HR Leave · Agentic AI Tutorial

ServiceNow Agentic AI + AI Control Tower

3-minute instance walkthrough: open the screens, know what each area does, and apply the same path on your lab. Inspired by Put AI to Work session themes (Autonomous IT / Otto, Control Tower, AI Agent Studio).

Time~3 minutes guided click-path
FocusAgentic AI · Control Tower · Agent Studio
URL token{BASE}
Deep guideChapter 14
Replace {BASE} then open AI Control Tower AI Agent Studio Now Assist / Otto surface Platform Analytics AI Agents list Flows

How to use this page: Keep it beside your instance. Each “screenshot” block shows what the UI area looks like conceptually. Click the matching direct URL, then follow the 3-minute timer. Plug-ins / Now Assist AI Agents must be licensed and installed — if a URL 404s, search the filter navigator for the product name.

0. Map in 20 seconds

Theme (event)Product areaOne-line job
Autonomous IT + Otto AI assistant for work Ask / act across ITSM·ITOM·SPM workflows
AI Control Tower + Workflow Data Fabric Governance + data for agents Inventory, risk, control, feed agents with trusted data
Build Agent + AI Agent Studio Create & manage agents Build agentic workflows and tools your digital workforce runs

1 Minute 1 — AI Control Tower

Summary: Control Tower is the command center for enterprise AI — see agents/models, govern risk, monitor runtime, and measure value. You cannot control what you cannot see.

Screenshot cue: Control Tower overview with inventory KPIs and left nav (Inventory / Governance / Runtime). Capture this after opening the workspace on your instance.

Open on instance

  1. Filter navigator → type AI Control Tower → open workspace
    Direct: {BASE}/now/ai-control-tower (alternate: Workspaces menu → AI Control Tower)
  2. Open AI Inventory — list agents, models, providers.
  3. Open one record — note lifecycle stage, owner, risk.
  4. Glance Governance / Assessments — policies and review status.

Functions (Control Tower)

Discover / Inventory

Find first- and third-party AI agents, models, MCP servers — reduce blind spots.

Govern risk

Assessments, policies, compliance alignment before agents go wide.

Identity & access

Track non-human / AI identities and what they can reach.

Monitor runtime

Security telemetry, prompt-injection signals, action logs.

Measure value

Executive view of adoption and outcome metrics.

CMDB / CSDM link

Tie AI assets to services and configuration context.

2 Minute 2 — AI Agent Studio (Build Agent)

Summary: AI Agent Studio is where you create, manage, and test AI Agents and Agentic Workflows — your digital workforce builders. Requires Now Assist AI Agents capability installed.

Screenshot cue: Create and manage list with Agentic Workflows / AI Agents tabs. Capture after clicking Create and manage.

Open on instance

  1. Filter navigator → AI Agent Studio
    Direct try: {BASE}/now/ai-agent-studio
  2. Click Create and manage.
  3. Open tab AI Agents — note existing agents (or empty on fresh lab).
  4. Open tab Agentic Workflows — open one sample if present.
  5. In Guided Setup, find Define Data Access:
    • Act as invoking user — runs with caller’s roles
    • AI User — dedicated identity (least privilege for agents)

Functions (Agent Studio)

Create AI Agent

Define goal, instructions, tools the agent may call.

Agentic Workflow

Orchestrate multi-step agent work toward a business outcome.

Tools / skills

Wire Flow actions, Script Includes, catalog, records — e.g. HR Leave utils.

Data Access / AI User

Choose execution identity — critical for secure agent runs.

Test & evaluate

Single tests and pattern evaluations before production.

Dashboard

Usage and performance of agents over time.

HR Leave tip: Build agents that call existing scoped APIs / Script Includes from chapters 01–12 — do not bypass ACLs. Deep build steps remain in 14_AI_SERVICENOW_AGENTS.html.

3 Minute 3 — Otto / Now Assist + Workflow Data Fabric

Focus of this minute: Use the AI assistant on the instance (Otto / Now Assist), run safe prompts, see which skill/agent ran, then open Data Fabric / Integration paths that feed agents with trusted data. This matches the event themes Autonomous IT + Otto and Workflow Data Fabric.

Summary: Otto (AI assistant for work) unifies assistant experiences so users ask and work gets done across workflows. Workflow Data Fabric connects enterprise / external data so agents and assist skills act on accurate, governed context — not guesses.

PieceWhat it isWho uses it
Otto / Now AssistSide panel / AI Experience — ask in natural language; platform executes via skills, agents, flowsEveryone (employee, agent, fulfiller)
Workflow Data FabricData layer that brings CMDB + external systems into AI/workflow contextArchitect, Integration, AI builder
Link to Control TowerAssist/agents you invoke should appear in AI Inventory for governancePlatform Owner / CoE
Screenshot 5 — Assist: Capture the open Otto/Now Assist panel with your prompt and the answer. Note any “skill / flow / agent” label shown.
Screenshot 6 — Data Fabric: Capture Integration Hub / Remote Tables / Data Fabric (whatever your release exposes) showing connected sources. If missing, capture navigator search results for those terms.

3A — Open Otto / Now Assist on the instance (≈90 sec)

  1. Open assist
    • Direct: {BASE}/now/assist
    • Or click the Otto / Now Assist / sparkle icon in the instance header (name varies by release).
    • Filter navigator fallback: type Now Assist or AI Experience.
  2. Confirm panel loads — Chat / Ask visible. If blocked, check license + Now Assist plugins.
  3. Run one safe lab prompt (pick one):
    PromptWhat you verify
    List my open incidentsReturns INC numbers only from accessible data
    Summarize this incident (open an INC first)Uses record context
    What is my leave balance? (if HR Leave skills exist)Scoped HR Leave skill / VA topic
    How do I request annual leave?Knowledge / catalog guidance — no write if not intended
  4. Note the trail — skill name, agent, or flow mentioned in the answer UI (you will match this in Control Tower inventory later).
  5. Optional action — if Otto offers “create / update / assign”, cancel on lab unless you intend to test write paths.

3B — Open Workflow Data Fabric / data paths (≈60 sec)

  1. Filter navigator search (try in order):
    • Data Fabric
    • Integration Hub
    • Remote Tables
    • Flow Data / Connections
  2. Useful direct URLs to try:
  3. Discuss with team: Which sources should agents read for HR Leave / ITSM? CMDB only? External HRIS? Block PII fields.

3C — Close the loop with Control Tower (≈30 sec)

  1. Re-open {BASE}/now/ai-control-tower
  2. Search AI Inventory for the skill/agent name you saw in Assist.
  3. Confirm: owner, risk stage, whether it is governed.

Functions — Otto / Assist (detailed)

Conversational ask

Natural language Q&A and “do this work” requests.

Context awareness

Uses open record / user / role to ground answers.

Skill routing

Picks Now Assist skill, agent, or flow for the intent.

Execute / finish work

Creates or updates work through platform workflows — not shadow IT.

Cross-product assist

ITSM · ITOM · SPM (Autonomous IT) in one assistant surface.

History & audit

Conversation and actions remain reviewable for support/governance.

Voice / multimodal*

Where licensed — voice agents & richer AI Experience.

Employee + agent UX

Same assistant pattern for self-service and fulfiller work.

Functions — Workflow Data Fabric (detailed)

Connect sources

CMDB, platform tables, external systems via Integration Hub / remote data.

Contextualize

Shape data so agents get usable, timely context.

Feed agentic AI

Supplies tools/skills with facts instead of hallucinations.

Access policy

Limit which identities/agents can read which sources.

Trust + CMDB

Align AI context with CSDM / CI relationships.

Ops readiness

Support Autonomous IT demos with real ops data paths.

Minute-3 apply checklist

✓ActionURL / where
☐Open Assist / Otto panel{BASE}/now/assist
☐Run one safe promptChat box
☐Write down skill/agent/flow name usedAnswer UI
☐Search navigator for Data Fabric / IH / Remote TablesFilter navigator
☐Open connections or REST aliases (if present){BASE}/sys_alias_list.do
☐Match skill/agent in Control Tower inventory{BASE}/now/ai-control-tower
☐Capture screenshots 5 + 6See screenshot pack

Lab safety: Prefer read-only prompts first. Do not approve mass updates from Assist on shared labs. For HR Leave production skills (balance, cancel, submit), follow the full build in Chapter 14.

Functions cheat-sheet (all three areas)

AreaPrimary usersCore functionsTypical open path
AI Control Tower AI CoE, Risk, Platform Owner Inventory, govern, monitor, measure Workspaces → AI Control Tower
AI Agent Studio Developer / Citizen builder Create agents, agentic workflows, tools, test Navigator → AI Agent Studio → Create and manage
Otto / Assist All employees / agents Ask, act, finish work via workflows Header assist / /now/assist
Workflow Data Fabric Architect / Integration Connect & contextualize data for AI Navigator search Data Fabric / IH

Screenshot pack (what to capture on your instance)

#CaptureFile name suggestionShows
1Control Tower overviewshot-01-control-tower.pngInventory KPIs + nav
2AI Inventory listshot-02-ai-inventory.pngAgents/models rows
3Agent Studio homeshot-03-agent-studio.pngCreate and manage entry
4Agentic Workflow Guided Setupshot-04-guided-setup.pngData Access / AI User
5Otto / Assist answershot-05-otto-assist.pngPrompt + response
6Data Fabric / integration contextshot-06-data-fabric.pngExternal data feed (if present)

Tip: Save screenshots under HR_Leave_App/html/assets/shots-agentic-ai/ later and replace the conceptual frames above with <img src="assets/shots-agentic-ai/shot-01-....png"> when you have real instance captures.

3-minute apply checklist

✓StepDirect URL
☐Open AI Control Tower — note inventory count{BASE}/now/ai-control-tower
☐Open AI Agent Studio → Create and manage{BASE}/now/ai-agent-studio
☐Open one agent or workflow — find Data Accessinside Studio
☐Open Assist / Otto — run one safe prompt{BASE}/now/assist
☐Confirm plugins licensed (Now Assist AI Agents)Plugin list / store
☐For deep HR Leave AI build, continue Chapter 1414_AI_SERVICENOW_AGENTS.html

Bottom line: Control Tower = see & govern AI. Agent Studio = build agents & agentic workflows. Otto/Assist = users ask and work finishes. Data Fabric = trusted data into those agents. Three minutes to open all three; Chapter 14 to implement HR Leave skills end-to-end.