Morrow: A Company-Memory Concept for AI-Native Teams
A design-led MVP concept for giving AI agents company memory they can trust. Morrow pulls Slack, Linear, GitHub, Notion, calls, and tickets into one reviewed context layer, turns it into executable skill files, and serves it to agents through an API. Every answer carries the evidence it came from.
Give agents context they can cite.
The Challenge
Companies don't have a data shortage. They have a context-trust problem. Decisions live in Slack threads, calls, tickets, docs, and pull requests. Search returns fragments, docs drift from how teams actually work, and agents either keep asking people or answer from sources that contradict each other. The concept set out to solve one thing: how do you give agents company memory that has been reviewed and kept current, so they can act on it without a person checking every response?
Our Solution
We designed and built Morrow as a full product story across two responsive screens. The landing page lays out the problem and the workflow behind it: ingest, resolve, approve, publish, and learn. The operator dashboard is where the work happens. It shows live memory-health metrics, source-connector status, a contradiction queue, conflict resolution backed by evidence, a human approval gate, generated skill files, and an audit trail. Conflicting sources turn into review tasks, and nothing reaches an agent until an owner signs off. Approved context then goes to agents through guarded API endpoints that expose reviewed memory and skill files only, never the open queue. The interface uses a constructed, lab-instrument look with animated context-graph visualizations, built in Next.js 16.
