Skip to main content
All CollectionsGeekbot AI features
Ask Geekbot - Opt-in Documentation
Ask Geekbot - Opt-in Documentation

A comprehensive breakdown of the benefits of Ask Geekbot, how it works, and relevant documentation resources.

Tasos Efkarpidis avatar
Written by Tasos Efkarpidis
Updated this week

Introduction

Ask Geekbot is an advanced analytics and insights tool that leverages artificial intelligence to help teams better understand their work patterns, progress, and dynamics through natural language queries. By analyzing your team's regular stand-up reports and other check-ins, Ask Geekbot provides valuable insights that would otherwise require manual analysis of hundreds of messages.

Core Concepts and Terminology

  • Tasks: Specific work items or activities reported by team members

  • Topics: Broader themes or areas of work

  • Blockers: Issues impeding progress

  • Sentiment: User or team morale and engagement

  • Meetings: Automatically identifies meeting references

Purpose and Benefits

  • Extract meaningful insights from daily standups and team communications

  • Save time by quickly accessing historical information

  • Identify patterns and trends in team activities

  • Track project progress and blockers effectively

  • Understand team sentiment and wellbeing

Key Features

  • Natural language query processing

  • Historical data analysis

  • Cross-team insights

  • Sentiment analysis

  • Blocker tracking

  • Time allocation analysis

Target Users

  • Team Leaders and Managers

  • Project Managers

  • Scrum Masters

  • Team Members

  • HR and People Operations

How Ask Geekbot Works

Data Collection and Processing

Ask Geekbot processes data from your existing Geekbot reports, including:

  • Daily standup responses

  • Sprint retrospectives

  • Team check-ins

  • Custom surveys

The system uses advanced natural language processing to:

  1. Analyze report content

  2. Identify key entities (people, projects, tasks)

  3. Extract meaningful insights

  4. Understand context and relationships

In-house models

While Ask Geekbot appears as a single conversational experience, it stands on top of a vast knowledge base consisting of data stored in both relational and graph databases. There is a number of in-house proprietary machine learning models and optimization algorithms that we have developed and use to process the raw data generated by users reports. These models deliver the knowledge needed for the bot to intelligently answer any type of questions related to a team's work.

Sending and getting data to/from OpenAI

We use OpenAI's API as it is hosted on Azure, which gives us seamless integration with other Azure services, alongside enterprise-grade security and full compliance. Given that it sits on Azure cloud, the data send remains within our control and is not used for model training by OpenAI.

Access Control and Scopes

Currently, a user only has access to the data (and any derived knowledge) from reports in the standups of which they are a member. This means users can only query and see another member's work if they participate in the same standups.

Example Queries

Team Level:

- "What did the frontend team work on last sprint?" -> task query

- "Show all blockers reported by the design team in Q1" -> blocker query

- "What's the current team sentiment trend?" -> sentiment query

Individual Level:

- "What has [user] been working on this month?" -> task query

- "Show [user]'s completed tasks from last week" -> task query

- "List all blockers reported by [user]" -> blocker query

Time-Based:

- "Show all completed projects in Q1" -> task query

- "What were the main blockers discussed last month?" -> blocker query

- "Compare team sentiment between January and February" -> sentiment query

💡 Tip: For a detailed guide on how to use Ask Geekbot, check it out here ➡️

Legal and Compliance Documents

Resources

Since Ask Geekbot is still in Beta, please contact our support team to have it enabled.✨

Did this answer your question?