Listed below are some common issues you may encounter while working with Workiva AI.
Requirements
- Workiva AI is available to any customer who has signed the Workiva AI Terms of Use. Once approved, Workiva AI is enabled by default for the entirety of the organization, including all existing and newly added workspaces and users.
- Org Admins can manage this enablement by restricting Workiva AI to specific workspaces and users or disabling it for the entire organization.
- Workiva AI is fully encrypted and never trained on Workiva content or customer data. Learn more: Workiva's approach to AI
Exceeding the token limit
Tokens are the basic units of text that AI language models use to process content. Depending on the language and model, a token can be a word, part of a word, a punctuation mark, or even a single character.
Each AI model has a token limit, which is the maximum number of tokens it can process at one time. This limit includes everything the model uses to generate a response, such as:
- Your question or prompt
- Any selected documents or files
- Supporting content retrieved by the system
- The AI’s response
Both the input (what you provide) and the output (what the AI generates) count toward this total.
How to work with token limits
Because models can process only a certain amount of content at once, large or complex requests may affect results. For example, if you analyze multiple large documents—such as several SEC filings at the same time—the model may not weigh each document equally within a single response.
To improve results, narrow the scope of your request:
- Select fewer documents
- Reduce the date range
- Focus on a specific section or topic
- Break a large request into smaller steps
Reducing the amount of content helps the model prioritize the most relevant information and generate more focused, detailed responses.