Horizon — the operator's home
Horizon brings together Home, the published app catalog and operational flows. Open it from the product switcher.
Horizon is Scrydon's Common Operating Picture (COP): a persistent, map-centric view of everything happening across your operation, right now. Detections appear as they arrive, tracks update live, and every dot on the map is backed by governed data — classified, masked, and clearance-filtered per viewer.
The Horizon app
Horizon ships as its own app next to Cortex, Agentic and Analytics, and it is on wherever the platform is deployed. Open it from the product switcher in the sidebar of any app, or from the Horizon card on the Platform Hub. Its sidebar is the platform sidebar: product switcher, workspace switcher, the Horizon sections, notifications, settings and account.
| Section | What it is today |
|---|---|
| Home | Your personal landing page: a greeting and Needs you — every open task across your flows that waits on you, overdue first, each one a link straight to that task. It reads your tasks, so it is exactly what a task assignment or an overdue notice in your inbox points at. |
| Flows | The process flows your organisation runs, operated from here: Needs me lists the flows with a task waiting on you, every row carries its stage strip, and the instance page is the full operator surface — wizard, kanban and list views, every task type, persona assignment and the knowledge base. A flow whose persona nobody holds says so in its header and offers to assign it. |
The Apps catalog lists discoverable published notebook releases with workspace-scoped search and audience filters. App runtime viewing, reports, obligations and the calendar are separate delivery work.
Flows opens as a list with the process name and template, status, ordered stage progress, assigned people, and the instance's last update time. People includes persona holders and directly assigned task owners; avatars show names on hover, with a count for additional people. Switch to Cards in the toolbar to use the card visualization without losing your filter or search.
The status tabs show counts, including Not started. Needs me opens first when you have work waiting; otherwise the page opens on All. Click a row or its process name to open the flow. The row menu retains confirmed deletion, and Funnel remains available for a template's stage conversion metrics.
Horizon is the only place process flows are operated. Agentic keeps the builder — workflows, automations, knowledge, extensions — and no longer has a Process Flows section; a link to a flow that was sent before this change (a distribution email, a receipt page, a bookmark) still opens the flow, in Horizon. Deployments that switch Horizon off keep the runtime and the builder but have no operator surface for flows.
The tactical picture is feature-flagged. The live map, feeds and digital
twins described below are the one part of Horizon that stays off by default
(HORIZON_ENABLED, enabled per deployment by your platform operator). The
Horizon app itself, Home and Flows are always on, and everything process
flows depend on — including map actions inside a running flow — works
regardless of the flag.
Execution vs. awareness
Horizon is a product; the rest of the platform is the engine underneath it.
| Agentic — the "Engine Room" | Horizon — the "Control Tower" | |
|---|---|---|
| Role | Workflow automation | Situational awareness |
| What it does | Runs the logic: state transitions, AI prompt chains, background system tasks | Displays the state of the world on a persistent visual interface |
| Question it answers | "How does the work get done, and who does it next?" | "What is happening across our entire operation right now?" |
| OODA phases | Act — plus the automated decisions you delegate to workflows | Observe, Orient, Decide — the human on the loop |
Analytics and the Ontology complete the engine: your data spaces feed governed managed tables, the ontology types them, and Horizon fuses them onto a single picture.
The OODA loop, compressed
Every operations doctrine since Boyd frames decision-making as the OODA loop — Observe, Orient, Decide, Act — and the organization that cycles it faster than the situation changes is the one that stays in control. Horizon and the engine underneath it are the platform's implementation of that loop, end to end and event-driven (no polling):
- Observe — a webhook or scheduled workflow receives telemetry, writes it to a governed managed table, and opens (or updates) a process-flow instance per track; duplicate detections collapse onto the existing instance automatically. A thin signal fans out instantly to every authorized operator watching the picture — it carries no sensitive payload, so clients re-read through governed endpoints at their own clearance.
- Orient — the picture gives the raw observation its meaning: the ontology types the track, Analytics joins it against everything already known, and AI assessments and advisories appear alongside the live map — fused, clearance-filtered context instead of a bare dot.
- Decide — the operator records a decision on the running process flow: classify the track, send it back for re-assessment, or approve the response. AI recommends; the human on the loop decides.
- Act — the decision writes through to the governed table and unlocks the downstream automation — distribution, tasking, reporting — back in the engine room. What the automation changes shows up on the picture, and the loop begins again.
The division of labor is the point: Horizon owns Observe → Orient → Decide (awareness and the human decision), Agentic owns Act (and any decision you explicitly automate). Compressing the loop from minutes to seconds is what a COP is for.
Everything on a Horizon picture respects the platform's governance stack: document clearance, column masking, and Rego authorisation apply to the map exactly as they do to tables and search. See Security.
Watch updates
A presence watch evaluates observations inside its configured area and time window. Stale or future-dated observations do not contribute to the detection. Repeated observations of the same entity do not multiply its contribution to the score.
During the cooldown, an unchanged signature is suppressed. If the watch permits updates for a new role or a sufficient score increase, it can update the existing detection even within the same minute. Historical ontology reads include derived detections, so earlier evidence remains available after an update.
Where a COP earns its keep
The COP pattern was born in defense and has spread to every industry that runs distributed physical operations. Horizon targets the same use cases:
Defense, intelligence & aerospace
The birthplace of the COP. Military operations fuse thousands of telemetry feeds — satellite imagery, drone video, blue-force (friendly) tracking, radar, weather — into a single, secure map interface. Use case: theater-wide situational awareness, mission planning, threat detection, and synchronized joint-force execution.
Emergency management & public safety
First responders, civil defense, and homeland security agencies rely on COPs to manage crises. Use case: during floods, wildfires, or industrial accidents, a COP bridges the communication gap between police, fire departments, medical teams, and NGOs — integrating dispatch (CAD) logs, field-agent GPS positions, traffic cameras, and real-time hazard mapping.
Energy, utilities & critical infrastructure
Managing large-scale physical networks requires an immediate understanding of system health and external risk. Use case: electrical grids, water networks, and oil & gas pipelines layer SCADA sensor alerts, asset health data, weather forecasts, and field-technician locations onto a single pane of glass to prevent — or rapidly respond to — outages and leaks.
Healthcare & hospital command centers
Modern hospital networks treat bed allocation and patient flow like air traffic control. Use case: fusing ER wait times, ICU bed availability, incoming ambulance telemetry, and staffing levels across regional facilities to optimize resources and handle surge or mass-casualty events.
Maritime, ports & supply-chain logistics
Ports and large logistics networks are inherently chaotic — dozens of moving parts, international data spaces, changing environments. Use case: managing container-ship arrivals, berth assignments, weather, AGV coordinates, and customs clearance in real time. Humanitarian logistics use the same pattern to coordinate cross-agency relief in disaster zones.
The common thread: data fusion
Across all of these, the value is interoperability and data fusion. A COP matters to an enterprise because it takes your data spaces (the siloed databases) and your analytics (the background crunching) and visualizes them on a single timeline or map — so that human operators and your Agentic workflows can act instantly.