes. Cancellation is one click in the customer portal and takes effect at the end of the current billing period. We don't pro-rate partial periods, but you keep Pro access until the period ends.

Privacy, security & GDPR

• Where is my data stored?

  Intendity runs on managed cloud infrastructure inside the EU with encryption at rest and in transit. Customer data is logically isolated per account using row-level security at the database layer.
• Do you use my data to train AI models?

  No. We never train third-party or proprietary models on your account data. Prompts are sent to model providers solely to obtain answers for your visibility report.
• Do the model providers (OpenAI, Anthropic, Google, Perplexity) train on my prompts?

  We use API access (not consumer accounts) for every provider. OpenAI, Anthropic and Google contractually do not train on API traffic by default. Perplexity's API has the same posture. Each provider's terms govern that interaction.
• Are you GDPR-compliant?

  Yes. We process personal data lawfully, support data access and deletion requests, and document our processing activities under a register of processing activities. See our Privacy Policy for full detail.
• Can I sign a DPA?

  Yes. A standard data processing agreement is available on Pro and required on Enterprise. Email [email protected].
• Who are your subprocessors?

  Our hosting provider, Stripe (payments), Resend (transactional email) and the LLM providers we query on your behalf (OpenAI, Anthropic, Google, Perplexity). The complete list with country of processing is published on our security page.
• Can I delete my account and data?

  Yes. You can delete your account from settings, which removes your brands, queries, runs, mentions and outputs from active systems. Backups are purged per our retention windows in the Privacy Policy.
• How is account access secured?

  Email + password with passwordless magic-link option, plus Google OAuth. Sessions are issued by Supabase Auth and revocable from any device. SSO is on the Enterprise plan.

Comparisons & alternatives

• Why not just check ChatGPT manually once a week?

  Manual checks miss the variance — answers shift between sessions, regions and model versions. Intendity runs the same prompts at scale, captures the variance, scores it, and tells you what changed. It's the difference between a snapshot and a baseline.
• How is this different from a brand monitoring tool?

  Brand monitoring listens for mentions on the public web and social. Intendity listens inside the AI itself — what the assistant actually says when asked, including answers that never produce a public mention.
• Do I still need SEO if I use Intendity?

  Yes. SEO and AEO are complementary. Some AEO recommendations explicitly include SEO work — for example, ranking on a Wikipedia-cited source page or reaching the top of a "best of" listicle that models pull from. See our SEO vs AEO comparison page for the full breakdown.
• Why not just use Google Search Console for this?

  Search Console measures clicks and impressions on Google. It tells you nothing about whether a model named you in an answer that never produced a click. AEO is the layer above — a buyer can read your name in a chat answer and never visit your site.
• How is Intendity different from a "rank tracker for AI"?

  Rank trackers focus on position. We track position and mention rate, sentiment, citations, share of voice against named competitors, and recommend the actual changes that move visibility — not just measure it.
• Can I use Intendity alongside my existing SEO stack (Ahrefs, Semrush, etc.)?

  Yes — they're complementary. AEO often drafts off the same source authority signals SEO tools surface, and our REST API + MCP server make it easy to push Intendity data into the same dashboards your team already uses.

API, MCP & integrations

• Is there a public API?

  Yes. A read-only REST API is available on Pro and Enterprise. It returns brands, queries, runs, mentions, daily visibility scores, recommendations and competitor share of voice — full reference at /docs.
• What is MCP and why does Intendity have an MCP server?

  MCP (Model Context Protocol) lets AI clients like Claude Desktop and Cursor call tools running on your infrastructure. Our MCP server exposes the same data the REST API does, but as tools the assistant can call directly during a conversation. Useful when you want an AI agent that can answer "how is our visibility trending?" in real time.
• How do I authenticate?

  Generate an API key in Settings → API keys inside the app. The plaintext is shown once. Pass it as a Bearer token on every REST request, or as a header on the MCP transport. Same key works for both.
• Are there rate limits?

  Yes — generous and account-scoped. The REST API targets 100 requests/minute per key by default, with cursor pagination on the roadmap for collections that grow past a single page.
• Can I push data into a warehouse?

  Yes. Most teams pull the daily visibility, mentions and competitor SoV endpoints into a warehouse on a schedule. Webhooks for real-time push are on the roadmap — tell us if that's blocking you and we'll prioritize.

Agencies & multi-brand teams

• Can one account manage multiple clients?

  Yes. On Pro you add a brand seat per client and they're fully isolated under one login. On Enterprise you get volume pricing, white-label PDF reports and a dedicated workspace.
• Can I share a client report without giving them an account?

  Yes. Every brand has a downloadable PDF report. Enterprise lets you white-label that report with your own logo and contact information.
• Do you have a partner program?

  Early-stage. If you're an agency offering AEO/GEO services and want to refer clients or co-sell, get in touch at [email protected].
• Can I get a single invoice for multiple clients?

  Yes — it's the default on Enterprise. Pro is invoiced per workspace; consolidation is an Enterprise feature.

Troubleshooting & support

• My first run produced no mentions — is something wrong?

  Probably not. A zero-mention result is a real signal — your brand isn't in the answer for those prompts today. Check the runs tab to see exactly which competitors were named instead, and the recommendations tab for what to publish to change that.
• A run failed or timed out — what now?

  Failed runs are flagged in the runs tab with the underlying error. The most common cause is an upstream provider hiccup — re-running the prompt usually succeeds. If a run consistently fails, email [email protected] with the run ID.
• Why is the parser saying I'm mentioned when I'm not (or vice versa)?

  Each mention has a confidence score. Low-confidence matches are typically ambiguous brand names or aliases that need disambiguation. Adding the official domain and aliases under brand settings resolves most cases.
• How do I get support?

  Email [email protected] from any plan; Pro and Enterprise customers get priority response. We answer most product questions inside one business day.