Agiloop's AI interview (found in the Invent section) guides you through defining a new feature by having a structured conversation with AI assistants. Rather than filling out a form, you describe your idea in plain language and the AI helps you think through requirements, technical approach, and scope — producing a ready-to-use specification at the end.
Starting a New Feature Interview
To begin defining a feature:
- Open your project and navigate to the Invent section.
- Select the option to start a new feature interview.
- You will be connected to an AI assistant and can begin describing your feature idea in your own words.
There is no required format — start with as much or as little detail as you have. The AI will ask follow-up questions to help you flesh out the idea.
The Two AI Assistants: Business Analyst and Architect
Your interview session includes two distinct AI assistants, each with a different focus:
- AI Business Analyst — Focuses on the user-facing side of your feature. It asks about user goals, workflows, edge cases, acceptance criteria, and business value. This assistant helps you think through what the feature should do and why it matters.
- AI Architect — Focuses on the technical side. It explores implementation approach, system design considerations, integrations, data structures, and potential technical constraints. This assistant helps define how the feature will be built.
Both assistants draw from the same conversation context, so you do not need to repeat yourself when switching between them.
Having a Conversation to Define Requirements
The interview is conversational — respond to questions naturally, add context, and ask for clarification if needed. There is no strict script.
Effective techniques during the interview:
- Describe the problem the feature solves before jumping to the solution.
- Mention any constraints you already know about (deadlines, existing systems it must integrate with, etc.).
- If you are unsure about a detail, say so — the AI can help you explore options rather than requiring a firm answer upfront.
- Push back or refine suggestions if the AI's direction does not match your intent.
The more context you provide, the more accurate and detailed your generated specifications will be.
Pausing and Resuming an Interview
If you need to stop mid-session, you can pause the interview and return to it later. Your conversation history is preserved exactly where you left off.
To resume, return to the Invent section and open your in-progress interview. You can pick up the conversation from where you stopped without losing any context.
This is useful when you need to gather more information, consult a teammate, or simply continue at a later time.
Discarding an Interview to Start Over
If the conversation has gone in the wrong direction or you want to reframe the feature entirely, you can discard the current interview. This clears the conversation and lets you begin fresh.
Note: Discarding an interview is permanent. The conversation history will not be recoverable after you confirm the discard.
Use this option when the requirements have changed significantly or when an earlier misunderstanding has made the existing conversation too difficult to course-correct.
What Gets Generated
Once your interview has covered enough ground, Agiloop generates a complete feature specification package from the conversation. This includes:
- Functional Specification — A structured description of what the feature does, including user stories, acceptance criteria, and behavior details surfaced during the Business Analyst conversation.
- Technical Specification — An architectural overview, implementation approach, and technical considerations drawn from the Architect conversation.
- Work Breakdown — A set of stories and tasks representing the development work required.
- Estimates — Story point estimates for the work breakdown, which can be generated by AI or set manually depending on your project's estimation mode setting.
You can download the specification as a document, or refresh and regenerate individual sections if you want to update them after further conversation.
Viewing the Feature Transcript
After an interview is complete, you can view the full transcript of the conversation. This is helpful for sharing context with teammates, reviewing the reasoning behind specific requirements, or auditing how a decision was reached.
The transcript is tied to the feature and remains accessible from the feature detail view even after specifications have been generated.
Related Articles
Was this article helpful?
That’s Great!
Thank you for your feedback
Sorry! We couldn't be helpful
Thank you for your feedback
Feedback sent
We appreciate your effort and will try to fix the article