A global CMDB and Discovery rollout at Dun & Bradstreet
Turning infrastructure data nobody trusted into the organization's first reliable, automated global inventory.
- Problem
- Infrastructure data spread across teams and tools, with no consistent automated discovery, no reliable CI identification, and fragmented ownership. Incident impact analysis was weak because the relationships simply weren't there.
- My role
- ServiceNow CMDB & Discovery Manager, 2023–2026. I led the rollout end to end.
- Key decisions
- Standardized credential patterns; a distributed MID Server design; governed discovery patterns and classification; strengthened identification and reconciliation rules; an ownership and certification cadence.
- Outcome
- The first reliable automated global infrastructure inventory. By my measure at the time, a 98% increase in CMDB accuracy and completeness.
The problem
Dun & Bradstreet had no consistent automated process for discovering infrastructure, identifying configuration items, or governing their lifecycle. The data existed, spread across teams and tools, but nobody could say with confidence what was actually running or who owned it. That shows up everywhere downstream: weak incident impact analysis, low confidence in change planning, and arguments about ownership.
What I designed
- Discovery coverage. Extended discovery into previously unmanaged infrastructure zones, working through network and platform dependencies with the teams that owned them.
- Credential architecture. Standardized credential patterns across Linux, Windows, network and API discovery — which is where most discovery failures actually come from.
- MID Server topology. A distributed design that balanced discovery load and held up across network segments.
- Pattern and classification governance. Tuned patterns so discovered components landed in the correct CI classes consistently, not case by case.
- Identification and reconciliation. Stronger identification rules to cut duplicate CIs and raise confidence across data sources.
- Ownership and certification. CI ownership expectations, a certification cadence, and reporting guardrails for compliance and operational health.
Alongside the build, I chaired a monthly CMDB advisory board with product stakeholders and sat on ServiceNow's product advisory for Discovery and CMDB.
The result
The organization's first reliable, automated global infrastructure inventory. The lasting outcomes were better incident impact analysis, more confident change planning, and cross-team agreement about who owns what.
Why it matters more now
The same foundation is what AI agents act on. ServiceNow's own 2026 AI releases describe the CMDB as what grounds their agents' decisions — the AI Control Tower is "anchored by the ServiceNow CMDB." An agent about to restart a server needs to know what depends on it and who owns it. That is a CMDB question, and the answer is only as good as the discovery, identification and ownership work underneath it.
An earlier, ServiceNow-focused version of my resume is also available: ServiceNow resume (PDF).