Case study / Automation architecture · AI reporting
AI Reporting Engine: from media to a report
A multi-route n8n workflow for extracting or transcribing content, preparing an AI prompt and routing a report to email, Telegram and Google Drive.
The problem
The workflow starts with files in Google Drive. An existing transcript and a video need different preparation before either can be used to generate a report.
Those input paths need to converge on a shared prompt and report-generation stage, then branch to the destinations where the report will be used.
What I built
- 01
Route the source file
The saved workflow shows a Google Drive trigger followed by file-type routing. The transcript branch validates, downloads and extracts text; the video branch prepares and downloads the video, then transcribes it with Whisper and formats the transcription.
- 02
Combine the material for the prompt
Both branches feed a merge step. The workflow then fetches a skills database from Google Sheets, combines the data and prepares the AI prompt.
- 03
Generate and deliver the report
An AI agent connected to an OpenRouter model feeds a report-generation step. The output branches are labelled Send Email, Send Telegram and Save to Drive.
Implementation evidence
- The workflow screenshot shows separate transcript and video preparation paths leading to one reporting sequence.
- Email, Telegram and Google Drive are represented as distinct output branches.
- This case study documents the workflow architecture. Measured time savings, delivery success rates and production performance are not included.
Have a similar problem?
Let’s make
something work.
Automation & AI projects start at J$60,000 per workflow. Tell me what you’re dealing with and you’ll get a clear proposal within a few days.
More case studies