Analytics Tab
Dashboard
Memory Graph
Conversation Replay
Replay-Confidence, Summary, ADR
Slack conversion (Agent in action)

Chronicle AI - Never Lose an Important Decision Again

Inspiration

Engineering teams make hundreds of important decisions every month inside Slack. Whether it's choosing an architecture, approving a migration, resolving an incident, or debating technical trade-offs, these discussions are often buried beneath thousands of messages.

Weeks or months later, engineers ask:

Why did we migrate this service?
Why did we reject that approach?
Who made this decision?
What were the trade-offs?

The answers usually require scrolling through endless Slack history - or worse, they are simply lost.

We wanted to solve this problem by giving organizations something they already have for code - version history but for engineering decisions.

That's how Chronicle AI was born.

What it does

Chronicle AI is an AI-powered Slack agent that transforms engineering conversations into structured Decision Replays.

Instead of simply summarizing messages, Chronicle reconstructs the reasoning behind important decisions by identifying:

The problem being solved
Proposed solutions
Supporting and opposing arguments
Final decisions
Trade-offs
Business impact
Confidence score

Each replay becomes part of a searchable organizational memory, allowing teams to quickly revisit historical decisions without manually digging through Slack conversations.

Chronicle also provides:

Interactive Decision Replay player
Searchable decision history
Organizational Memory Graph
Architecture Decision Record (ADR) generation
Analytics dashboard for organizational insights

How we built it

Chronicle AI combines Slack with AI-powered reasoning to automatically capture engineering knowledge.

The system listens to engineering discussions inside Slack, retrieves the relevant conversation context, and uses Google's Gemini models to analyze the discussion and generate structured Decision Replays.

The application consists of:

Slack Bolt for event handling and Slack integration
Express.js backend
React + TypeScript frontend
FastAPI AI services
Gemini API for reasoning and structured decision generation
SQLite for replay persistence
Interactive visualization components for replay playback and organizational memory

The generated replays are immediately available through a modern dashboard where engineers can explore previous architectural decisions and their supporting context.

Challenges we ran into

One of the biggest challenges was designing AI prompts that generated structured, consistent decision records instead of generic summaries.

Another challenge was integrating real-time Slack events with asynchronous AI reasoning while keeping the experience responsive inside Slack.

We also spent significant effort designing an interface that presents complex engineering discussions in a way that is easy to understand for both existing team members and new engineers joining a project.

Accomplishments that we're proud of

Built a complete end-to-end Slack AI workflow
Designed the Decision Replay system from scratch
Generated structured engineering decisions instead of simple summaries
Created an interactive replay interface that visualizes how discussions evolved
Built a searchable organizational memory experience
Developed an enterprise-inspired UI focused on engineering productivity

What we learned

Building Chronicle AI reinforced that AI is most valuable when it preserves organizational knowledge rather than simply answering questions.

We also learned how challenging it is to convert messy, real-world engineering conversations into structured, trustworthy knowledge while keeping the workflow natural inside Slack.