How to Create a New Project in Agiloop

Starting a new project in Agiloop is guided by an AI-powered interview that helps you define your product clearly before writing a single line of code. This article walks you through creating a project, working with the AI assistants, and what to expect once your project is ready.

Before You Begin

Make sure you are signed in to your Agiloop account. You will need sufficient credits if you plan to use AI-powered build features later, though no credits are required just to create a project.

Step 1: Start a New Project

1. Open the project menu from the main navigation.
2. Click Add Project. and select the desired workflow (interview or repo analysis for existing projects)

Step 2: The AI-Powered Interview

After choosing the Interview process, Agiloop launches the Invent interview — a conversational session designed to help you articulate what you are building, who it is for, and how it should work.

You are joined by two AI assistants:

• AI Business Analyst — focuses on your product goals, user needs, workflows, and functional requirements.
• AI Architect — explores the technical side: integrations, data models, infrastructure considerations, and constraints.

The interview unfolds as a natural conversation. The assistants ask questions one at a time, building on your answers to develop a thorough understanding of your product. You do not need to have everything figured out in advance — the interview is designed to surface ideas and fill in gaps as you go.

Note: There are no wrong answers. The more detail you provide, the richer your generated specs and features will be. If you are unsure about something, say so — the assistants can help you think it through.

What to Expect During the Interview

• Questions cover your target users, core use cases, key workflows, and any technical preferences or constraints.
• You can ask the assistants clarifying questions or push back on their suggestions.
• The conversation typically takes 10–30 minutes depending on the complexity of your product.

Pausing and Resuming an Interview

You do not have to complete the interview in one sitting.

• To pause, simply navigate away from the interview. Your progress is saved automatically.
• To resume, open your project from the project menu and return to the Invent section. The conversation picks up exactly where you left off.

Note: You can return to an interview and continue adding detail at any point before generating your specs. The more context you provide, the more accurate and complete your output will be.

Step 3: Generating Your Project Specs and Features

When you are satisfied with the conversation, you can instruct the AI to generate your project output. Agiloop produces:

• Functional Specification — a structured description of what your product does, covering user goals, features, and workflows.
• Technical Specification — an overview of the recommended technical approach, architecture, and implementation considerations.
• Initial Feature List — a set of features derived from the interview, each ready to be refined, prioritized, and eventually built.

These assets give your team a shared foundation to work from, whether you are planning sprints, briefing engineers, or scoping an MVP.

Note: You can refresh or regenerate individual specs later if your product direction changes. Agiloop also supports importing projects from GitHub or GitLab if you are starting from an existing codebase.

After Your Project Is Created

Once your project is set up and your initial features are generated, you have several natural next steps:

• Review and refine features — Open the feature list to review what was generated, reorder items, update statuses, or add new features manually.
• Dive into specifications — Each feature has its own functional spec, technical spec, and work breakdown with story-point estimates.
• Invite your team — Go to project settings to invite team members by email or link, assigning roles such as admin, maintainer, or member.
• Configure your project — Set your estimation mode (AI or human), adjust project details, and connect any integrations for exporting to tools like Jira, Trello, or Azure DevOps.
• Start building — When a feature is ready, use the Implement section to kick off AI-powered code generation

Related Articles

• Understanding the Agiloop Interview

• How to Define Features Using Agiloop's AI Interview

• Project Settings: A Complete Guide

• Managing Team Members on a Project

• What Is a Feature Bundle