Four Core Capabilities

🗄️
Unified Data Layer
  • Aggregates structured, unstructured & streaming data
  • Backbone of the workflow data fabric
  • Seamless access & contextualisation for humans and AI
  • Eliminates silos across the enterprise
🤖
Agentic AI Enablement
  • Real-time, governed access to enterprise data
  • Powers AI agents beyond simple automation
  • Workflows become adaptive & context-aware
  • Supports continuous, autonomous improvement
⚡
Real-Time, Zero-Copy Access
  • Direct access to external sources — no duplication
  • Improves performance, reduces integration complexity
  • Minimises security risk from unnecessary data movement
  • Workflows & agents always use the freshest data
🔐
Governance & Security
  • Embedded into the architecture from the ground up
  • All data interactions compliant & auditable
  • Protection against unauthorised access
  • Enables confident scaling of automation

Use Cases & Benefits

1
Accelerating enterprise-wide integration by connecting disparate systems more quickly and efficiently
2
Enabling AI-powered process automation by unifying data so agents can act on complex business processes
3
Reducing integration total cost of ownership by simplifying architecture and cutting maintenance overhead
4
Improving compliance and security posture through centralised data management across integrated systems
5
Driving continuous process improvement using real-time insights to identify bottlenecks and optimise
🧬 Foundational Shift: Workflow Data Fabric replaces brittle point-to-point integrations with data virtualisation, federated access, and dynamic data mapping — the prerequisite layer that makes every downstream design and automation choice possible.

Tool Comparison

🌊
Flow Designer
The primary low-code automation builder. Excels at repetitive tasks, approvals, and record operations via an intuitive, visual interface accessible to technical and non-technical users alike.
Best For →
New ITSM / HR / CSM processes Approvals, routing, notifications Event-triggered automation
Upgrade-safe, reusable logic Needs plugins for advanced integration
📋
Decision Tables
Decouples business logic from code, turning multiple conditions into a single outcome. Rules are managed in a structured, business-friendly format rather than hardcoded into flows.
Best For →
Priority calculation (impact + urgency) Assignment / routing rules Approval matrix selection
Non-technical users can edit rules Needs Flow Designer to call it
🦾
RPA Hub
Orchestrates attended and unattended robots to automate manual, repetitive tasks in legacy or external systems where APIs are unavailable or insufficient, handing off between digital and robotic work.
Best For →
Data entry into legacy apps Reading PDF / email into forms End-to-end onboarding steps
Works with non-API systems Needs stable target UI

Designing for Reusability

🧵
Subflows
  • Reusable sequences of actions, inputs & outputs
  • Invoked by flows, other subflows, or scripts
  • Ideal for encapsulating approvals & notifications
  • Design with clear inputs/outputs & documentation
⚙️
Actions
  • Reusable, no-code operations on platform features
  • Encapsulate configuration for consistency
  • Scope appropriately & protect from unwanted edits
  • Design with clear, typed inputs & outputs
🔌
Spokes
  • Scoped apps packaging related actions & subflows
  • Manage integrations & specific tables at scale
  • Activate, update, or retire independently
  • Promote governance through clear boundaries
💡 Workflow Studio ties it together: Flow Designer, Decision Tables, and RPA Hub are all authored and orchestrated inside Workflow Studio — pick the tool per decision point, not per project, and reuse subflows/actions/spokes across all three.

Process Mining vs Task Mining

🔍 Process Mining System-Level
Examines system-level event logs across the end-to-end business process, revealing actual process paths, bottlenecks, and deviations. Gives Process Excellence teams a data-driven map for targeted optimisation and quantifies the "whys behind the KPIs".
🖱️ Task Mining User-Level
Analyses user-level activity — mouse clicks, keystrokes, application usage — to understand how individual tasks are actually executed. Introduced in the Zurich release to complement Process Mining.

Platform Analytics Toolkit

📈
Analytics Center
  • Overview of bookmarked dashboards
  • Central place to ask questions about analytics
🔎
KPI Details
  • Exploratory view of a single indicator
  • Trends, predictions, breakdowns & related records
🔦
Spotlight
  • Ranks records by weighted criteria
  • Surfaces records otherwise overlooked
📡
KPI Signals
  • Flags significant, "special cause" variation
  • Distinguishes abnormal shifts from normal noise

Workflow Analytics: Do's and Don'ts

✅  Decide exactly what you want to learn — focus on actionable insight, not data for data's sake
✅  Tailor metrics to each workflow phase — intake speed early, resolution quality later
✅  Break results down by department, region, or role to expose patterns and anomalies
❌  Tracking dozens of metrics without a clear goal overwhelms teams and dilutes focus
❌  Averages hide outliers — a 3-day "average" can mask 30-day stragglers
❌  A KPI with no link to a business goal becomes a vanity measure, not a driver

Continual Improvement Management (CIM)

1
Identify & prioritise improvements from any source — audits, incidents, problems, surveys, or strategic goals
2
Align with business objectives by linking initiatives to corporate strategy, KPIs, and OKRs
3
Plan & execute using guided templates, tasks, and milestones with assigned responsibilities
4
Track & govern — monitor status, risk, and benefit through dashboards while standardising approval and review
📊 Key Factor: Successful analytics are built into a workflow from the very beginning. Retrofitting after development is complete costs more time and rarely delivers an ideal solution — embed reviews into ongoing operations, not a one-off report at go-live.

Design-Time Elements

📄
Process Definition
  • A trigger + sequence of stages + activities
  • Configured & organised by the process owner
▶️
Trigger
  • Specifies when to start running the process
  • Templates authored by admin / pd_trigger_author
🛤️
Stage (Lane)
  • Logical grouping of activities in the board view
  • Owner sets the start rule for the stage
🧩
Activity Definition
  • Maps Flow Designer inputs/outputs to the activity
  • Defines the automation plan & user-facing view
⏱️
Start Rule
  • When process/stage starts, or
  • After specific stages/activities finish

Roles in Playbooks

👨‍💻
Developer
  • Uses Flow Designer to build flows, actions & activities
🧑‍💼
Process Owner
  • Uses Playbooks to organise pieces into one workflow
🛠️
Workspace Administrator
  • Configures the Playbook Experience views for users
🎧
Agents
  • Work individual tasks via a Configurable Workspace
🙋
End-Users
  • Guided, streamlined intake via Playbooks for Portals
🪄 Playbook Assist: Part of Now Assist for Creator — generate a first-draft playbook from text, an image (even a whiteboard photo), or both, then refine it in Workflow Studio.

Runtime Execution: Six Steps

1
Evaluates any conditions specified in the trigger definition and processes the trigger
2
Processes the event and starts running the process definition in the background
3
Builds the automation plans from each activity into an entire process plan
4
Runs the process plan for the process definition
5
Stores the process execution information in the Process Execution [sys_pd_context] table
6
Supplies data for a user-facing view of the process execution

Example: Manual → Digitized Complaint Process

📥
1 · Intake
  • Agent logs the complaint & notifies the customer
🚦
2 · Triage
  • Case created, type determined, routed to a contact
🔬
3 · Research
  • Assignee investigates & moves case to Work in Progress
💬
4 · Respond
  • Assignee emails the customer as progress is made
✅
5 · Resolve
  • Resolution sent, case state updated to Resolved
🔒
6 · Close
  • Case closed once the customer accepts resolution

Benefits vs Limitations

✅ Benefits
Reuses existing Flow Designer content across processes · Enables seamless, cross-department data flow with no redundant entry · Gives a visual, task-based view that's easy to track · Enforces a standardised record lifecycle with step-by-step guidance
❌  Conditions on lanes are not currently supported
❌  Prebuilt activities may not cover every business need — custom activities required
❌  Execution is tied to Flow Designer logic — a missing/misconfigured subflow limits it
❌  Process execution is asynchronous — sequencing must be designed carefully
Takeaway 01
Unify Before You Automate
Workflow Data Fabric is the prerequisite layer — real-time, governed, zero-copy access to enterprise data is what allows Flow Designer, Decision Tables, and agentic AI to act on trustworthy information.
Takeaway 02
Match the Tool to the Logic, Then Measure It
Flow Designer, Decision Tables, and RPA Hub each solve a different problem. Analytics and process mining must be designed in from day one — retrofitted measurement rarely finds the real bottleneck.
Takeaway 03
Playbooks Turn Automation into a Guided Experience
Playbooks organise Flow Designer content into a single cross-enterprise process — giving process owners governance, agents task-level guidance, and end-users a structured, save-and-resume intake.