New to Elementum? If you prefer a hands-on approach, try our Quick Start Guide first, then return here for deeper understanding.
Processes & Apps
Apps are the containers that hold your business logic and configuration. They organize your data, configuration, automations, and user permissions into unified spaces tailored to specific business processes—like onboarding, incident management, or vendor reviews. Apps work hand-in-hand with Flow, a visual process builder that lets you design your workflow with stages, decision points, automations, and user inputs. Think of Apps as the workspace and Flow as the roadmap: Apps provide the structure and resources, while Flow defines how work moves through your process.What Makes Up an App
Every App brings together several key components:- Data structure — CloudLinks connect to existing data sources, Elements define your business entities, and relationships link everything together.
- Process design — Flow maps your workflow stages, approval processes handle reviews, and assignment rules route work automatically.
- User experience — Layouts control how users interact with data, forms capture input, and views display information clearly.
- Automation engine — Event-driven rules trigger actions, Agents handle complex decisions, and integrations connect external systems.
How Workflows Start
Apps provide multiple ways to initiate workflows:- User-initiated — Service portals, manual creation, and forms let people start workflows when needed.
- System-triggered — API endpoints, scheduled processes, and data mining automatically start workflows based on conditions.
- Smart automation — AI agents and email ingestion intelligently create and route workflows based on content and context.
Design Principles for Apps
- Process-first: Map your business workflow before configuring the App
- User-centered: Design for the people who will use it daily
- Data-driven: Connect to authoritative data sources rather than creating duplicates
- Iterative: Start simple and add complexity as users adopt the workflow
- Integrated: Plan how the App will connect to your existing systems
Data & Elements
Elements are structured data objects that represent the key business entities in your organization—such as vendors, products, incidents, or customers. Each Element consists of defined fields that capture the essential information for that entity. By standardizing how business data is organized and accessed, Elements provide a consistent foundation for your workflows and processes. CloudLinks connect Elementum to your existing data repositories (like Snowflake, BigQuery, or Databricks) to use as a data source without copying or syncing. External data becomes instantly usable inside workflows, unlocking real-time solutions and faster value delivery. CloudLink data can be made available to your workflows as Element objects.What Makes Up an Element
- Fields and properties — Data types (including JSON columns for unstructured data), validation rules, default values, and calculated fields define what information the Element can contain.
- Relationships — Connections between Elements mirror real-world business relationships, like customers linked to orders or incidents linked to products.
- Permissions and access — Role-based controls determine who can view, edit, or delete Element records within your workflows.
- Automation triggers — Event-driven rules automatically respond to Element changes, updates, or state transitions.
How Elements Connect to Data Sources
Elements can be populated and updated from multiple sources:- CloudLinks — External databases like Snowflake that serve as authoritative data sources without copying data.
- Manual entry — Forms and interfaces where users directly create and update Element records.
- AI-assisted processing — AI agents that automatically generate Element records, extract structure from unstructured data, and populate JSON fields from loose data sources.
Element Types and Patterns
Common Element patterns include:- Master data Elements — Core business entities (Customers, Products, Vendors) that other Elements reference. Relatively stable and often sourced from external systems.
- Transactional Elements — High-volume business events and data records (Transaction Logs, Usage Records, Performance Metrics) used for analysis and reporting.
- Configuration Elements — Settings and rules (Approval Thresholds, SLA Definitions, Routing Rules) that govern process behavior and change infrequently.
- Historical Elements — Append-only records (Change Logs, Version History, Activity Records) used for audit trails and compliance.
- Semi-structured Elements — Flexible data with varying structures (Form Responses, API Payloads, Survey Data) that use JSON columns, with AI helping extract structure from unstructured inputs.
Design Principles for Elements
- Business-first: Model real business entities, not technical database tables
- Validation-rich: Include rules that prevent invalid data from entering your workflows
- Relationship-aware: Design connections that reflect actual business relationships
- Evolution-ready: Structure Elements to grow with your business needs
Automations
Automation in Elementum is event-driven. Rules define what should happen when something changes—like a field update, a form submission, or a time-based condition. Triggers start workflows, assign tasks, loop in an AI Agent, or modify records automatically. Use platform APIs and prebuilt automation actions to send or receive data between Elementum and other systems—triggering workflows, updating records, or enriching data.Core Building Blocks
- Triggers — Start automations on record changes, status transitions, comments, file events, or schedules.
- Conditions — Target the right scenarios with filters, comparisons, and branch (if/else) logic.
- Actions — Update fields, create related records, send notifications, request approvals, call webhooks, or invoke AI.
- Schedules and SLAs — Time-based triggers and deadline tracking to escalate or remind when work is overdue.
- Error handling — Built-in retries and failure paths keep processes resilient; log outcomes for audits.
- Observability — View automation runs, success rates, and bottlenecks to improve reliability.
Automation Patterns
Event-driven rules execute immediately when data changes occur. Use them for data validation, field calculations, notifications, and workflow state transitions. Common triggers include field updates, status changes, record creation, and comment additions. Scheduled processing runs at fixed intervals for batch operations, maintenance tasks, and time-based business logic. Schedule options include hourly, daily, weekly, and monthly intervals. Use for SLA monitoring, data reconciliation, periodic reports, and cleanup tasks. Human-in-the-loop workflows combine automated processing with human decision points using assignments, approvals, and conditional routing for complex decisions, compliance requirements, and exception handling. System integrations synchronize data and trigger actions across external systems using webhooks, APIs, and database connections. Always implement retry logic and error handling for external system calls.Design Principles for Automations
- Single responsibility: Each automation should handle one specific business rule or process step
- Idempotency: Design actions to be safely repeatable—multiple executions produce the same result
- Observability: Log execution details, measure success rates, and monitor performance metrics
- Failure resilience: Implement retry policies, error handling, and fallback paths for failed operations
Performance tip: Batch similar operations and use conditional logic to minimize unnecessary executions. Monitor automation frequency to avoid overwhelming system resources.
AI & Intelligence
AI in Elementum provides programmatic reasoning, data extraction, and content generation within your workflows. There are two primary ways to add intelligence:- Agents — Reusable AI components with tools, memory, and policy controls. Ideal for multi-step reasoning and actions within a workflow.
- AI Actions — Lightweight actions powered by LLMs for one-off tasks like categorization, extraction, and transformation.
Common Use Cases
- Triage and routing — Classify incoming requests, detect intent, and assign to the right team with confidence scores.
- Information extraction — Pull structured fields from unstructured content (emails, PDFs, log files) into Elements.
- Summarization and drafting — Generate summaries, replies, or knowledge base entries with human approval steps.
- Search and reasoning — Retrieve relevant context and reason over it to propose next steps or detect anomalies.
- Vendor outreach — Agents that email vendors for missing documents, validate responses, update Element fields, and advance workflows automatically.
- Support triage — Conversational agents that answer FAQs, gather context, triage requests, and escalate with structured handoffs when needed.
Enterprise AI Orchestration
The key to successful AI implementation is embedding non-deterministic AI capabilities within deterministic workflow structures. This approach enables reliable, auditable, and scalable AI deployment. Governance and controls include:- Workflow boundaries — Clear input/output contracts and validation rules that AI actions must respect.
- Security and permissions — Respect roles and data access controls when reading or writing records.
- Transparency — Log prompts, context, and outputs for auditability and improvement.
- Human oversight — Require approvals for high-impact actions; build review queues into your flow.
- Compliance assurance — Built-in audit trails, approval workflows, and policy enforcement for regulatory requirements.
Confidence Scoring
AI actions return confidence scores that determine workflow routing and human oversight requirements:- High confidence (90–100%) — Auto-approve and execute actions
- Medium confidence (70–89%) — Route to human review queue
- Low confidence (0–69%) — Escalate to exception handling or manual processing
User Collaboration
Collaboration in Elementum enables human oversight and decision-making within automated workflows. Configure role-based interfaces, approval chains, and notification systems to maintain accountability while preserving automation benefits.Human-in-the-loop doesn’t mean manual. Use automation to prepare, validate, and route; reserve human judgment for decisions.
Building Blocks
- Layouts and views — Configure tables, timelines, and cards to present the right information for each role.
- Forms — Capture structured inputs with validations, conditionals, and dynamic defaults.
- Approvals — Define approval chains with thresholds, parallel steps, and escalation policies.
- Notifications — Keep stakeholders informed via in-app, email, and chat channels; respect SLAs.
- Roles and permissions — Control who can see and change data with fine-grained access rules.
- Activity and audit — Track history, comments, and changes for compliance and collaboration.
Collaboration Patterns
Assignment and handoffs route work based on user skills, workload, or business priority. Create role-specific views that highlight pending tasks and next actions. Automatic escalation ensures nothing falls through the cracks. Approval workflows support conditional logic—parallel approvals for speed, sequential for hierarchy—with timeout handling and automatic escalation paths. Self-service portals offer intake forms with conditional fields, real-time validation, and status tracking so users can monitor request progress without requiring admin access.Design Principles for Collaboration
- Clarity: Make next actions obvious with clear buttons, status indicators, and task assignments
- Progressive disclosure: Show essential information first, detailed data on demand
- Least privilege: Grant minimal access by default, use approval workflows to elevate permissions when needed
- Feedback loops: Send notifications with specific actions required, not just status updates
How Core Concepts Work Together
Understanding how these concepts connect is key to building effective workflows. Each layer builds on the previous to create sophisticated, automated business processes.Define Your Workflow Structure
Start with Apps to organize your workspace. Then use Flow to design your visual process with stages, decision points, and automations.Example: Create a “Vendor Onboarding” app with a flow that moves vendors through Application, Review, Approval, and Activation stages.
Model Your Business Data
Connect to existing data sources using CloudLinks without copying or syncing. Link data Elements from your CloudLink as structured objects to represent your business entities.Example: Connect to your CRM database via CloudLink and create Elements for Customers, Contracts, and Support Tickets.
Implement Event-Driven Automation
Set up Automations with rules that trigger when conditions are met. Use external system integrations via APIs and prebuilt actions to connect with other platforms.Example: Automatically create a support ticket in your help desk system when a customer submits a complaint form.
Add AI Intelligence
Deploy Agents for complex workflow steps that require reasoning. Use AI Actions for lightweight, single-use tasks like categorization and data transformation.Example: Use an Agent to analyze customer feedback, automatically categorize issues, then route them to the appropriate team.
Enable Human Collaboration
Design human-in-the-loop workflows where people maintain control through roles, approvals, and notifications. Create flexible interfaces with forms, comments, and configurable views for real-time collaboration.Example: Set up approval workflows where managers review and approve expense requests, with automated notifications and deadline tracking.
Next Steps
- Quick Start Guide — Build your first App and see these concepts in action
- Best Practices — Learn design patterns for building effective workflows
- Workflow Design — Explore layouts, automations, approvals, and more
- API Reference — Technical documentation for connecting external systems and building integrations