Build AI context in awork Docs
Organize your agency, project, and customer knowledge in Docs Spaces to provide your AI Agent with accurate, relevant context across your workspace
Want the awork Agent to work with your project or customer knowledge? Start with the sources you already maintain. Use awork Docs for knowledge related to a customer, a project, or a topic. If the authoritative source is in another tool, you can leave it there, as long as the Agent can actually access it through a suitable connector. Check this for the specific file and with the access permissions of the person using the Agent. For knowledge that applies across the entire workspace, use Workspace-Kontext.
Key benefits and use cases
Good AI results start with the right context. Your Agents need information about your agency, your customers, and the way you work.
The central place for this is the Docs section. This is where you build the knowledge structure your Agents work with.
Agents also use the context they're running in:
-
Project and task
-
Mentioned documents
-
Mentioned customers
The better maintained your Docs and projects are, the better the results. Create Spaces with the relevant information.
You don't need to include knowledge documented in your Docs in every prompt.
Set up agency context
We recommend organizing your agency information into clearly separated areas. This keeps the context easy to navigate and prevents individual Docs from becoming unnecessarily large.
For agency context, we recommend creating a Space with the following areas, each as a separate Doc:
-
Profile and services
-
Ways of working and methods
-
Rate cards and pricing
-
Internal policies
-
Corporate design and tone
1. Profile and services
Briefly describe what your agency stands for and what you offer.
For example, include:
-
Agency profile: Industry, focus, and relevant customers
-
Services: Your core offerings
-
Team structure: Departments, disciplines, or other relevant structures
The information should be clear enough to help a new team member quickly get up to speed.
2. Ways of working and methods
Document your standard processes and recurring ways of working.
Here's an example of a standard project workflow:
-
Brief received
-
Brief review
-
Rebriefing call with the customer
-
Concept and estimate
-
Design and copy
-
Internal quality assurance
-
Written customer approval
-
Production, final artwork, or print-ready files
Also document any binding principles for your projects here.
Examples from the setup:
-
One round of revisions is included as standard
-
Approvals are always given in writing
-
A legal review is carried out before any print-ready files are produced
3. Rate cards and pricing
Add your standard hourly rates.
You can organize these by discipline, seniority level, or your internal structure.
Important: Customer-specific exceptions belong in the relevant customer profile and take precedence.
4. Internal policies
Document rules that are important to your day-to-day collaboration.
These may include:
-
File storage and naming
-
Handling confidential data
-
Use of AI
-
Customer communication
Tip: Imagine you're onboarding a new team member. Anything they need to know that isn't self-evident should be documented here.
5. Corporate design and tone
Add the information your AI needs to create content and communicate.
This includes:
-
Brand idea
-
Logo and logo guidelines
-
Primary and accent colors
-
Typography
-
Tone of voice
This is especially useful when your AI drafts emails or puts together cost estimates, for example.
Set up customer context
Create a Space for customer context and add a separate area as a Doc for each customer, with its own nested Docs. Try to use the same structure for every customer.
Recommended Docs for each customer:
-
Brand and tone of voice
-
Stakeholders and approval process
-
Terms and special conditions
-
Personas
1. Brand and tone of voice
Add the most important information about the brand.
This includes:
-
Brand essence
-
Tone of voice
-
Visual identity
-
Legal and regulatory requirements
2. Stakeholders and approval process
Document all relevant contacts and their responsibilities.
Include:
-
Names
-
Roles
-
Responsibilities
-
Approval processes
-
Response times
This makes it clear who can approve which content.
3. Terms and special conditions
Add individual agreements that differ from your standard terms.
These may include:
-
Adjusted hourly rates
-
Retainer terms
-
Monthly flat fees
-
Special collaboration rules
-
Periods with particular priorities
Examples:
-
No visuals featuring competitors' products
-
Spring and Christmas seasonal campaigns take priority
-
Use an informal tone when communicating
4. Personas
Document your customers' personas in awork Docs.
This lets both your team and your AI access the same information.
Add project-specific context
Projects can also include information for Agents. For example, a project can have a Doc called Für Agenten.
It can include:
-
Who is responsible for the project
-
How tasks should be handled
-
When the project is considered critical or unhealthy
-
Which rules apply specifically to the project
Example of a project rule: A project is considered unhealthy if more than 20 hours have been logged in the current calendar month or a specific task list contains more than five tasks.
The Agent can use these project rules together with the Workspace Context.
Create workspace context
With awork Docs, you can build a structured knowledge base for your AI containing all the important details about your agency, customers, methods, and specific policies.
Workspace-Kontext (Einstellungen > Workspace-Kontext) provides a shared foundation for the entire workspace. Add the most important information about your organization, your ways of working, and your general rules here:
- who we are
- how we work
- which terms we use
- which general rules and standards apply
- which roles and responsibilities we have
This information is available to all AI Agents as shared context.
You can organize and maintain more detailed, specific information in your Docs.
Users with access to workspace settings and AI permissions can access Workspace-Kontext in the settings.
👉 Learn more about setting up Workspace-Kontext in awork: Workspace-Kontext für AI.
Best practices for AI context
- Don't build all your context at once. Start with the information that matters most to your agency and customers.
- Keep general agency knowledge clearly separate from customer-specific information.
- Document rules, roles, and ways of working so they're understandable to a new team member.
- The more relevant knowledge you've already documented in awork Docs, the less you'll need to explain repeatedly in prompts.
Test your setup with a real task
✅ Choose a typical question, such as asking for the currently approved brief.
✅ Start in the relevant project or add the relevant Doc directly as context. If you're using an external source, check whether the Agent can read the specific file.
✅ Check that the Agent is using the current source and that the people who need to use the setup have access to it.
✅ Keep only knowledge intended for the entire workspace in Workspace-Kontext.
FAQ
What information belongs in agency context?
This includes, for example, your profile and services, ways of working and methods, rate cards and pricing, internal policies, and corporate design and tone.
What information belongs in customer context?
For each customer, we recommend adding information about their brand and tone of voice, stakeholders and approval processes, terms and special conditions, and personas.
Do I need to include information in every prompt?
No. You don't need to include information that's already in your documents in every prompt.
What additional information do Agents use?
In addition to Docs, Agents use the context they're running in. This includes the project and task, mentioned documents, and mentioned customers.
Can an Agent remember information permanently?
Only if you explicitly confirm it.
How detailed should the context be?
Start with the basic information and expand your context step by step. Focus on the information, rules, and ways of working that matter in your day-to-day work.


