How to Automate ServiceNow

ServiceNow turns a large organisation's requests into records

Many large organisations run their service desks on ServiceNow's workflow platform. An outage becomes an incident, a new laptop becomes a request, and a server patch becomes a change request with an approval trail, each with an assignment group, a priority and an SLA timer.

IT Service Management is the best-known part, and service desk analysts work their queues in Service Operations Workspace. In a hospital that runs on ServiceNow, the biomedical engineering queue for broken devices can live there too.

Behind it sits a platform team: administrators and developers who build the flows, catalogue items and integrations everyone else relies on, working through a backlog of requests from every department. The change manager runs the weekly change advisory board from the same instance.

Many of the tickets that matter start outside the instance

In a hospital, the dispensing cabinet logs a controlled-substance waste with no witness signature, and the telemetry system shows a transmitter battery running low on a monitored bed. At shift change, the narcotic count does not match between two nurses. Where those systems have no connection to ServiceNow, a charge nurse phones pharmacy or biomed or writes it on the handover sheet, and someone types the ticket, if there is one, hours later.

IT has its own version. A change request is scheduled for Saturday night, three of its five approvers have not responded by Thursday, and the change manager finds out on Friday afternoon, too late to move the maintenance window or find a deputy approver.

An incident's SLA passes 75 per cent at two in the morning. The notification goes to an assignee who is asleep and a manager who reads it at nine, while the on-call group that could have picked it up is watching its phones, not the instance.

ServiceNow reaches other systems through a spoke, custom code or a robot

Workflow Studio flows start when a record is created or updated, on a schedule, from an inbound email or from an SLA task, and act on records anywhere in the instance. SLA definitions notify the assignee at 50 per cent of the target, and the assignee and their manager at 75 per cent and on breach. In Service Operations Workspace, ServiceNow's own walkthrough shows ServiceNow Otto, its agentic AI, triaging incoming incidents, writing summaries and reply suggestions, and generating resolution notes.

Reaching other systems takes Integration Hub, sold as a separate subscription: its spokes connect flows to outside services, its REST triggers need the Enterprise package, and developers can write custom integrations on it. RPA Hub runs attended or unattended software robots built in RPA Desktop Design Studio.

Every connection to a dispensing cabinet or a telemetry system therefore needs a spoke that exists, an integration someone writes or a robot someone designs and maintains, and each one waits in the platform team's backlog next to every other department's request. The 75 per cent notice goes to the assignee and their manager by default, not to the on-call group's phones.

Pharmacy and biomed work cabinet and telemetry alerts as ServiceNow tickets

A Microsoft Teams message reaches the nurse and the witness within the hour when a waste in the dispensing cabinet still lacks a witness signature, and anything still open at the end of the shift becomes a ServiceNow incident for pharmacy naming the cabinet, the drug and the time.

At each shift change, both nurses get a text to recount when narcotic counts do not match, and pharmacy receives a ticket if the gap is still there thirty minutes later. Biomed picks up low transmitter batteries as tickets with the unit and bed named, and the on-call technician gets a text for anything critical, so nobody phones the ward first.

On the IT side, the change manager sees on Telegram each morning the changes whose approvals will not be in before their planned start, with a text for the high-risk ones. The on-call group follows incidents close to their SLA target on its Trello board, updated every hour, and gets a WhatsApp message before the breach.

To ServiceNow, WebRun is one more user with roles and groups

Reading the cabinet and telemetry screens, then raising an incident in the right assignment group, is a coordinator's job today. WebRun is an AI agent that works a real Chrome browser, signed in as you. ServiceNow sees an ordinary user, so the roles and groups on that account decide what it can read and which queues it can open tickets in, and the cabinet and telemetry systems open in the same browser with the logins the hospital already issues.

Approving a change, closing an incident and editing a controlled-substance record are left to the people who own them, so its ServiceNow account needs no approval or admin role. On the hospital side the jobs start from Omnicell and Norav Medical screens, and on the IT side from the change queue and the SLA list.