Practice
You are licensed for far more platform than you are running.
ServiceNow has become an AI platform with a service desk attached — agents, a data fabric, process mining, a control tower. We implement and extend the whole of it, and we engineer on it the way the rest of your organisation builds software.
- ITSM
- ITOM
- ITAM
- HRSD
- CSM
- FSM
- SecOps
- IRM
- SPM
- App Engine
- Now Assist
- AI Agent Studio
- Workflow Data Fabric
- ATF
The platform, as it stands
ServiceNow is five layers now. Most estates run one of them.
The platform your organisation bought for incident management has grown a data layer, an AI layer and an agentic layer underneath and around the workflows. Knowing where the seams are is most of the job.
Source: ServiceNow product documentation, reviewed September 2026. Availability of individual capabilities depends on your licence and family release.
What we hear
The conversations that bring people here.
We bought the platform. We are using about a fifth of it, and paying for all of it.
IT Director
Our last partner configured it by clicking. Nobody can tell me why any of it is the way it is.
Head of Service Management
The board wants AI agents on the service desk. Our CMDB is not good enough to trust one.
CIO
Workflow coverage
Where on the platform we work.
Implementation, extension and rescue across these families — on new instances, and on estates that have been in production for years and show it.
IT Service Management
Incident, problem, change, request and knowledge, plus Service Operations Workspace — with the CMDB underneath them modelled properly rather than populated once and abandoned.
IT Operations Management
Discovery, Service Mapping, Event Management and AIOps across hybrid estates, including the identification and reconciliation rules that decide whether the CMDB is trustworthy.
Asset Management
Software and hardware asset lifecycle, licence position, reclamation, and Enterprise Asset Management where the assets are not just IT.
Employee Experience
HR Service Delivery, case management, onboarding journeys and the employee portal — the fastest route to proving the platform is worth more than a service desk.
Customer & Field Service
Customer Service Management, agent workspace, Field Service dispatch and scheduling, and Sales CRM or CPQ where the platform is being asked to carry revenue processes.
Security, Risk & Compliance
Security Incident Response, vulnerability and exposure management, Integrated Risk Management, third-party risk and Privacy Management — with evidence collected continuously rather than reassembled each quarter.
Strategic Portfolio Management
Demand, roadmaps, resource and financials, so the platform reports on what IT is spending its year doing as well as what broke today.
Workplace & Legal
Workplace Service Delivery and Legal Service Delivery — desk booking, facilities and legal intake, for organisations extending the platform past IT and HR.
Custom applications
Scoped applications for the processes no module covers — built to survive a family upgrade rather than to be rebuilt after one.
The AI layer
The part of the platform nobody has a five-year methodology for.
ServiceNow now ships generative assistance, autonomous agents, an agent build environment, an orchestrator and a governance layer. This is recent, it changes with every family release, and it is where the value in your licence currently sits unclaimed.
Now Assist
- Case and incident summarisation
- Generated resolution notes
- Knowledge article drafting from resolved cases
- Code generation inside the platform
- Rolled out where handling time is measured before and after
AI Agents & Studio
- Agents built against your own processes
- Tools and APIs an agent may call, defined explicitly
- Guardrails and escalation paths to a human
- Evaluation before anything is allowed to act
- Agent Orchestrator for work that crosses teams
Data & grounding
- Workflow Data Fabric for data that lives outside the platform
- CMDB and CSDM quality work first — agents inherit your data problems
- Knowledge base curation and retirement
- Process Mining to find where the work actually goes
Governance
- AI Control Tower: which agents exist and what they touch
- Permissions, scoping and data boundaries per agent
- Audit trail your risk function will accept
- ServiceNow Vault where data sensitivity requires it
What we will tell you not to do
- Do not put agents on a process you have not mined. Automating a broken path makes it faster, not better. Process Mining first, then decide what deserves an agent.
- Do not let an agent act on an untrustworthy CMDB. If reconciliation is not sound, an autonomous action is a production incident waiting for a trigger.
- Start with assistance, not autonomy. Summarisation and drafting earn trust and produce the measurement baseline you will need before anyone approves an agent that acts.
- Do not buy capability you have no owner for. Every agent needs someone accountable for what it does. If that person does not exist, neither should the agent.
Technical capability
What our engineers actually do on the platform.
The specifics, for the architect who will be reviewing whoever you shortlist.
Application development
Scoped applications and App Engine, Flow Designer and Workflow Studio, UI Builder and Next Experience, Service Portal widgets, business rules, script includes, GlideAjax, ACLs, roles and domain separation.
Connecting the estate
IntegrationHub spokes and custom actions, scripted REST APIs and inbound actions, MID Server topology and credentials, import sets and transform maps, event and webhook ingestion, SSO and provisioning via Entra ID or Okta.
CMDB & reporting
CSDM alignment and class modelling, Discovery patterns, Service Mapping, identification and reconciliation engine rules, de-duplication, data certification, Performance Analytics, table rotation and archiving.
Keeping it healthy
ATF suites and regression coverage, instance performance and query tuning, clone schedules and data preservers, family-release upgrade planning, technical debt review, managed administration and HI case handling.
Engagement shapes
What people actually call us in for.
| Situation | What we do | Typical shape |
|---|---|---|
| New implementation | Stand up ITSM, HRSD, CSM or SecOps on a clean instance, configured to how your teams work rather than to the demo, with the CMDB and integrations designed up front. | Fixed-price phases |
| Stalled rollout | Take over an implementation that has slipped or been abandoned. We audit what exists, keep what is sound, and give you an honest account of what has to be rebuilt. | Two-week audit, then phases |
| Upgrade backlog | Clear a multi-release gap: customisation inventory, ATF coverage written, skipped releases planned and taken in sequence. | Fixed-price per release |
| Custom application | A scoped app for a process no module covers — built in source control, tested, and documented so your own admins can extend it. | Fixed-price milestone |
| AI enablement | Process mining, data readiness, then Now Assist and agents introduced where there is a measured baseline and a named owner. | Assessment, then pilot |
| Managed support | Ongoing administration, enhancement backlog, instance health and upgrade cadence, with a named engineer rather than a queue. | Monthly |
Selected work
What changed, in the client’s terms.
Three recent engagements from this practice.
Untangling a complex IT environment
A complex IT environment on legacy ITSM tools that could not meet growing demand. Incidents, changes and problems were handled largely by hand, consuming a significant share of the IT team’s time. Read the case study→ Private cloud · Billing integration
Charging for what customers actually consume
The provider needed to bill on actual consumption — CPU, memory, storage and bandwidth — and the systems holding that consumption data were not the systems producing invoices. Read the case study→ Healthcare · Systems integration
EMR, HIS and PM talking to each other
The provider ran an Electronic Medical Record system, a Hospital Information System and a Practice Management system that exchanged nothing. Staff moved data between them by hand. Read the case study→A two-week paid pilot
Fixed price, on a real piece of your backlog. You keep the output whether or not you continue.