With d.velop pilot, you can already use AI today to analyze document content and search for information. With agent-based skills, we’re further expanding the possibilities of d.velop pilot: AI no longer just helps you answer individual questions, but can also handle more complex tasks based on your information for processing.
Agents can search through document collections in the d.velop platform, evaluate document content, and use this information for processing of your request. In addition, you can integrate agents into d.velop process studio and thus use them as part of structured business processes.
From a Simple Question to a Complex Task
Until now, working with AI has often centered on a simple question, such as:
“What is in this document?”
Instead, an agent can autonomously derive and process a more complex task, such as:
“Review this contract, including its attachments, based on this list of criteria, and create a summary of the differences.”
To do this, the agent can combine multiple steps, research relevant information, and analyze document content. The classic pattern Document → Question → Answer is thus increasingly shifting to Task → Research → Analysis → Result.
Expert Knowledge for Your Agents
With skills, you can provide agents with additional expert knowledge and specific work instructions for recurring tasks.
A skill functions similarly to an onboarding guide for new employees: It describes how a specific task should be completed. This may include review criteria, standards and regulations, references to relevant documents, or desired output formats. This allows agents to be better aligned with your company’s requirements and workflows.
Reusing Results with d.velop pages
An AI analysis doesn’t have to end in the chat. Results can be reused, for example, as a structured summary or review note. In combination with d.velop pages, this can generate content that can then be further edited and utilized within d.velop documents.
Using Agents in d.velop process studio
You can also integrate agents into d.velop process studio and use them as part of structured business processes. Using this integration requires separate licensing for d.velop process studio.
Example
A new case is received in a service process. An agent analyzes the available information and searches for relevant document content. The result can then be used in the subsequent stages of the process. This allows agent skills to be utilized not only interactively in chat but also within digital business processes.
Availability
d.velop pilot is part of d.velop documents platform edition business and ultimate. Current base quotas are available to respective users who are licensed. The previous token-based allocation system will be replaced by d.velop AI Credits upon acceptance of the supplemental agreement.
The new d.velop pilot features are deployed in both cloud and hybrid deployment models. Currently, the agents are available only to AWS and Azure customers. A deployment for T Cloud Public is in the works.
Billing via d.velop AI Credits
Using d.velop pilot, including its agent skills, consumes d.velop AI Credits. Actual consumption depends on the functions executed and the processing operations performed.
Credits are used, for example, for processing messages and document content, the agent-based execution of tasks, the generation of output, and processing operations via d.velop process studio or MCP.
Available credits are accumulated within a tenant for the respective calendar month. Unused credits expire at the end of the month. If the available quota is exceeded, the additional usage is billed as an overage at the applicable rates. Additional d.velop AI Credits can be purchased in packages.
How can I view my specific credit usage?
To estimate the credit usage of a task, we currently recommend a before-and-after comparison in d.velop consumption tracker. At the same time, we’re working on making usage even more detailed in the future. You are welcome to give us feedback on what insights would be particularly helpful to you.
d.velop consumption tracker is accessible via the configuration.
Important: Confirm the Supplemental Agreement
Before you can use agent skills in d.velop pilot, you must confirm the supplemental agreement for using the new features.
The agreement specifically governs the expanded feature set, the transition from tokens to d.velop AI Credits, and usage-based billing.
Without confirming the supplemental agreement, you will not have access to the agent skills. In this case, your existing contractual agreements regarding the use of d.velop pilot remain unchanged.
Enable Agents
To use the d.velop pilot agents, you must first enable them.
This is how it works
- On the d.velop documents start page, open the Configuration section.
- Under d.velop pilot, select the option Activate d.velop pilot agents.
- Review and confirm the contract documents displayed.
Set limits for d.velop AI Credits
We recommend that you also set the desired usage limits for your tenant as part of the activation process. This allows you to control the usage of d.velop AI Credits according to your requirements.
Where can you use d.velop pilot agents?
Agent-based support is particularly well-suited for recurring tasks that involve searching for information, analyzing documents, applying technical guidelines, and generating results based on them.
Possible examples include contract reviews, research related to service cases, quality checks, or consolidating information from multiple documents and dossiers.
Existing checklists, work instructions, or review templates are therefore a good starting point for your first use case.
From “What’s in this document?” to “Help me with this task”
With d.velop pilot agents, AI support within the d.velop work environment continues to evolve.
From:
“What’s in this document?”
to:
“Help me finish this task based on our information.”
In this way, d.velop pilot combines the existing information in d.velop documents with agent-based task processing—from interactive work with documents to the integration of agents into digital business processes.