Case 01
Building IT's first real platform, against a Zendesk renewal clock.
A six-month greenfield rollout for an organisation that had outgrown what its existing tooling could carry. The constraint was real: a renewal deadline, no internal platform experience, and a single phase to land three modules at once.
The organisation had outgrown what its existing tooling could carry. Service desk lived in Zendesk, asset records lived in three different places, and IT was operating mostly out of inboxes and spreadsheets. The Zendesk contract was up for renewal, the upgrade path didn't justify the cost, and leadership wanted to do something real instead.
There was no one internally who'd implemented ServiceNow before. I came in as architect and delivery lead, the only person on the client side with platform depth. We brought in a build partner to add hours, but the architecture, the design decisions, and the technical ownership were mine. Under six months. A single phase. ITSM, Hardware Asset Management, and Software Asset Management standing up together.
The build
I designed the CMDB class model and identification rules from scratch, which meant careful work upfront defining what a CI actually was inside this organisation and how Discovery should be allowed to dedupe records. Process design across Incident, Request, and Change went the same way: written down, walked through with the people who'd live in it, and pulled close to platform defaults wherever I could defend it.
I built the integrations into Active Directory, telephony, and the endpoint and procurement tools. Where the business genuinely needed something the platform didn't offer, I wrote scoped applications and custom scripts rather than warping the core. Discovery I tuned myself, including the MID-server topology and schedules. The HAM lifecycle workflows (procure, deploy, retire) and the SAM reconciliation logic matching installs against entitlements were both my designs and my builds.
The pushback
The hardest architectural call came mid-project, when there was real pressure to recreate Zendesk-style workflows inside ServiceNow because they felt familiar. I pushed back. We'd be paying for the comfort every release for years, and we'd be teaching the team to fight the platform instead of use it.
I won the argument by walking leadership through the recurring cost of customisation in plain numbers, not platform jargon. We kept the flows out-of-the-box. That decision is what made the six-month timeline survivable. It's also what made the system durable after I rolled off.
The harder work was operational
Up to that point, IT had been organised by site. Each office had its own queue, its own conventions, its own people. The platform forced a different shape. Once everything lived in one system, the support function reorganised itself as a single national team working a shared queue.
I led the change management that got them there: rebuilding the operating model, retraining the team on a new way of dispatching work, and porting the Zendesk history they actually needed. The migration itself was tedious but tractable. The cultural shift was where the real work sat.
What changed
ServiceNow became the single front door for IT across the organisation. Incidents, requests, hardware, software — one place, one process, one source of truth.
MTTR dropped to between two and four hours, mostly because tickets stopped being trapped behind the wall of whoever happened to be sitting in that office that day. Software licensing moved from quarterly fire drills to a managed practice. The IT function went from reactive and fragmented to something the rest of the business could actually rely on.
The platform is still running. The customisations I refused on day one are still not there.