AI Use Policy

In short

• AI services are optional and paid with points; points are charged only when the service succeeds.
• Requests are processed by Anthropic, OpenAI or Google outside Egypt, mainly in the USA, after personal data is automatically removed from them; your account data is not sent to them.
• Outputs may be wrong; they are general information, not investment, legal or tax advice.
• Your project's metrics enter an investor's report only with your explicit consent after accepting the intro.
• The platform makes no automated decision with legal or similarly significant effect on you.

1. Services Covered

1. All services are optional and paid with points under the "Subscriptions, Points and Refunds Policy".
2. In the investor analyst, numerical KPIs are computed by deterministic code; the model only explains and phrases them.

2. Model Providers and Data Processing

1. Requests are processed by one of the following providers: Anthropic, OpenAI and Google, under paid API accounts. Their servers are outside Egypt, mainly in the USA.
2. The platform routes each request to a model according to task complexity and cost, and falls back to another provider if the first fails. Your request may therefore be processed by any of the listed providers.
3. Under these providers' API terms, data sent is not used to train their models. [For review: verify each provider's current terms and their data retention for abuse monitoring]
4. We log usage metadata: provider, model, token counts and cost.
5. Inputs and outputs are stored with your sessions and reports and are deleted when you delete your account.
6. Only anonymised content, as described in section 3 below, is transferred to providers; the "Privacy Policy" sets out the international transfer provisions.

3. Removing Personal Data Before Sending

1. The platform applies an automatic anonymisation layer to every AI request before it leaves the platform, so that providers receive anonymised content without personal data.
2. Account data (name, email address and mobile number) is never sent to providers.
3. Your full name and parts of it, your email address and your mobile number are replaced with placeholders wherever they appear in a request you send, as are the other party's details when a request contains content relating to them. Placeholders are restored inside the platform when the reply is shown to you, without the provider ever seeing the original.
4. The anonymisation layer detects the following identifiers in any text and replaces them with placeholders: email addresses, phone numbers, Egyptian national ID numbers (14 digits), numbers resembling Saudi and UAE ID numbers, payment card numbers (after a Luhn check), international bank account numbers (IBANs), and links to personal social media profiles.
5. We log the number of replacements made in each request, not the replaced values.
6. In the AI review within a Verified Handover, only the scan report summary and the README file are sent, after the same anonymisation, including the seller's account details; the code is never sent, and the review takes place only with the seller's permission. See the "Verified Handover Terms".
7. Limits of anonymisation: automatic detection cannot guarantee the removal of every indirect identifier written in free text, such as names of other people typed by the user. We therefore ask you not to include other people's personal data in AI requests.

4. Project Data in Analysis

1. An investor's report includes only project data made available to them after the intro was accepted.
2. Monthly metrics are used in an analysis for an investor only if the founder enabled "Allow AI analysis" for that project.
3. If the founder withdraws consent or deletes their account, the content of investor reports built on that data is redacted.
4. The sponsor market report is built on aggregated, anonymised data, and any cell with fewer than 3 projects is hidden.

5. Limits on Relying on Outputs

1. Outputs may be inaccurate, incomplete or out of date, even when based on sources.
2. Outputs are general information, not investment, legal, tax or accounting advice, and do not replace a qualified adviser.
3. Execution assistant answers are general procedural information. Always check with the competent official authority before taking any step.
4. Match scores and ratings are informational only.
5. You are responsible for reviewing outputs before relying on or sharing them.

6. Automated Decisions

1. The platform makes no automated decision producing legal or similarly significant effects on any user.
2. Accepting or declining intro requests, investment decisions and accepting handover items are decided by people.
3. AI outputs are not used to approve or suspend accounts or to verify identity.

7. Ownership and Use of Outputs

1. You may use outputs you requested for your lawful purposes, subject to third-party rights and the providers' terms.
2. We do not guarantee that outputs are original or free from similarity to other users' outputs.

8. Prohibited Use

Using AI tools to extract other users' data, bypass restrictions or for unlawful purposes is prohibited under the "Acceptable Use Policy".

9. Reporting

If you notice a harmful or seriously wrong output, report it to support@octasolutions.net.