Illustrative enterprise prototype

What are we working on?

Ask a question, create an output, or build an app that works across your tools and governed enterprise data.

Get started across the organization

Build a cyber resilience dashboard across customer-facing applications, APIs, airport and station services, workforce access, and third-party connectivity.

AI

I need scoped access to the enterprise service catalog and summarized security telemetry.

Gatekeeper: Security telemetry + service catalogRequesting read-only access to service ownership, control coverage, policy exceptions, and aggregated security events. Credentials remain isolated and every query is logged.

Approved. Do not expose raw event, passenger, or employee data.

AI

Access is limited to aggregated resilience metadata.

Executed code

const posture = await env.SECURITY_GRAPH.assess({
  domains: ["customer-digital", "api", "station-edge",
    "workforce", "third-party"],
  controls: ["ddos", "waf", "api-security",
    "zero-trust", "dns"]
});
const priorities = posture.findings.filter(f =>
  ["critical", "high"].includes(f.severity)
);
workspace.render(CyberResilience, { posture, priorities });

AI

Cyber Resilience Dashboard is ready with illustrative data. It highlights three architecture decisions without exposing raw telemetry.

Draft a target-state operational resilience architecture for customer digital, airport and station, workforce, cloud, and third-party services.

AI

I will use centrally governed standards from the context library.

Context library: Architecture + operationsLoading read-only architecture principles, resilience standards, service dependencies, data governance, and vendor requirements.

AI

The blueprint needs three explicit decisions:

1. Business-service ownership and measurable resilience objectives
2. Decoupling boundaries between customer journeys and core operational systems
3. Graceful degradation and recovery patterns for critical services

AI

Operational Resilience Reference Architecture v1.0 is ready. The document maps service domains to target patterns, guardrails, and validation evidence.

Build a 90-day AI portfolio roadmap that moves the highest-value use cases into governed pilots and creates a reusable enterprise platform.

AI

I will request read-only portfolio and architecture data.

Gatekeepers: AI intake + model registry + architecture repositoryRead use-case stage, business owner, risk tier, model, evaluation status, and milestone. Exclude prompts, passenger data, and employee content.

Approved for aggregate portfolio planning only.

AI

Executed code

const useCases = await env.AI_PORTFOLIO.list({
  horizon: "90-days", status: ["proposed", "pilot"]
});
const controls = await env.MODEL_REGISTRY.controlStatus();
deck.generate(AIPortfolioRoadmap, {
  useCases, controls, dataMode: "aggregate-only"
});

AI

90-Day AI Portfolio Roadmap is ready. It organizes foundation, pilot, evaluation, and scale gates around five workstreams.

Enterprise Cyber Resilience

Illustrative dataCISO live app

94%

Critical digital services with standard edge controls

4

High-priority resilience decisions

81%

Airport and station policy coverage

30m

Executive incident triage target

Priority architecture findings

Day-of-travel APIs: Two illustrative legacy paths sit outside standard API discovery and schema validation policy.

Third-party support: Four persistent vendor paths do not enforce device posture or time-bound authorization.

Airport and station edge: Nine illustrative locations use direct-origin failover that bypasses centralized inspection.

Control coverage by domain

Southwest.com + mobile96%Protected

Rapid Rewards + identity92%Protected

Corporate workforce88%Managed

Airport + station edge81%In progress

Third-party access73%Review

Legacy origin paths64%Priority

Generated application query

const posture = await env.SECURITY_GRAPH.assess({
  domains: ["customer-digital", "api", "station-edge", "workforce", "third-party"],
  controls: ["ddos", "waf", "api-security", "zero-trust", "dns"],
  dataMode: "aggregate-only"
});

return render(<CyberResilienceDashboard data={posture} />);

Operational Resilience Reference Architecture v1.0

CIO draft

Enterprise Operational Resilience Reference Architecture

Version 1.0 · Illustrative draft · August 2026 · Owner: Enterprise Technology

1. Purpose

Define a consistent target state for customer digital, airport and station, workforce, cloud, data, and third-party services. The architecture organizes technology around measurable business services and reduces coupling between experience layers and critical operating systems.

2. Design principles

Design around customer and employee journeys rather than application boundaries.
Use stable service and event contracts to isolate change.
Assign explicit ownership and resilience objectives to critical business services.
Build observability, graceful degradation, and tested recovery into target patterns.
Centralize control evidence while minimizing retained sensitive data.

3. Architecture decisions

Service domain	Target-state decision	Required guardrail	Validation evidence
Customer digital	Experience services decoupled from core operational systems	Stable APIs, event contracts, caching, and graceful degradation	Journey SLOs and dependency tests
Airport and station	Resilient edge services with centrally governed policy	Offline-safe patterns, segmented access, and remote observability	Station recovery exercises
Workforce and crew	Identity-aware access to named applications	Strong authentication, device posture, and least privilege	Access and continuity tests
Third-party services	Contract-first integration with bounded failure modes	Time-bound access, circuit breakers, and service ownership	Supplier dependency reviews
AI and automation	Shared model access, evaluation, and tool-control plane	Service identity, governed data, approval, and audit	Model and agent evaluations

Architecture decision: Degraded-mode patterns must preserve operational safety and service continuity without creating permanent bypasses or unmanaged data paths.

4. Validation sequence

Map critical business services, normalize ownership, define target patterns and evidence, validate representative digital and station flows, test failure modes, and use measured outcomes as the gate for scaled modernization.

90-Day Governed AI Portfolio Roadmap

Head of AI deck

Slide 1 of 4

90-Day Governed AI Portfolio Roadmap

From isolated experiments to reusable, trusted enterprise capabilities

Slide 2 of 4

Priority workstreams

Workstream	Accountable team	Stage	Progress
Model gateway + observability	AI Platform	Design	40%
AI use-case governance	AI Council	Pilot	60%
Enterprise retrieval	Data + AI	Discovery	30%
Agentic workflow controls	Security + AI	Pilot	45%
AI value + FinOps	Strategy + Finance	Design	35%

Illustrative portfolio data for demonstration.

Slide 3 of 4

Decision-gate success measures

100%Production models inventoried

>90%Required evaluation pass rate

100%High-impact actions approved

3Reusable pilot components

Slide 4 of 4

90-day execution path

Days 1-30Baseline. Inventory use cases and models; agree on risk tiers, evaluation, and ownership.

Days 31-60Build. Launch shared model access, approved tools, observability, and the first pilot cohort.

Days 61-90Validate. Measure quality, risk, cost, adoption, and realized value in production-like use.

DecisionScale gate. Productionize winners, retire low-value experiments, and fund shared capabilities.

Aggregated source records

System	Scope	Fields used	Permission
AI use-case portfolio	Proposed and pilot initiatives	Owner, stage, value, risk tier	Read-only
Model registry	Approved model endpoints	Model, evaluation, control status	Read-only
Architecture repository	AI platform decisions	ADR title, state, review date	Read-only

Planning cadence, reporting definitions, approval policy, audit requirements, and internal controls.

Read-onlyUpdated Aug 2026

Skills

Reusable, governed workflows available across business functions.

Name	Description	Group	Source
meeting-prep	Combine calendar, communications, and approved context into concise briefing material.	General	Enterprise
weekly-report	Compile cross-tool activity, outcomes, risks, and next steps into a weekly summary.	General	Enterprise
incident-response	Draft and update response procedures grounded in approved standards and control mappings.	Security	Enterprise
vendor-assessment	Generate security questionnaires, gather evidence, and summarize third-party risk.	Security	Enterprise
architecture-review	Build decision-oriented reviews from project, service catalog, and architecture data.	Architecture	Enterprise
change-impact	Map dependencies, affected services, business owners, and implementation risk.	Architecture	Enterprise
disruption-planning	Model coordinated actions, dependencies, and customer impacts for operating scenarios.	Operations	Enterprise
service-dependency-map	Trace a business service through applications, data, infrastructure, and vendors.	Operations	Enterprise
runbook-draft	Convert approved procedures and system context into reviewable operating runbooks.	Operations	Enterprise
customer-insights	Analyze approved loyalty and experience trends using aggregated customer measures.	Customer	Enterprise
journey-analysis	Find friction and improvement opportunities across planning, travel, service, and loyalty.	Customer	Enterprise
budget-analysis	Compare actuals to plan, explain variances, and forecast period-end outcomes.	Finance	Enterprise
job-posting-draft	Draft role descriptions using approved competencies, templates, and compensation guidance.	HR	Enterprise
onboarding-guide	Create role-based onboarding plans from policy, team context, and access procedures.	HR	Enterprise
model-risk-review	Apply proportional controls for data, models, outputs, actions, and human oversight.	AI	Enterprise
agent-design	Define an agent's tools, boundaries, approvals, evaluation, and operating ownership.	AI	Enterprise

Integrations

Illustrative enterprise connections. Gatekeepers isolate credentials, enforce scoped permissions, and log agent actions.

Enterprise systems

M365

Microsoft 365

Mail, calendar, documents, spreadsheets, presentations, and files.

TM

Microsoft Teams

Messages, channels, meetings, and governed collaboration workflows.

J

Jira

Projects, epics, issues, milestones, and delivery status.

C

Confluence

Enterprise knowledge, project spaces, and approved documentation.

SN

ServiceNow

IT services, HR cases, change requests, service catalog, and facilities.

W

Workday

Employee records, organizational structure, payroll, leave, and benefits.

SAP

Finance, supply chain, procurement, assets, and planning.

SF

Salesforce

Business relationships, service workflows, and approved customer programs.

GH

GitHub

Repositories, pull requests, issues, releases, and engineering metadata.

DB

Snowflake

Governed enterprise data, analytics, and cross-functional reporting.

SIEM

Security analytics

Alerts, summarized events, control telemetry, and security correlation.

BI

Power BI / Tableau

Dashboards, executive reporting, and governed visualization.

MCP servers

Customer Digital Servicesmcp://customer-digital

Authorized

Operational Data Fabricmcp://operational-data

Needs approval

Airport & Station Systemsmcp://airport-and-station

Needs approval

Employee Directorymcp://employee-directory

Authorized

Cloudflare APIhttps://mcp.cloudflare.com/mcp

Authorized

AI Gateway

Visibility and controls across every AI provider Southwest uses — one console.

Requests

128,400

▲ 11% vs last mo

Tokens

342M

▲ 8% vs last mo

Est. spend

$9,120

76% of budget

Cache-hit

27%

▲ saves ~$2.4k

Error rate

0.6%

▼ 0.2 pts

p50 latency

480 ms

across providers

Models in Use

This month

Model	Route	Tokens	Spend	Share	p50 latency
Llama 3.3 70B	Workers AI	156M	$2,140	42%	310 ms
Claude	via AI Gateway	98M	$3,980	24%	720 ms
GPT-4o	via AI Gateway	61M	$2,510	18%	640 ms
Workers AI embeddings (bge)	Workers AI	27M	$190	16%	40 ms

Spend vs. Budget

9 days remaining

$9,120spent of $12,000 cap

76%

On track · ~$2,880 left with 9 days

Top Users

Aravind M.42M tok $1,180

Marcus R.31M tok $960

Aisha L.28M tok $840

Dev K.22M tok $610

Usage by Workspace / Team

342M tokens total

AI Platform

121M tokens · $3,240

Enterprise Technology

89M tokens · $2,460

Customer Digital Services

62M tokens · $1,510

Airport & Station Systems

41M tokens · $1,020

Security + AI

29M tokens · $890

Governance

Guardrails enforced by Gatekeepers + AI Gateway, so Security + AI teams can operate safely at scale.

Per-team allowed models

Restrict which providers each workspace can call.

Monthly spend caps

Hard limits per team; agents stop before overrun.

AI Platform $4,000Enterprise Technology $3,000

PII redaction

Strip sensitive fields from prompts before they leave.

Enabled

Prompt / response logging

Full request logs retained for audit & review.

Enabled · 90-day retention

Rate limits

Per-team request ceilings to protect budgets.

600 req / min|burst 1,000

Raise AI Platform cap to $6,000

Change queued by an agent — needs a human sign-off.

Requires approval