Zenkit AI: Usage and Pricing Model

1. Overview

Zenkit offers AI features for two core areas:

the integrated AI assistant
AI features in workflows and processes

The AI assistant supports users directly within the platform. AI features in workflows and processes can be used as process or node types to analyze content, structure data, or support automated workflows.

AI features are intended as supporting tools. They do not replace professional, legal, organizational, or human review.

2. Organization-Based AI Usage

Zenkit AI is used in the context of an organization.

A user can be a member of multiple organizations. What matters is the organization in which the user is currently working.

This means:

AI features are enabled for organizations.
Permissions and limits apply in the respective organizational context.
AI credits are charged to the organization in whose context the AI feature is used.
Content and data access follow the roles and permissions of the respective organization.

The organization thus retains control over activation, usage, permissions, limits, and costs.

3. Activation and Control of Zenkit AI

Enablement by Zenkit

Zenkit may initially enable AI features for individual organizations, for example as part of a pilot phase, a specific plan, or an individual agreement.

Only when an organization has been technically enabled for AI can an organization admin activate AI features in the admin area.

Later, AI features may also be made generally available to additional organizations.

Activation by the Organization Admin

After enablement, an authorized organization admin can activate AI features for the organization.

In doing so, the admin confirms in particular:

that they are authorized to activate AI features for the organization,
that they accept the AI terms of use,
that they have taken note of information regarding security, data protection, costs, and responsibility,
that AI-generated content and actions should be reviewed,
that AI usage consumes credits and may incur costs,
that automated workflows can repeatedly consume credits.

Activation applies to the respective organization. Users can only use AI features if they are activated in that organization and the respective permissions exist.

First Use by Individual Users

When a user uses Zenkit AI for the first time, they may additionally be asked to confirm a brief usage notice.

This confirmation is user-specific and generally one-time. It is not an activation of AI for an organization, but a personal notice regarding safe usage.

The user confirms in particular that AI-generated results may be incorrect, incomplete, or inappropriate and should be reviewed before important use.

Personal On/Off Toggle

Users can turn AI features on or off for their own usage, for example via the “More” menu or a comparable settings area.

If a user personally turns off AI, this only affects their own usage. AI remains available for the organization and other authorized users, provided it is activated there.

The user can turn AI back on later, provided the organization continues to have AI features activated and the user has the necessary permissions.

Precedence of Organization Settings

Organization settings take precedence over personal user settings.

If an organization deactivates AI features or revokes a user’s permission, the user cannot use AI, even if they have personally turned it on.

4. The AI Assistant

The integrated AI assistant supports users directly within Zenkit.

Depending on permissions and configuration, the AI assistant can, among other things:

create or modify content,
analyze or summarize information,
create or modify processes,
create databases, lists, and structures,
modify fields and configurations,
create, edit, or delete data,
prepare or execute workflow-related actions.

The AI assistant operates within the roles, permissions, and access controls of the respective organization.

The AI assistant is currently optimized in particular for process-related tasks, for example creating, modifying, analyzing, or structuring processes. General or very open-ended requests are possible, but can currently be comparatively cost-intensive and should be used deliberately.

5. AI in Workflows and Processes

AI features can be used within workflows and processes as process or node types.

Typical use cases include:

analysis of content,
processing of external inputs,
generation of text or structured data,
classification of information,
preparation of workflow-based actions,
support of automated processes.

Depending on the configuration, AI-supported process steps can be executed automatically.

For automated workflows, not only consumption per execution is relevant, but also the frequency of execution. An AI node that runs regularly or on many events can accordingly consume credits repeatedly.

We recommend testing AI nodes in workflows initially with limited data volumes, clear test cases, and appropriate limits.

6. Bring Your Own Model in Workflows and Processes

For AI features within workflows and processes, an organization may, depending on configuration, use its own or external AI models or model providers.

This means: In workflows and processes, the organization can determine which model provider or model is used, provided the corresponding integration is available and configured.

In this case, the organization is responsible for selection, configuration, data protection, security, costs, and legally compliant use of the external model provider.

Depending on the provider, different conditions may apply, particularly regarding:

data processing,
storage locations,
use of data for training,
costs,
availability,
regulatory requirements.

Costs, data protection conditions, and billing models of external providers are outside Zenkit credit billing, unless expressly agreed otherwise.

7. Processing External Content and Third-Party Data

Particular caution is required when AI features process external or unverified content.

This includes in particular:

emails,
form inputs,
webhooks,
external documents,
messages,
automatically imported content,
other third-party data.

External content may contain incorrect, misleading, or unexpected information that can affect AI results or automated processes.

We recommend not processing external content automatically without appropriate security, review, and approval mechanisms.

This applies especially when AI results automatically trigger further actions, modify data, send messages, or interact with external systems.

8. Human Review and Approvals

Zenkit supports human review and approval mechanisms.

Depending on configuration, AI-generated content, data changes, or process actions can be reviewed and approved before execution.

Certain automations may, however, also be executed without an additional manual approval step if the organization configures it that way.

For sensitive, business-critical, legal, financial, or personal data processes, we explicitly recommend human review and approval steps.

9. Storage and Visibility of AI Interactions

AI interactions may be stored within the respective organization.

This may include:

inputs to the AI assistant,
responses from the AI assistant,
chat histories,
inputs and outputs of AI features in workflows and processes,
technical execution data,
activity and change histories.

Chats with the AI assistant are stored in the context of the respective organization, but are generally only visible to the user who conducted the respective chat.

AI interactions in workflows and processes may be stored in the respective process, workflow, or activity context and follow the roles, permissions, and access controls applicable there.

10. Transparency, History, and Recovery

Zenkit supports traceable and verifiable workflows.

Changes by users or AI features can remain traceable via activity logs, histories, or state tracking.

Historical states and activity data can help organizations manually trace changes or, where applicable, undo them.

However, complete automatic recovery cannot be guaranteed in all cases. Zenkit does not provide a universal “magic undo button” for all AI-supported actions.

11. Data Processing, Models, and Hosting

The integrated AI assistant is provided via Amazon Bedrock on AWS infrastructure within Europe.

AI models within Amazon Bedrock may be used, for example Claude models from Anthropic.

According to current provider information, inputs and responses of the integrated AI assistant are not used for training or improving the underlying AI models.

Content is technically processed to provide AI features.

Despite high security and data protection standards, residual risks with cloud-based processing can never be completely excluded.

Please continue to treat confidential, personal, or particularly sensitive information with appropriate caution.

For Bring Your Own Model configurations in workflows and processes, the data processing conditions of the external model provider chosen by the organization apply additionally or alternatively.

12. Security and Access Control

Zenkit supports modern security and access control mechanisms for use in organizations and enterprises.

These include, among others:

Single Sign-On,
SAML-based authentication,
SCIM provisioning,
multi-factor authentication,
role-based access control,
organization and permission models,
secure session and authentication mechanisms,
activity and change logs,
configurable approval and review processes.

AI features operate within the roles, permissions, and access controls configured by the organization.

13. What Are AI Credits?

AI credits are an organization’s usage balance for AI features within Zenkit.

Credits are consumed when AI features are executed, for example by the AI assistant or by AI nodes in workflows and processes.

One AI credit corresponds to a usage value of 1 US cent.

AI credits are assigned to the organization account in whose context the AI feature is used.

14. Currency and Local Billing

AI credits are based on US dollars, as most underlying AI models and model providers currently bill their usage costs in US dollars.

Depending on country, contract, payment method, or billing model, prices may be displayed and calculated in a local currency, for example in euros.

The price displayed in the order process, administration area, or contract is authoritative.

15. What Does Credit Consumption Depend On?

Our AI credits are based on the costs of the respective AI models and providers used.

Depending on model, function, and usage, a platform markup may be added.

This markup serves in particular the provision, integration, security, billing, monitoring, scaling, and further development of AI features.

Actual consumption may vary depending on the use case, particularly depending on:

the model used,
the length of the input,
the length of the response,
the amount of data processed,
the complexity of the task,
the use of workflows, tools, or automations.

16. Consumption Guidelines

The AI assistant is currently optimized in particular for process-related tasks, for example creating, modifying, analyzing, or structuring processes.

As a rough guideline, a process-related request can currently range from approximately 20 to 80 credits per call, depending on scope, model, and execution. This corresponds to approximately $0.20 to $0.80 USD.

General or very open-ended requests to the AI assistant are possible, but can currently be comparatively cost-intensive. We therefore recommend using such requests deliberately and initially primarily for testing and pilot purposes.

These figures are non-binding guidelines. Actual consumption may differ.

17. Cost Control Through Limits

Organizations can set limits to manage AI usage and control costs.

In particular, the following are possible:

limits for individual users,
limits for groups,
daily limits,
weekly limits,
monthly limits,
organization-wide credit quotas,
restrictions for workflows and processes.

If no individual limit is set, AI usage may charge the organization’s general credit balance.

Automated workflows can repeatedly consume credits. We therefore recommend setting appropriate limits, especially for AI features in workflows.

18. Purchasing and Managing Credits

Organizations can purchase, top up, or receive AI credits as part of a plan.

When purchasing AI credits, the selected credit package, price, currency, applicable taxes, and payment terms are displayed in the order process.

Credits are assigned to the respective organization account.

AI credits are exclusively for the use of AI features within Zenkit. They do not constitute a means of payment and cannot be paid out or transferred, unless expressly agreed otherwise.

19. Recommended Best Practices

We recommend in particular:

testing AI features initially with clear test cases,
using human review and approval steps for sensitive processes,
granting permissions restrictively,
processing confidential or particularly sensitive information only with appropriate caution,
reviewing AI-generated results before execution or publication,
regularly monitoring automated processes,
not processing external content automatically without security and approval mechanisms,
carefully reviewing external AI models for data protection, security, costs, and compliance,
setting daily, weekly, or monthly limits for users, groups, and automated workflows.

20. Legally Binding Documents

This page is for informational purposes and is intended to explain the use of Zenkit AI in an understandable way.

The following are legally binding in particular:

the AI terms of use,
the general terms of use or EULA,
the data processing agreement, where applicable,
individual contracts or agreements,
the prices and conditions displayed in the order process, administration area, or contract.