FAQ
Questions people actually ask us.
Grouped rather than stacked in one list. Where an answer needs more room than a paragraph, it links to the post that works through it properly — including the ones where our answer is that the evidence is thinner than the category pretends.
The basics
What it is
What does snowdust.ai actually do?
It measures how AI answers describe, compare and recommend your brand across a fixed set of prompts, explains why a competitor gets recommended instead, and turns that into specific fixes. Three groups of capability: Analytics to see it, Optimization to act on it, Search Intelligence to connect it to your existing search data.
What is AI Discovery Intelligence?
Our name for tracking brand position inside AI answers over time, rather than taking a single reading. It runs on a four-part engine — Brief (where you stand), Lab (why), Actions (what to change) and the AI Discovery Analyst (a person to argue with).
Which AI models do you track?
Every plan covers all major platforms — ChatGPT, Perplexity, Claude, Google AI Overviews, Google AI Mode, Google Gemini and Microsoft Copilot. Each is scored separately rather than blended. A brand can be recommended on one and absent from another, and a single average hides exactly the gap you need to see.
How is this different from an SEO tool?
SEO tools measure whether you rank on a results page. This measures whether you are in the answer, and whether you are the recommendation. They are related — which is why Google Search Console and DataForSEO plug in — but strong rankings do not guarantee an assistant names you.
Measurement
How the numbers work
AI answers change every time. How is any of this reliable?
A single check is an anecdote. Tracked prompts are sampled continuously across models, so what you read is a trend built from many responses rather than one lucky run. Every figure stays traceable to the individual answers behind it.
What is the difference between a mention and a recommendation?
A mention is your brand appearing anywhere in the answer, including as the cheaper alternative. A recommendation is the answer telling the reader to choose you. They move independently, and conflating them is the most common reporting mistake in this category.
How many prompts should we track?
Enough to cover the decisions your buyers actually make, and enough that one odd answer cannot swing the percentage. Starter covers up to 50 and Growth up to 200; Enterprise is unlimited.
Can we compare your numbers with another tool's?
Not meaningfully. The category has no shared definition of share of voice, citation rate or position — vendors use the same words for different formulas. Pick one tool and track the trend inside it.
Does adding prompts improve our score?
It changes what you measure, not how you perform. Adding branded prompts in particular inflates a visibility number without anything about your brand changing. We keep branded and unbranded separate for that reason — the long version is here.
Plans & pricing
What it costs
What does it cost?
Starter is $49/mo for one brand and up to 50 prompts. Growth is $155/mo for up to four brands and 200 prompts, and adds Shopping Analytics. Paying for the year up front brings those to $39 and $145 a month. Enterprise is custom. Full comparison on pricing.
Is there a free option?
Every plan starts with a 14-day free trial, so you can measure your own brand before paying for anything.
What does Enterprise add?
Unlimited brands and prompts, white labelling, SSO/SAML, and no cooldown between on-demand scans. See enterprise.
Can we change plan later?
Yes. Pricing is monthly and you can upgrade, downgrade or cancel. Bear in mind that changing the tracked question set breaks the continuity of your trend line, so plan changes and question-set changes are worth doing deliberately.
Data & security
Data, security and getting it out
Do you have SOC 2 or ISO 27001?
We are not going to claim a certification on a marketing page. Write to security@snowdust.ai and you will get a straight answer about where we actually are, which is a much better outcome than discovering it during procurement.
Where does the data come from?
From the answers the models give to your tracked prompts, collected continuously, plus the search data you choose to connect through Google Search Console and DataForSEO.
Can we get the data out?
Ask us about export before you sign rather than after. It matters more than people expect: because metric definitions differ between vendors, the raw answers and your prompt set are the parts that actually carry across a switch.
Can other tools read our snowdust.ai data?
That is what the MCP Server is for — it exposes your visibility data to the assistants and agents your team already works in, and lets you pull your own systems in alongside it.