Interview automation agent

Client: Personal project
Type of project: AI
Role: AI Creative
Tool: N8N
Year: 2024-2025

The challenge – from raw text to usable insights
Interviews generate valuable knowledge, but the process of cleaning, analyzing, and presenting findings is often slow and repetitive. The challenge was to see if a sequence of AI agents could collaborate like a small team, turning transcripts into structured insights and clear presentations.

Solution
I built a chain of agents that worked step by step. The first agent cleaned raw transcripts, removing filler words and errors. The second analyzed the material, drawing conclusions and highlighting patterns, trends, and correlations. The third applied thematic analysis and coding to structure the findings, and the final agent created a presentation that summarized key insights with clarity and narrative flow.

Outcome
The interview automation agent showed how multi-step workflows can transform unstructured data into useful, shareable knowledge. It demonstrated that agents can act less like isolated tools and more like colleagues who contribute different skills. The prototype suggested new possibilities for scaling research and making insights more accessible.

The diagram illustrates the interview agent workflow. A raw interview transcript first passes through a cleaning agent, producing a clean version. An analyzing agent processes this into an analyzed document. The content then flows into a theming agent, which extracts key themes, and finally into a presentation agent, which produces a finished presentation. Each stage builds on the previous, simulating collaboration across a small team of agents.

Contact

Email: pierre@bremell.com
Phone: +46 709353902
LinkedIn: bremell