Skip to main content

Core Concepts

Understanding Align's data model helps you get the most out of the platform.

Decisions

A Decision is any choice made by your team that affects your product, architecture, process, or organization. Examples:

  • "We'll use GraphQL for the public API"
  • "QA sign-off for release 2.4.1"
  • "Approved: increase Redis cluster to 6 nodes"
  • "Postponing dark mode to Q3"

Decisions aren't just architecture choices. They include approvals, sign-offs, trade-offs, and prioritization calls.

Decision Snapshots

A Decision Snapshot is Align's structured record of a decision. It contains:

FieldDescription
TitleAI-generated summary of the decision
DescriptionFull context and rationale
Decision TypeArchitecture, Approval, Trade-off, etc.
StatusActive, Superseded, Reverted
ParticipantsPeople involved in making the decision
EvidenceLinks to source conversations, tickets, PRs
TagsCategorization for search and filtering
Created AtWhen the decision was captured

Evidence

Evidence links a Decision Snapshot to its source material:

  • Chat messages - The conversation where the decision was made (Slack, Teams, etc.)
  • Jira tickets - Related issues or epics
  • GitHub PRs - Implementation or ADR documents
  • Linear issues - Related tasks
  • Documents - Confluence pages, Google Docs, etc.

Evidence provides the audit trail, so you can always trace back to the original conversation.

Capture Methods

Real-time Capture

Mention @align in your connected communication tools when a decision is made:

@align We decided to migrate to Kubernetes for better
scaling. The trade-off is increased operational complexity.

Discover (Historical Import)

Use Discover to scan your existing tools for past decisions:

  1. Connect a source (Jira, GitHub, etc.)
  2. Align's AI scans for decision-like content
  3. Review and confirm suggested decisions
  4. Import them as Decision Snapshots

Manual Entry

Create decisions directly in the Align UI when needed.

Decision Spaces

A Decision Space is an organizational context that decisions live within. Spaces form a hierarchy that mirrors your org structure:

Organization
-- Engineering
| -- Platform Team
| | -- Q1 Migration Initiative
| -- Product Team Alpha
-- Security

Decisions can belong to multiple spaces. Cross-space relationships (e.g., Platform Team's decision affecting Product Team Alpha) are the highest-value edges in the graph - they reveal hidden dependencies, conflicts, and alignment patterns across your organization.

Spaces have types: organization, department, team, project, and initiative. Hierarchy depth is limited to 4 levels.

See Decision Spaces for the full guide.

Decision Tiers

Decisions are automatically classified into three tiers:

TierScopeExample
StrategicOrg-wide, long-lived"Migrate to microservices"
TacticalTeam-scoped, medium-lived"Use Redis for session cache"
OperationalIndividual, short-lived"Name this endpoint /v2/users"

Classification uses a gradient boosting model trained on structural signals (link count, cross-platform references, participant count, language analysis) - no additional LLM calls. Tiers update automatically as relationships change.

See Decision Intelligence - Tiers for details.

Relationships

Decisions are connected through typed relationships that form the decision graph:

TypeMeaning
relatesGeneral relationship (most common)
supersedesThis decision replaces an older one
conflicts_withContradictory decisions
duplicatesSame decision from different sources
refinesAdds detail or narrows scope
clarifiesResolves ambiguity
supportsReinforces or implements another decision
questionsRaises concerns about a decision

Relationships are detected both deterministically (cross-platform reference matching) and via LLM analysis (semantic similarity).

Organizations & Tenants

Align uses Organizations (also called tenants) to isolate data:

  • Each org has its own decisions, users, and integrations
  • Users can belong to multiple organizations
  • Data is never shared between organizations

User Roles

RolePermissions
AdminFull access, manage integrations, invite users
MemberCreate and edit decisions, view all decisions
ViewerRead-only access to decisions

Decision Lifecycle

┌──────────┐ ┌──────────┐ ┌────────────┐
│ Captured │ ──▶ │ Active │ ──▶ │ Superseded │
└──────────┘ └──────────┘ └────────────┘


┌──────────┐
│ Reverted │
└──────────┘
  • Active - Current and in effect
  • Superseded - Replaced by a newer decision
  • Reverted - Decision was rolled back

Search & Discovery

Find decisions using:

  • Full-text search - Search across titles, descriptions, and evidence
  • Semantic search - Natural language queries matched against decision embeddings
  • Filters - By type, status, tier, space, date range, participants, tags
  • Related decisions - AI-suggested related decisions
  • Timeline view - See decisions over time
  • Relevance feed - Personalized feed based on your spaces and graph proximity

Decision Intelligence

Align provides several analytical views on top of the decision graph:

  • Health Scores - 4-component scoring (freshness, evidence, conflicts, connectivity) per decision and per space
  • Graph Analysis - Clusters, centrality (PageRank), orphan detection
  • Cascade View - Track how strategic decisions flow into tactical implementations
  • Temporal Analysis - Decision velocity, anomaly detection, trend analysis
  • Impact Simulation - Blast radius analysis before changing a decision
  • Space Benchmarking - Compare teams on health, velocity, and conflict metrics
  • Drift Detection - Surface decisions that have gone stale or diverged from intent

See Decision Intelligence for the full guide.