Org Prompt Configuration tunes WorkBoard's AI agents — the Chief of Staff Agent and the Leadership Coach Agent — to reflect your organization's tone, priorities, values, and operating boundaries. The guidance you provide is considered by the agents in every interaction, helping ensure responses stay aligned with how your organization operates.
Example Scenario
An organization wants to ensure the Chief of Staff Agent does not produce performance judgments about individuals.
Configuration: Specify that the agent must not assess, rank, or evaluate individuals based on their work, and must redirect performance-adjacent questions toward factual OKR and Action Item context.
Result: Once applied and tested, the agent consistently:
Declines performance-based assessments
Notes that available information should not be used for performance evaluation
Offers factual updates on OKR progress and Action Item ownership instead
This pattern can align agent behavior with internal policies, governance considerations, or organizational principles.
How Tuning Works
WorkBoard support implements the configuration in partnership with your team. Customers identify the guidance they want the agents to reflect; WorkBoard support applies it to your organization's instance.
This collaborative approach ensures:
Prompts are structured for the strongest possible effect
Configuration changes are tested before they affect end users
Tradeoffs and potential impacts on agent performance are reviewed up front
To update or refine your configuration, contact your WorkBoard team.
Things to Consider Before Tuning
Org Prompt Configuration is powerful, and changes should not be made lightly.
Prompts affect agent performance broadly. Changes influence how the agents respond across all interactions, not just the scenarios you have in mind.
Response variation is expected. Generative AI is probabilistic. Adding prompts may introduce variation in areas you did not intend to change.
Testing is essential. Every prompt change requires additional testing to confirm the agents behave as intended. WorkBoard partners with your team on this testing.
Less is more. Concise, principle-level guidance is more effective than long passages of detailed content. The field limits reinforce this.
What You Can Configure
Each field has a defined limit to keep guidance focused and high-signal.
Company Context — High-level description of your organization to give the agents grounding context. Limit: 800 characters.
Company Values — Values and cultural principles the agents should reinforce. Limit: 800 characters.
Additional Guardrails & Boundaries — What the agents should not do, when to remain silent and redirect to humans, and how to phrase uncertainty or risk. Limit: 800 characters.
Company Priorities — Top company goals and strategic themes the agents should reinforce when helping people set aligned OKRs, scan for risk, and plan their time. Limit: 800 characters.
Additional Instructions — Custom instruction categories for specific scenarios or workflows. Each category has a name and an instructions field. Limit: 80 words per category. There can be multiple categories.
Best Use Cases
Org Prompt Configuration is a tuning layer, not a knowledge base. It works best for:
High-level company focus areas the agents should reinforce
Company values the agents should reflect
Guardrails or boundaries on what the agents should not do
For detailed reference content or document-level knowledge, continue to use your existing knowledge sources.
Getting Started
Contact your WorkBoard team to explore Org Prompt Configuration for your organization. They will help you:
Identify the right starting points
Draft the configuration
Coordinate testing before changes go live