Architecture Overview

Hissuno is built on a 3-layer architecture. The Base Graph captures and connects all product data. The Automation automates processing and enrichment. The Execution layer delivers value through agents and interfaces.

Layer 1: Base Graph

The Base Graph is an interconnected knowledge graph containing five resource types:

Feedback - Customer conversations from the widget, Slack, Intercom, Gong, and other channels
Issues - Bugs, feature requests, and change requests extracted from feedback
Contacts & Companies - The people and organizations behind the feedback
Knowledge Sources - Your codebase, documentation, and reference materials
Product Scopes - Your product ontology of areas and initiatives

These resources are connected through relationships and embeddings. Whenever an entity enters the graph, graph evaluation automatically discovers connections - linking new feedback to the right scope, related issues, relevant knowledge, and the customer who sent it.

Layer 2: Automation

The Automation layer is a set of workflows that run without manual intervention:

Resource Ingestion - When feedback closes, the review workflow classifies it with tags, resolves the customer contact, runs graph evaluation, and decides whether to create or upvote an issue.
Issue Triage - When an issue is created or upvoted, the analysis workflow gathers linked feedback and customer context, scores technical impact and effort, and generates a product spec when thresholds are met.

Layer 3: Execution

Agents and interfaces consume the graph to deliver value:

Support Agent - A customer-facing AI that answers questions using compiled knowledge packages, past conversations, and product context. Deployed via the embeddable widget or Slack.
PM Copilot - A team-facing AI assistant for PMs, founders, and engineers. Available in-app or via Slack.
Interfaces - CLI, Skills, API, and the embeddable Widget.