Using the AI Chat to Build Projects

Farline AI's AI chat helps you create and modify projects using natural language. No YAML knowledge required!

How It Works

The AI chat understands project planning concepts and can:

Create new projects from scratch
Add workstreams and work items
Define dependencies between tasks
Create scenarios for comparison
Modify existing projects
Answer questions about your forecast

Getting Started with AI Chat

Creating a New Project

Simply describe your project in plain English:

You: "Create a project for building a mobile app. We have 2 iOS developers and 1 backend developer. The project should take about 12 weeks."

AI: Creates a project with appropriate workstreams, capacity, and initial work items.

Adding Work Items

Tell the AI what needs to be done:

You: "Add a work item for user authentication that will take 3 weeks"

AI: Adds the work item with appropriate size based on your team's capacity.

Setting Up Dependencies

Explain which tasks depend on others:

You: "The UI components can't start until the design is finished"

AI: Sets up the dependency relationship between those work items.

Creating Scenarios

Compare different approaches:

You: "Create an optimistic scenario where we hire 2 more developers"

AI: Duplicates your baseline and adjusts the capacity.

Example Conversation

Here's a complete example of building a project through chat:

You: "I'm planning a website redesign project"

AI: "Great! I'll help you plan that. How many people will be working on this project?"

You: "We have 3 frontend developers and 2 backend developers"

AI: Creates project with two workstreams

You: "The main work items are: design, build new components, migrate content, testing, and deployment. Design should happen first."

AI: Adds 5 work items with appropriate dependencies

You: "How long will this take?"

AI: "Based on the current plan, the project will complete in approximately 8 weeks. Would you like to adjust anything?"

You: "What if we only had 2 frontend developers instead of 3?"

AI: Adjusts capacity and shows new timeline

Tips for Better AI Interactions

Be Specific About Scope

Good: "Add a work item for building the checkout flow, including payment integration and order confirmation. It should take about 2 weeks."

Less Clear: "Add checkout stuff"

Mention Dependencies Clearly

Good: "The deployment work item depends on both testing and documentation being complete"

Less Clear: "Deployment needs some other things done first"

Use Real Numbers When Possible

Good: "We have 5 engineers with a velocity of 30 story points per sprint"

Less Clear: "We have a few people and they're pretty fast"

Ask for Explanations

The AI can explain its decisions:

You: "Why is the project completion date so far out?"

AI: Explains the critical path and bottlenecks

Common Commands

Creating and Modifying

"Create a new project called [name]"
"Add a workstream called [name] with capacity [number]"
"Add a work item for [description]"
"Set the size of [work item] to [number]"
"Make [item A] depend on [item B]"
"Remove the [work item name]"

Scenarios

"Create a scenario called [name]"
"Duplicate the baseline scenario"
"In the [scenario name], change [something]"

Analysis

"When will this project finish?"
"What's the critical path?"
"Which work items can run in parallel?"
"What if we added more people to [workstream]?"
"Show me the total project cost"

Getting Help

"Explain how dependencies work"
"What's a workstream?"
"Show me an example project"
"What can you help me with?"

Multi-Turn Conversations

The AI remembers context within a conversation:

You: "Create a project for a mobile app launch"

AI: Creates project

You: "Add testing work items"

AI: Adds testing to the existing project (remembers we're working on the mobile app)

You: "Now create a faster version"

AI: Creates new scenario with optimizations (still working on the same project)

Editing the Project Definition Directly

You can always switch between chat and direct editing:

Use the chat to create the basic structure
Switch to the project definition editor for fine-tuning
Ask the AI to explain any part of the project definition you don't understand

AI Providers

Farline AI supports multiple AI providers. Each has different strengths:

OpenAI (GPT): Fast and widely used
Anthropic (Claude): Excellent at structured tasks
Google (Gemini): Very cost-effective

Your administrator configures which provider is used. The experience is similar across all providers.

Limitations

The AI chat works best for:

✅ Creating project structures
✅ Adding and modifying work items
✅ Setting up dependencies
✅ Generating scenarios
✅ Answering planning questions

The AI cannot:

❌ Execute the project for you
❌ Make business decisions (which features to build)
❌ Access external data (unless you import from Jira)
❌ Integrate with other tools directly

Troubleshooting

"The AI didn't understand my request"

Try rephrasing with more specific details
Break complex requests into smaller steps
Use the examples in this article as templates

"The AI made incorrect assumptions"

Provide more context upfront
Correct the AI by saying "Actually, [clarification]"
Review the generated project definition to verify

"I want to undo what the AI did"

Ask: "Undo that last change"
Or manually edit the project definition to revert

Getting Started with Farline AI - Overview of core concepts
Understanding Project Definition Structure - Learn to read and edit project definitions
Importing From Jira - Import existing work from Jira

Last updated: 2025-12-19