Back office ops · Production

BKFC: an agentic Python notebook that extracts structured knowledge from Google Chat using Gemini

The problem

Extracting specific information from weeks or months of team chat history is inefficient and error-prone: important context gets buried, action items are forgotten, and valuable knowledge becomes siloed.

Workflow diagram · grounded in source
1
Authenticate and configure
trigger
“enabling the Google Chat and Vertex AI APIs in a GCP project and creating OAuth credentials (specifically for a Desktop app) to allow the script to securely access chat data on your behalf”
2
Fetch active spaces and messages
integration
“It lists all chat spaces the authenticated user has access to. It filters these spaces to only include those active within a defined recent period (e.g., the last 100 days, configurable via SINCE_DAYS). For each relevant space, it fetche…”
3
Sort, group, and concatenate messages
integration
“Messages are sorted chronologically within their respective spaces and threads. They are then grouped by the parent chat space. For each space, the formatted text of the messages is concatenated into a single "page" or document represent…”
4
Gemini structured extraction
ai_action
“a prompt is sent to the Gemini model (gemini-2.0-flash is a great choice for speed and cost-effectiveness here). Crucially, the API call specifies the desired response_mime_type as application/json and provides the response_schema (our C…”
5
Generate Markdown and save jsonlines
output
“Readable Markdown summaries are generated for each chat space, organizing the extracted information under clear headings (Summary, Q&A, Projects, Actions, etc.). The raw structured data is also saved to a jsonlines file, which is ideal f…”
Reported outcome

BKFC transforms conversational noise into actionable insights — producing structured summaries, Q&A pairs, action items, and project updates per chat space — saving considerable time compared to manual review.

Reported metrics
Time saved vs manual reviewSaves considerable time compared to manual review
Reported stack
Google Chat APIVertex AIGeminiGoogle Colab
Source
https://mlops.community/blog/bkfc-an-agentic-workflow-for-gathering-knowledge-from-google-chat
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

BKFC transforms conversational noise into actionable insights — producing structured summaries, Q&A pairs, action items, and project updates per chat space — saving considerable time compared to manual review.

What tools did this team use?

Google Chat API, Vertex AI, Gemini, Google Colab.

What results were reported?

Time saved vs manual review: Saves considerable time compared to manual review (source-reported, not independently verified).

How is this back office ops AI workflow structured?

Authenticate and configure → Fetch active spaces and messages → Sort, group, and concatenate messages → Gemini structured extraction → Generate Markdown and save jsonlines.

Related back office ops cases