About our AI assistant
Last updated 21 November 2026
You are talking to an AI
The AI guide, the project chat, the portfolio builder and "Help me invest safer" answer with an AI model, not a person. Every chat says so in its first message.
What it uses
It answers only from Crowdinform's own data: our platform and project pages, scores, risk levels, checked red flags, Deep Risk Reviews and the documents a platform publishes for a project. It is instructed not to use general knowledge for figures. When we have no data, it should say so.
What it can get wrong
AI can misread a document, mix up two companies, quote an out-of-date figure or state something with more confidence than it deserves. Check any fact that matters to you on the linked page or in the platform's own documents. Tell us about a wrong answer with "Report a mistake" or at [email protected]; we use reports to improve it.
What it does not do
- It does not give personal investment advice. It does not know your finances, goals or experience and must not ask for them.
- It does not tell you to buy, keep, sell, move or withdraw any investment, and it does not judge your portfolio as a whole.
- An example plan from the portfolio builder is an illustration made from today's projects, not a plan for you.
- It does not place orders or contact platforms for you.
"Help me invest safer"
You type the names of the platforms and lenders you use. The assistant shows, for each one, the same Platform Score, Crowdinform Score, risk level and checked red flags that every visitor can see, in the order you gave, and points you to the Deep Risk Reviews. It never asks for amounts. It may offer to add the items to your watchlist; nothing is added unless you press the button. It is information on our published method, not advice about your money.
Allowances
- Explore: 5 AI questions a month. One "Help me invest safer" run counts as one; each follow-up question counts as a new one.
- Pro: up to 30 AI questions a day.
- Portfolio builder: Explore 1 a month, Pro 10 a month.
The allowances that apply when you use the assistant are shown on the plans page.
Fair use
Do not use the assistant to copy our data in bulk, to get around its instructions or limits, or to post its answers as our opinion or as professional analysis. Do not type other people's personal data, passwords or card details.
Your data
Your questions and the answers are processed by our AI providers (Anthropic, OpenAI) to produce the answer, and kept for 365 days in your account, then deleted automatically. We review samples, without your name or email, to test and improve the assistant. Details: Privacy policy, sections 4.8, 10 and 15.
Who is responsible
Crowdinform OÜ chooses the models, writes their instructions and decides what data they may use. We test the assistant with a fixed set of questions before each change. This notice is part of our Terms of use (section 13.3).