Skip to main content

Working with Agents

Agents are AI workers that can autonomously execute tasks in DreamSolve.

What are Agents?​

Agents are specialized AI systems that can:

  • Research and gather information
  • Write content and documentation
  • Analyze data and provide insights
  • Execute defined workflows
  • Interface with external systems

Agent Types​

DreamSolve supports various agent types:

Research Agents​

  • Gather information from multiple sources
  • Synthesize findings into reports
  • Great for market research, competitive analysis

Writing Agents​

  • Create content based on specifications
  • Edit and refine existing text
  • Handle documentation tasks

Analysis Agents​

  • Process data and find patterns
  • Generate insights and recommendations
  • Create visualizations

Custom Agents​

  • Configure agents for specific workflows
  • Define capabilities and constraints
  • Import agent configurations

Assigning Agents to Tasks​

  1. Select a task node
  2. Click Assign Agent in the task actions
  3. Choose an appropriate agent type
  4. Configure any agent-specific settings
  5. Start the agent

Monitoring Agent Work​

Once an agent starts working:

  • Status Indicator: Shows if agent is idle, working, or complete
  • Progress Updates: Real-time updates on what the agent is doing
  • Output Preview: See generated content as it's created

Reviewing Agent Output​

Before agent work becomes final:

  1. Review the generated content
  2. Make any necessary edits
  3. Approve or request revisions
  4. Mark the task as complete

Workflows​

Agents can be chained together into workflows - multi-step sequences where each agent builds on the previous agent's output. For example, the Product Analysis workflow chains ideation, market analysis, and pitch creation into a single execution.

Learn more in the Workflows feature guide.

Best Practices​

  • Be Specific: Clear task descriptions lead to better results
  • Provide Context: Attach relevant files and background
  • Start Small: Test agents on smaller tasks first
  • Review Carefully: Always verify agent output before approving
  • Iterate: Use feedback to improve future agent work