AI & Automated Access Policy

1. Purpose of SAHIBAT’s AI

SAHIBAT develops and improves AI tools intended to assist participating businesses with customer acquisition, marketing and advertising. Uses may include business positioning, promotional copy, campaign ideas, service descriptions and recommendations for improving business visibility.

Subject to the permission described below, eligible content may be selected, cleaned, classified and transformed into datasets for training, fine-tuning, evaluation and improvement of these tools.

A model may learn from multiple participating businesses and assist other SAHIBAT members. Participation does not create an exclusive model for your business or guarantee access to a paid AI feature outside your purchased plan.

2. Eligible public business content

Subject to your opt-in and authority to grant permission, eligible content consists of business information you submit to SAHIBAT for public display, including:

• Business and trading names and business categories.
• Business descriptions and brand positioning.
• Descriptions of services, products and public offers.
• General service areas and market coverage.
• Business logos, product images and other business media for which you hold the necessary training rights.

Public availability alone does not make content eligible. Third-party rights, personal information, confidentiality and the exclusions below still apply.

Customer reviews, third-party testimonials and content copied from other websites are not covered by a listing owner’s permission alone.

3. Information excluded from this programme

The public-content training permission does not cover:

• Passwords, password hashes, authentication tokens, recovery codes or other access secrets.
• Card details, bank details, payment credentials, private invoices or financial account records.
• Private account fields, unpublished drafts, identity documents, verification records or confidential business documents.
• Private messages, support emails, enquiries, customer lists or individual lead information.
• Personal phone numbers, personal email addresses, home addresses or precise personal location information, even if displayed publicly.
• Sensitive personal information, children’s information, biometric identifiers or identifiable faces and voices.
• Visitor browsing histories, advertising identifiers or individual behavioural profiles.

These categories must be excluded from training datasets under this programme. A future feature requiring a different category of data would need a separate assessment, notice and any additional permission required by law.

4. Separate consent and content permission

Participation requires a separate affirmative opt-in identifying this training purpose. Opening an account, buying a membership, publishing a listing or accepting general terms is not by itself an opt-in to this programme.

Existing members whose content was collected without this permission must be given the relevant choice before their content enters the programme.

By opting in, you grant SM Marketing Management LLC a non-exclusive, royalty-free permission to reproduce, prepare, adapt and process your eligible public business content solely to develop, train, evaluate and improve SAHIBAT’s AI tools for the purposes described here.

This permission extends only to authorised providers acting on SAHIBAT’s behalf for those purposes. It does not transfer ownership of your original content, grant an unrestricted resale right, or authorise third parties to train their own independent general-purpose models.

You must own the relevant rights or have authority to grant this permission. Accepting these terms on behalf of a business does not automatically authorise processing of another individual’s personal information.

Personal-data consent and intellectual-property permission remain legally distinct. Neither removes mandatory rights under applicable law. No royalty or other payment is due merely because content participates in the programme, unless separately agreed in writing.

Refusing or withdrawing optional training permission does not prevent ordinary directory membership. A separately requested personalised AI feature may need its own limited processing to function, which must be explained when that feature is offered.

5. Processing safeguards and providers

Training must use a defined set of eligible content rather than unrestricted access to account databases, email inboxes or verification records. Access to datasets must be limited to authorised personnel and providers.

Providers acting for SAHIBAT must be subject to appropriate contractual restrictions on confidentiality, security, purpose, onward disclosure and independent model training.

Our website is hosted with Hostinger in the United States, while the company’s email server uses Hetzner in Germany. These arrangements do not establish where separate AI training takes place.

Before transferring participating content to an additional AI provider, SAHIBAT will disclose relevant processing and destination information and establish the lawful transfer arrangements and permissions required.

This permission does not authorise publication or sale of training datasets containing member content. Measures to reduce memorisation and inappropriate reproduction must form part of model evaluation.

6. Withdrawal, removal and existing models

You may withdraw training consent and the associated permission for future training by emailing support@sahibat.com. Identify your account or business listing; do not send your password.

Following a verified withdrawal, SAHIBAT will stop adding your content to training and exclude it from future training runs. Content must be removed from active training datasets unless a specific lawful retention obligation applies. Retained legal records must not be reused for training.

Removing a listing or requesting removal of particular public content also triggers review of that content’s continued inclusion in the programme.

Withdrawal does not make earlier lawful processing unlawful. However, deleting a source record does not necessarily remove information already learned by a model.

Where a model retains or can reproduce personal information, SAHIBAT must assess the request and take measures required by applicable law. Depending on the circumstances, these may include output suppression, access restrictions, unlearning, retraining or withdrawal of an affected model.

Technical difficulty does not create a blanket exemption from data protection duties. Genuinely anonymous learning may be retained where lawful, but a model will not be treated as anonymous merely because source records or names were removed.

7. AI-assisted outputs and user responsibilities

AI outputs may be inaccurate, incomplete, outdated or similar to material produced for others. They do not guarantee leads, sales, advertising approval, search rankings or commercial results.

Review generated material before using or publishing it. Check factual claims, prices, permissions, intellectual-property rights and applicable advertising requirements. Do not assume generated content is exclusive or automatically copyrightable.

Do not use the tools for deceptive advertising, impersonation, unlawful discrimination, unsolicited unlawful messaging or prohibited profiling. Do not submit confidential or sensitive information through a feature intended for public business content.

Training consent does not authorise autonomous ad spending, publication in your name or access to your customer accounts. Such actions require separate authority.

This policy does not authorise solely automated decisions with legal or similarly significant effects on individuals. Any such feature requires separate legal assessment, disclosure and applicable safeguards.

8. Automated access by third parties

Except where expressly authorised by SAHIBAT or permitted by mandatory law, you must not:

• Scrape, bulk-download, systematically extract or replicate listings, member content or contact information.
• Use platform content to assemble external training datasets, train independent AI models or build a competing directory.
• Harvest contact details for spam, profiling or unauthorised marketing.
• Bypass authentication, access restrictions, rate limits, anti-bot controls or other technical safeguards.
• Use bots to create accounts, submit listings, manipulate reviews, generate fraudulent referrals or distort analytics.
• Attempt to extract confidential prompts, personal data, restricted model information or another user’s private content.

Ordinary search-engine indexing may be permitted subject to applicable access instructions. Such indexing is not a general licence for AI training or commercial republication. Technical accessibility does not itself grant permission.

Request permission for integrations, research or automated access at support@sahibat.com.

9. Enforcement, changes and contact

SAHIBAT may investigate suspected misuse, restrict automated traffic, suspend access in accordance with applicable terms, preserve relevant evidence and seek available legal remedies. Enforcement remains subject to applicable law and mandatory rights.

Changes that materially expand training purposes, eligible information or independent recipients require appropriate notice and fresh permission where required. An updated policy does not retroactively grant that permission.

This policy supplements the Privacy Policy and Content & Intellectual Property Policy . The specific permission here applies only to the described opt-in programme.