What are we working on?
Ask a question, create an output, or build an app that works with your tools and data.
Get Started
Build an AI infrastructure operations dashboard
Track XPU programs, AI model deployments, and compute utilization
App
Track XPU programs, AI model deployments, and compute utilization
Draft a VMware Cloud Foundation migration policy
Ground the policy in VCF 6.0 standards and compliance requirements
Document
Ground the policy in VCF 6.0 standards and compliance requirements
Build a Q3 AI & platform strategy review deck
Pull data from Jira and Confluence into slides
Slide Deck
Pull data from Jira and Confluence into slides
Connect to our AI infrastructure systems and build a dashboard showing XPU custom silicon program status, AI model deployments across our hyperscaler partnerships, GPU/TPU cluster utilization, and inference latency across data centers.
AI
I'll need access to your AI infrastructure platform. Let me request that through a Gatekeeper.
🔒 Gatekeeper: XPU Program Tracker API
Requesting read-only access to custom silicon program status, tape-out schedules, and compute cluster metrics. The Gatekeeper holds the credential and logs every query.
Requesting read-only access to custom silicon program status, tape-out schedules, and compute cluster metrics. The Gatekeeper holds the credential and logs every query.
Approved.
AI
Access granted. Pulling AI infrastructure telemetry now.
Executed code
const programs = await env.XPU_TRACKER.getProgramStatus({
scope: "all-active",
include: ["silicon", "deployments", "utilization", "latency"]
});
const alerts = programs.clusters.filter(
c => c.utilizationPct > 90 || c.latencyP95 > c.slaTarget
);
app.render(AIInfraDashboard, { programs, alerts });
AI
AI Infrastructure Operations Dashboard is live. Connected via Gatekeeper -- read-only, every query logged.
I see 2 clusters above 90% utilization and one XPU tape-out milestone approaching. Want me to set up Slack alerts when any cluster exceeds 95% capacity?
I see 2 clusters above 90% utilization and one XPU tape-out milestone approaching. Want me to set up Slack alerts when any cluster exceeds 95% capacity?
+
Draft an enterprise-wide VMware Cloud Foundation migration policy. Ground it in our VCF 6.0 standards, address workload classification, and include compliance mapping for SOX and data sovereignty requirements across our global operations.
AI
I'll pull from your company's context library to ground this in Broadcom's standards.
📚 Context Library: Infrastructure & Compliance Standards
Loading Broadcom's VMware Cloud Foundation migration playbook (v6.0), enterprise architecture principles, SOX compliance requirements, and data sovereignty guidelines. These are curated by your infrastructure and compliance teams and read-only to the agent.
Loading Broadcom's VMware Cloud Foundation migration playbook (v6.0), enterprise architecture principles, SOX compliance requirements, and data sovereignty guidelines. These are curated by your infrastructure and compliance teams and read-only to the agent.
AI
Context loaded. I've identified 3 gaps in the current migration framework:
1. Workload classification criteria missing for AI/ML training pipelines
2. Data sovereignty rules not mapped to VCF availability zones for EMEA and APAC
3. Rollback procedures incomplete for stateful containerized workloads
I'll address all three in the updated policy.
1. Workload classification criteria missing for AI/ML training pipelines
2. Data sovereignty rules not mapped to VCF availability zones for EMEA and APAC
3. Rollback procedures incomplete for stateful containerized workloads
I'll address all three in the updated policy.
AI
VMware Cloud Foundation Migration Policy v3.0 draft is ready. Structured with Broadcom-specific workload tiers, compliance mappings, and phased migration timeline.
You can edit directly, export to Confluence, or share for review. The document stays connected to the context library -- if standards change, I can flag sections that need updating.
You can edit directly, export to Confluence, or share for review. The document stays connected to the context library -- if standards change, I can flag sections that need updating.
+
Build a Q3 AI and platform strategy review deck summarizing our XPU silicon programs, VMware Cloud Foundation migration progress, security posture, and upcoming initiatives. Pull data from Jira and Confluence.
AI
I'll need access to your project tracking and wiki systems.
🔒 Gatekeeper: Jira (read-only)
Requesting access to the XPU, PLATFORM, and SECURITY projects. Read epics, status, and completion data only.
Requesting access to the XPU, PLATFORM, and SECURITY projects. Read epics, status, and completion data only.
🔒 Gatekeeper: Confluence (read-only)
Requesting access to the AI Strategy and Infrastructure Architecture spaces. Read pages only.
Requesting access to the AI Strategy and Infrastructure Architecture spaces. Read pages only.
Both approved.
AI
Executed code
const epics = await env.JIRA.getEpics({
projects: ["XPU", "PLATFORM", "SECURITY"],
quarter: "Q3-2026"
});
const adrs = await env.CONFLUENCE.getPages({
space: "ai-strategy",
since: "2026-07-01"
});
deck.generate(QuarterlyReview, { epics, adrs });
AI
Q3 AI & Platform Strategy Review deck is ready -- 4 slides from live Jira and Confluence data.
You can regenerate anytime to pick up changes, or edit individual slides. Export to Google Slides or PDF when ready to present.
You can regenerate anytime to pick up changes, or edit individual slides. Export to Google Slides or PDF when ready to present.
+
Gadget
Code
Connections
AI Infrastructure Operations Dashboard
7
Active XPU Programs
142
AI Model Deployments
87.3%
Avg Cluster Utilization
4.2ms
P95 Inference Latency
Capacity Alerts
⚠️ GCP us-central1 -- TPU v5p Cluster -- Utilization at 96.4%. 2 training jobs queued. Auto-scaling request submitted.
⚠️ Azure East US -- GPU A100 Pool -- Utilization at 93.1%. Inference latency P95 elevated to 6.8ms (SLA: 5ms).
⚠️ XPU Gen-4 Tape-out -- Milestone T-14 days. TSMC N3E mask review pending sign-off from Silicon Design Lead.
Compute Cluster Status
GCP us-central1 (TPU v5p)
96.4%
High
Azure East US (A100)
93.1%
High
AWS us-west-2 (H100)
84.0%
Monitor
GCP europe-west4 (TPU v5e)
78.0%
Monitor
On-Prem Palo Alto (XPU Test)
62.0%
Healthy
AWS ap-northeast-1 (L40S)
54.0%
Healthy
On-Prem Singapore (Validation)
41.0%
Healthy
Page
Source
Connections
VMware Cloud Foundation Migration Policy v3.0
Broadcom -- VMware Cloud Foundation Migration Policy
1. Purpose
This policy establishes the standards, procedures, and governance framework for migrating enterprise workloads to VMware Cloud Foundation (VCF) 6.0. It ensures consistent workload classification, compliance adherence, and risk mitigation across Broadcom's global infrastructure operations.
2. Scope
- All on-premises vSphere environments targeted for VCF consolidation
- Hybrid cloud workloads spanning VMware Cloud on AWS, Azure VMware Solution, and Google Cloud VMware Engine
- AI/ML training and inference pipelines running on VMware-managed infrastructure [NEW]
- CA Technologies mainframe-to-cloud bridge workloads
- Symantec security appliance VMs and endpoint management platforms
3. Workload Classification
| Tier | Description | Migration Priority | Downtime Tolerance |
|---|---|---|---|
| Tier 0 -- Critical | XPU design EDA pipelines, customer-facing SaaS platforms, financial systems | Phase 3 (validated) | Zero (live migration only) |
| Tier 1 -- Business | AI model training clusters, CI/CD pipelines, Symantec management consoles | Phase 2 | < 30 min |
| Tier 2 -- Standard | Internal tools, dev/test environments, analytics and reporting workloads | Phase 1 | < 4 hours |
| Tier 3 -- Archive | Legacy CA Technologies workloads, deprecated services, cold storage VMs | Phase 1 (lift & shift) | < 24 hours |
AI/ML Workload Gate: Any workload classified as AI/ML training or inference must undergo GPU/XPU passthrough compatibility validation and NUMA topology assessment before migration scheduling. The AI Infrastructure team must sign off on performance benchmarks post-migration.
4. Data Sovereignty Requirements
| Region | Regulation | VCF Availability Zone | Data Residency |
|---|---|---|---|
| EMEA | GDPR, EU Data Act | eu-west-1 (Frankfurt) | EU-only storage & processing |
| APAC | PDPA, PIPL | ap-southeast-1 (Singapore) | In-region processing required |
| Americas | SOX, CCPA | us-east-1 (Virginia) | US-only for financial data |
5. Rollback Procedures
- Stateless workloads: automated VCF snapshot-based rollback within 15 minutes
- Stateful containers: persistent volume snapshot + application-consistent quiesce before cutover
- Database workloads: parallel-run validation period of 72 hours before decommissioning source
- AI training jobs: checkpoint-based recovery with distributed storage synchronization [NEW]
Slides
Source
Connections
Q3 AI & Platform Strategy Review
Integrations
Connect external services to Broadcom OS. Gatekeepers govern access, scope permissions, and log every action.
Gatekeepers
Google Workspace
Gmail, Docs, Sheets, Slides, Calendar, Drive
Slack
Send messages, read channels, manage workflows
Jira
Projects, epics, issues, sprints, and boards
Confluence
Read and write wiki pages, search knowledge spaces
ServiceNow
IT tickets, change requests, CMDB, incident management
Workday
Employee data, org charts, payroll, time off, benefits
SAP
Finance, supply chain, procurement, and inventory management
Salesforce
Enterprise accounts, partner pipeline, customer success
GitHub Enterprise
Repositories, pull requests, CI/CD pipelines, infrastructure-as-code
Snowflake
Data warehouse, analytics, cross-department reporting
Symantec / Splunk SIEM
Security events, alerts, endpoint protection, and threat correlation
Power BI / Tableau
Dashboards, executive reporting, and data visualization
MCP Servers
Remote MCP servers available to all workspaces.
XPU Program Tracker
https://xpu.mcp.broadcom.internal/mcp
Auto
VMware vCenter API
https://vcenter.mcp.broadcom.internal/mcp
Needs auth
Symantec Security Console
https://symantec.mcp.broadcom.internal/mcp
Needs auth
Employee Directory
https://directory.mcp.broadcom.internal/mcp
Auto
Cloudflare API
https://mcp.cloudflare.com/mcp
Auto
Context
Curated reference documents that ground your agent in Broadcom's knowledge. Published centrally, read-only to all agents and workspaces.
ai-semiconductor-strategy-fy2027.md
XPU roadmap, custom silicon partnerships, AI accelerator priorities, and hyperscaler engagement plans
brand-voice-guidelines.md
Tone, terminology, customer-facing communication standards, and approved messaging templates
security-standards.md
Symantec enterprise security policies, Zero Trust framework, endpoint protection, and SASE guidelines
architecture-principles.md
Enterprise architecture standards, VMware Cloud Foundation guidelines, ADR templates, and technology radar
vmware-migration-playbook.md
VCF 6.0 migration procedures, workload classification criteria, compatibility matrix, and rollback standards
vendor-partner-standards.md
Foundry partnerships (TSMC, Samsung), hyperscaler collaboration standards, and IP governance
infrastructure-inventory.md
System catalog, data center inventory, SLAs, dependency mappings, and change advisory contacts
software-licensing-framework.md
VMware, CA Technologies, and Symantec license governance, entitlement tracking, and compliance rules
hr-policies.md
Leave policies, performance review criteria, hiring procedures, and compensation guidelines
financial-reporting-standards.md
Chart of accounts, reporting cadence, audit requirements, and SOX compliance procedures
Skills
| Name | Description | Group | Source |
|---|---|---|---|
| meeting-prep | Scan calendar, gather context from connected systems, and generate briefing docs | General | Broadcom |
| weekly-report | Compile team activity summaries from Jira, Slack, email, and calendar data | General | Broadcom |
| xpu-program-tracker | Track custom silicon programs, tape-out schedules, yield metrics, and foundry milestones | AI / Silicon | Broadcom |
| ai-cluster-ops | Monitor GPU/TPU/XPU compute clusters, utilization trends, and capacity planning forecasts | AI / Silicon | Broadcom |
| chip-design-review | Design rule check automation, silicon validation reports, and post-tapeout verification | Semiconductor | Broadcom |
| incident-response | Draft or update incident response policies grounded in Broadcom security standards | Security | Broadcom |
| security-posture | Symantec endpoint and network security assessment, compliance scoring, and SASE readiness | Security | Broadcom |
| vendor-assessment | Foundry and supplier security/risk scoring against Broadcom's tiered vendor requirements | Security | Broadcom |
| vmware-migration | Plan and track VCF workload migrations with dependency mapping and compatibility validation | Software | Broadcom |
| license-optimization | Software license utilization analysis across VMware, CA Technologies, and Symantec portfolios | Software | Broadcom |
| architecture-review | Build quarterly architecture and strategy review decks from Jira and Confluence data | Architecture | Broadcom |
| change-impact | Analyze change impact across infrastructure, map dependencies, and identify affected teams | Architecture | Broadcom |
| cost-optimization | Cloud and data center spend analysis, trend visualization, and optimization recommendations | Operations | Broadcom |
| runbook-automation | Convert static runbooks into interactive step-by-step tools with automated pre-checks | Operations | Broadcom |
| budget-analysis | Pull spend data from SAP, compare to budget, flag variances, and forecast quarter-end | Finance | Broadcom |
| job-posting-draft | Draft job descriptions from role requirements, team context, and compensation guidelines | HR | Broadcom |
| onboarding-guide | Generate new hire documentation from wikis, org charts, and system access procedures | HR | Broadcom |