Skip to main content

// signature workflow

The five-minute loop your team will not run manually again.

Narya Command is an automation application, not a chat wrapper. Most steps are deterministic connector actions; AI prompt steps appear only where extraction or judgment is required. Tokens, timing, and outcomes are in the run log.

WorkflowsRunsSchedulesConnectionsSkillsTemplatesToolsPrompt

dashboard

Command center

Monitor notifications, live runs, and recent outcomes from one surface.

workflows

Workflow builder

Compose connector, validation, and AI steps into reusable automations.

runs

Run history

Inspect every run with trigger, duration, status, and pause points.

// the signature workflow

  1. 01
    ScheduleRun this workflow every five minutes
  2. 02
    Email scanCheck inbox for new CygNet help requests
  3. 03
    AI analyzeClassify the request and extract device, symptom, urgency into strict JSON
  4. 04
    Azure DevOpsOpen or update a work item with the request and assigned engineer
  5. 05
    CygNet diagnosticsQuery points, recent alarms, and trend context for the device
  6. 06
    Write findingsUpdate the Azure DevOps work item with the diagnostic summary
  7. 07
    Notify the Engineer in the Command CenterA new item needs to be triaged

See the workflow in Narya Command

// how it works

How Narya Command works

Build a workflow once — email triage, work item creation, SCADA investigation — then run it on a schedule with preview/apply safety and a full audit trail across every connector step.

01
Schedule the run

A workflow starts on a schedule, from email, or from an operator. Set it to run every five minutes, scan the inbox for CygNet help requests, and route to the right connector actions.

"Every five minutes, scan inbox for CygNet help requests and route to the right workflow."

02
Cross-system steps execute

Deterministic connector steps handle the known work — query CygNet, open an Azure DevOps work item, validate diagnostics. AI prompt steps run only where extraction or judgment is needed, with strict schemas and logged outputs.

Schedule → email scan → classify → Azure DevOps work item → CygNet query.

03
Update and notify

Update the Azure DevOps work item with findings, persist the run log, and notify the on-call engineer. Writes use preview/apply when they touch SCADA.

Findings → work item update → notify engineer → run log retained.

// before vs after

- before

Engineer reads email at 8 AM, copies the request into a work item, opens CygNet Studio, runs four queries by hand, types findings back into the work item, forwards to the next person. Repeats at 9. And 10.

+ after

A workflow runs every five minutes. It reads new email, classifies CygNet requests, opens an Azure DevOps work item, runs diagnostics, writes findings to the work item, and notifies the engineer that triage is waiting.

// why workflows, not chat-only

Engineers, multiplied

Not by replacing them with a chatbot. By giving them an automation application where known work runs deterministically and the model is invoked only when judgment is needed. We built this after model-first approaches produced hallucinations and inconsistent results in production.

Deterministic connector steps
Clone a device, bulk-update facilities, ack an alarm, create a work item — explicit actions with predictable outcomes. No tokens burned. Same result every run.
connector_actionloopcalculatorhuman_gate
Intelligence where it earns its place
AI prompt steps for classification, extraction, and ambiguous context — strict JSON schemas, evidence in the run log, human review before anything touches SCADA.
ai_promptclassifyextract
Lessons from production
Leaving everything to the model does not scale in OT. Repeatable ops need repeatable steps; judgment calls get a prompt step and a gate — not a free-form chat thread.
preview_applychangesetchronicle

// same workflow, mixed step types

+ deterministic

Schedule trigger → read inbox → Azure DevOps work item → CygNet diagnostics → notify engineer

+ ai_prompt (one step)

Classify the email and extract device, symptom, and urgency into strict JSON — then back to connector steps

// the stack

App, intelligence, connectors

Narya Command is the automation app. WAIS is the intelligence layer. Built-in connectors reach email, Azure DevOps, and CygNet today. Snowflake is on the roadmap. Inference is swappable — cloud by default, on-prem when policy demands it.

// capabilities

One automation app. Real connectors.

Workflows, runs, connector plugins, Skills, preview/apply safety, and WAIS learning — everything you need to automate operational work across email, SCADA, and engineering systems without losing governance.

Workflows & automations
Mix deterministic connector steps with AI prompt steps only where you need them. Triggers, loops, validation, and human gates — save templates and rerun with the same outcome every time.
Runs & audit trail
Every run logs step-by-step events: tokens, duration, connector results, and approval decisions. 30-day changeset retention for reviewed writes.
Connector plugins
CygNet, Email (Outlook Classic or Microsoft 365), and Azure DevOps built in today. Snowflake on the roadmap. One run engine, one safety model, one audit trail.
Skills, not scripts
Reusable playbooks triggered by slash commands. Built on the open Agent Skills format — portable operational knowledge your team owns.
Preview / apply safety
Writes go through preview → confirm → changeset. Rollback flags conflicts and can restore historian values when VHS history is selected.
WAIS — your site's chronicle
Corrections become lessons. Lessons reinforce into conventions. Conventions graduate into Skills. Centralize WAIS and every operator's knowledge compounds.

// use cases

Three teams. One automation application.

Cross-system triage, deep CygNet administration, and platform governance — all from the same app.

Ops triage
Route field email through classification, Azure DevOps work items, and SCADA investigation — one workflow, one audit trail.
  • >Scheduled inbox scan every five minutes
  • >Azure DevOps work item auto-creation
  • >Human gate before SCADA writes
SCADA engineers
CygNet is the deep SCADA connector — clone devices, bulk-update facilities, and query points with preview/confirm/rollback built in.
  • >New-site setup measured in minutes
  • >Spreadsheet-driven bulk operations
  • >Preview / confirm / rollback on writes
IT / OT platform owners
Deploy a governed automation application on Windows — connector plugins, workflow catalog, and WAIS chronicle without a separate SaaS rollout.
  • >Runs locally on the workstation
  • >Workflow catalog for recurring jobs
  • >Built-in knowledge base and audit

// architecture

App, intelligence, connectors

The operator works in Narya Command, an automation application for workflows and scheduled runs. WAIS retrieves knowledge and grows your site's chronicle. Connector plugins reach email, Azure DevOps, CygNet, and more.

0102030405OPERATOREngineer · Ops manager · Field tech · Triage queueNARYA COMMANDWorkflows · Runs · Tools · Skills · Connections · Templates · AutomationsWAISHybrid retrieval · Memory · Lessons · Conventions · Golden rules · SafetyCONNECTOR PLUGINSCygNet · Email · Azure DevOps (shipping) · Snowflake (roadmap)EXTERNAL SYSTEMSSCADA · Email · Ticketing · Chat · Data warehouse · Field devices

One safety model. One audit trail. Connectors talk to external systems; for CygNet, CygNet ACS stays the authority for permissions and audit — we don't replace it. Read the full architecture →

// the flywheel

How WAIS gets smarter at your site

Every CygNet system is configured differently. WAIS observes how your team uses it, captures the corrections, and grows the chronicle — so the next administrator gets the answer instead of relearning it.

  1. 01

    Use

    Engineer runs a skill or asks Narya to handle a task.

  2. 02

    Correction

    Failure-then-success patterns are detected automatically — without the user having to flag them.

  3. 03

    Lesson

    A candidate lesson is captured and tagged with scope, provenance, and a starting confidence.

  4. 04

    Reinforcement

    Repeated observations grow confidence. Senior administrators carry more weight, so trusted knowledge promotes faster.

  5. 05

    Convention

    Reinforced lessons graduate into site conventions. Entity-level corrections lift to class-level rules over time.

  6. 06

    Skill

    Promoted into a reusable Skill so every administrator gets the answer next time — not just the one who learned it.

// the moat

Centralize WAIS, and every admin's knowledge becomes every admin's knowledge.

Run one Westmarch server for the team and every correction, every convention, every learned alias is immediately available to every other administrator on your site. Not next quarter. Not after a model retrain. Now. See how WAIS works →

// north star

80% resolvable from accumulated knowledge

The long-term goal is a coverage-driven triage layer: 80% of incoming operational problems at a mature site resolvable from your accumulated CygNet knowledge alone — without fresh learning every time. WAIS is the substrate. The Triage Agent is the surface — shipping in Narya Command today.

// the metric

coverage_ratio — distinct (device_type × action × outcome) tuples observed at your site, divided by the known-possible surface. A measurable quantity, not a vibe.

// safety stays

Higher coverage never means looser safety. Preview, confirm, changeset, rollback, and human-in-the-loop approval remain the default at every autonomy level — including scheduled triage workflows.

// what's shipping today

Real surface area, not vague promises

0+
connectors
CygNet, Email, and Azure DevOps built in; Snowflake on roadmap
0+
step types
trigger, loop, AI prompt, connector, gate, save
0
CygNet services
deep SCADA coverage in the first expert wedge

// ready?

Put Narya Command in front of your team

Start with the beta for governed workflows, connector plugins, Skills, and rollback-ready writes — with WAIS learning your site conventions from day one.