Built for teams that need to remember
Distributed software teams, agencies, and consulting firms make hundreds of decisions every week. Those decisions scatter across Slack threads, Zoom calls, email chains, and Jira tickets — then disappear into silos where no one can find them again.
The result: teams waste hours searching for why a decision was made, who approved it, and what objections were raised. New employees ask the same questions. Settled decisions get unknowingly reversed. Institutional knowledge walks out the door when someone leaves.
DecisionTrace was built to solve this. It is a cross-tool decision memory system that uses AI to extract, link, and retrieve institutional knowledge from unstructured conversations. Connect your tools with OAuth, and the system starts indexing in the background — no training, no tagging, no manual curation.
When someone asks “Why did we switch from Stripe to Adyen?”, DecisionTrace returns the Zoom meeting where it was discussed, lists who approved it, surfaces objections about integration complexity, and links the Jira ticket tracking the migration. Full context, in seconds.
Our mission
Eliminate repeated conversations, accelerate onboarding, and prevent teams from unknowingly reversing settled decisions. We believe every decision deserves a retrievable record — linked to the people and context that shaped it.
What guides us
Precise
Every result returns only the decisions, participants, and artifacts that directly answer the query. No noise, no irrelevant matches.
Grounded
All answers trace back to original conversations, meetings, or documents with clickable source links. We report what was recorded, not what we interpret.
Unobtrusive
The system works in the background, indexing continuously without requiring manual tagging or curation. Your team keeps working; context builds itself.
Trustworthy
Results respect permissions and never surface decisions a user is not authorized to see. When context is incomplete, the system says so clearly.
Design principles
Index, don't interrupt
Ingestion happens silently. Users never tag, categorize, or train the system during daily work.
Query where you already are
Slack and Teams bots answer questions inline. The web app is for deep exploration, not the only entry point.
Show the seams
Every answer links to original sources and lists participants so users can verify or dig deeper without guessing.
Scope before you search
Filters for project, date range, and stakeholder appear at the top of every view, not buried in settings.
Fail forward with clarity
When context is incomplete or ambiguous, the system says so and offers related queries instead of pretending certainty.
Ready to build your decision memory?
Connect your tools and start surfacing past decisions with full context in seconds.