Creatio.ai FAQ
How does Creatio no-code platform integrate with Creatio.ai capabilities?
Creatio.ai is natively implemented within Creatio no-code platform, thus the platform lets you embed AI-driven capabilities directly into workflows and applications. You can create intelligent processes without writing code. Learn more: Develop Creatio.ai Skill.
How does the token-based pricing model for Creatio.ai work?
Tokens in Creatio.ai represent units of information processed by large language models (LLMs). Each time you send a request to the AI or receive a response, the system counts how many tokens were used. This includes both the input you provide and the output generated by the model. The platform tracks token usage in real time, helping you scale AI adoption while keeping full transparency over costs. Token allocation varies by edition, with the Growth, Enterprise, and Unlimited plans offering 100,000, 200,000, and 300,000 tokens per user, respectively. You can purchase additional token packages on a pay-as-you-go basis so that you can use AI features without interruption. Learn more: Creatio composable pricing. Track token consumption, set thresholds, and manage usage effectively in the AI Command Center.
AI capabilities are included in the base license. Creatio.ai has no hidden fees.
How does Creatio ensure the data privacy of customer data to make sure that data is not shared with third-party LLMs?
Creatio.ai architecture ensures that customer data is not stored by the LLM provider or somewhere beyond Creatio. This means that neither the LLM provider nor Creatio uses customer data for training. Furthermore, Creatio employees do not have access to customer data or their dialogues with Creatio.ai, and Creatio.ai assists with compliance regarding user access rights and restrictions. Learn more: Data privacy in Creatio.ai.
Can I configure Creatio.ai for access by specific people only?
Yes. To do this, grant permissions to the system operation that controls Creatio.ai access:
- For version 8.2.0 and earlier, use "Can run Copilot" ("CanRunCopilot" code)
- For version 8.2.1 and later, use "Can run Creatio.ai" ("CanRunCreatioAI" code)
Learn more: System operation permissions.
Can I edit or modify AI Skills?
Yes, you can edit and create your own AI Skills without any limitations.
What are tokens?
Tokens are units of text that represent chunks of words, characters, or symbols that a language model, like GPT, processes when interpreting input and generating output.
What determines the number of tokens used by Creatio.ai for AI Skill execution?
The number of tokens used depends on the following factors: the length of your dialogue with Creatio.ai, the amount of information passed to the model, the prompt length, the number of executed actions, and all necessary LLM operations needed to generate a response for the user.
How many tokens can be used on a single AI Skill?
View examples of how many tokens an AI Skill can use in the table below.
Small AI Skill | Medium AI Skill | Large AI Skill |
Create vacation request. Handles vacation requests by checking whether it is sick leave, determining vacation dates, verifying available dates, and creating the request if sufficient days are available. The skill confirms the action with the user and checks for satisfaction afterwards. | Meeting preparation. Prepares a user for the meeting by gathering information on the last 5 meetings, generating a plan, confirming details with the user, scheduling the next meeting, and adding past participants if approved. | Search product of the product description. Helps the user to find data by using search parameters, processing the user request, searching the platform, displaying results, and adding confirmed items to cart in specified quantities. |
5800 tokens | 11500 tokens | 19000 tokens |
Why does a single question use so many tokens?
Creatio.ai takes into account factors and context that might be natural for human yet need to be precisely given to LLM itself. For example, if you ask Creatio.ai to create an email draft, Creatio.ai takes the following steps:
- Search for available skills.
- Read the skill prompt and available actions.
- Read data from the page.
- Check system settings to confirm that the message can be sent.
- Generate a draft message and display it in the chat. This is the only visible step.
- Perform additional actions, such as opening a page, etc.
While Steps 1-4 are not visible in they UI, they also consume tokens.
Which table stores information about used tokens?
The CreatioaiRequestEnt database table lets you track the number of tokens spent on a request. However, one chat message does not always equal one request to the LLM.
How is the number of tokens used counted?
As a rule of thumb, one token generally corresponds to approximately 4 characters of a common English text. Learn more about tokenization: Understand tokens (official Microsoft documentation), Tokenizer (official OpenAI documentation).
Can I process system level data using Creatio.ai?
You can work with system-level data such as setting values, current user information, available business processes, etc., using our low-code tools. Learn more: Creatio.ai (developer documentation).
What type of system performance can be expected with Creatio.ai usage? Will there be any additional impact on performance once it begins to be utilized widely?
Creatio.ai has no impact on the Creatio platform or its performance.
Can I use AI Skills in Mobile Creatio?
Yes, however Mobile Creatio has certain limitations:
- The mobile app does not support displaying web UI elements to the user. This includes dialog messages, web pages, and any other UI components that require user interaction with the UI. As such, skills that rely on these elements will not function in the mobile app.
- While business processes can be executed, they must be limited to backend logic only. If a process includes interactive elements, it will not work in the mobile app. However, if the business process consists solely of data processing and requires no user interaction, it will run correctly, and the skill will function as intended.