AI Usage Policy

1. How Loopency Uses AI

Loopency, provided by Jowgik, is AI native. AI agents and the AI assistant answer questions about your business, find what needs attention, and prepare changes. This policy explains what they can do, what data they use, and what stays with people.

Wherever you work with an agent or an AI answer, the Service shows that it is AI. Agents appear by name as agents, and changes they prepare are marked as prepared by an agent.

2. Agents Work Inside Your Permissions

An agent is an authorized actor in your workspace, subject to the same rules as a person:

• It can only read records the role it works for can read, and only the fields that role may see.
• It can run the reports it is given and look up records automatically.
• It prepares changes to your records but does not apply them. A person with the right permission approves each change, one at a time or all together.
• It cannot approve a change it prepared, and it cannot bypass approvals, separation of duties or the audit trail.
• Every run, its steps and what it found are recorded in your workspace.

3. What Data AI Uses

To answer a question or run an agent, Loopency sends the model the instruction and the records needed for that task. It uses your data only to do the task you or your agent asked for.

Personal data is masked before it leaves

Before any text is sent to a model provider, Loopency replaces personal names, email addresses, phone numbers, card numbers, national ID numbers (such as Korean resident registration numbers and US social security numbers) and bank account numbers (IBANs) with placeholders. The real values are restored in the answer and in any action an agent takes, so the work uses them, but the provider does not receive them. Names are recognized in English and Korean text by a language model, so an unusual name, or a name written in another language, can be missed.

No training on your data

Your data is not used to train AI models, by us or by the model provider.

4. Model Providers

Loopency uses a paid API account with OpenAI, reached through a gateway on our own servers, and, for one experimental setup feature that reads your business description during onboarding, TypeSafe AI, reached directly. Before anything is sent to any of them, personal names, email addresses, phone numbers, card numbers, national ID numbers and bank account numbers are replaced with placeholders, so the providers receive placeholders, not those values. Each AI provider is listed in our Subprocessors list, with notice, before it processes any of your data.

5. Agent Memory

Agents learn from the changes you approve and decline, so they prepare better changes next time. What an agent remembers is kept inside your workspace, is never shared with another customer, and can be deleted from the agent's page.

6. Accuracy and Your Decisions

AI output can be wrong, incomplete or out of date. Check it before you rely on it, especially for financial, legal, employment and safety matters. A change you approve is your decision.

Using AI consumes AI credits. Each plan includes a monthly allowance, shown on the pricing page and in your billing settings.

7. Prohibited Uses

In addition to the Acceptable Use Policy, do not use AI in Loopency to:

• Make decisions with legal or similarly significant effects on a person, such as hiring, dismissal or credit, without meaningful human review.
• Generate content that deceives people about whether a human or an AI wrote it, where that matters.
• Try to extract other customers' data, the system's instructions, or data your role cannot see.
• Generate unlawful, harmful or infringing content.

8. Contact

Questions about AI in Loopency: contact@loopency.com