The data quality is sufficient for AI, with key fields like 'assignment_group' and 'priority' well-populated and showing strong predictive signals. However, the high process fragmentation and severe rework indicate low process maturity, suggesting change management will be a key factor for success.
AI Potential: High · Moderate confidence
AI Readiness: Medium
The data quality is sufficient for AI, with key fields like 'assignment_group' and 'priority' well-populated and showing strong predictive signals. However, the high process fragmentation and severe rework indicate low process maturity, suggesting change management will be a key factor for success.
ServiceNow capabilities
Now Assist
Predictive Intelligence
Virtual Agent
Flow Designer
Automation Engine
Assignment Rules
SLA/Performance Analytics
AI Readiness: MediumThe data quality is sufficient for AI, with key fields like 'assignment_group' and 'priority' well-populated and showing strong predictive signals. However, the high process fragmentation and severe rework indicate low process maturity, suggesting change management will be a key factor for success.
🎯
AI Opportunities
Opportunity 1
Automate Incident Routing to Reduce Reassignment and SLA Breaches
CriticalImpact: Very HighEffort: Medium✓ High confidenceRouting
Predictive Intelligence
Problem signal
The high volume of 'Assigned - Active' rework transitions (712 occurrences) and the 67% SLA breach rate indicate that incidents are frequently routed to the wrong teams, causing significant delays.
How Predictive Intelligence helps
Predictive Intelligence analyzes historical incident data (like short descriptions and categories) to predict and automatically populate the correct assignment group, ensuring incidents reach the right team the first time.
Trigger
Task created
Expected benefit
Faster assignment, reduced time to resolution, and a significant reduction in SLA breaches by eliminating manual triage delays and misrouting.
Owner: ServiceNow Platform Owner, Incident Process Owner
Config: Create a classification model in Predictive Intelligence to predict 'assignment_group' based on fields like 'short_description' and 'contact_type', then trigger it with a Business Rule on incident creation.
Evidence
Rework transition 'Assigned - Active' occurred 712 times
67% of incidents with SLAs have breached
Assignment_group has 46 distinct values, indicating a complex routing environment
Opportunity 2
Proactively Manage and Escalate Incidents at Risk of SLA Breach
HighImpact: HighEffort: Medium✓ High confidenceRisk Detection
Predictive Intelligence
Problem signal
An extremely high 67% of sampled incidents have breached their SLAs. This reactive approach to SLA management puts service delivery at risk and indicates a need for proactive monitoring.
How Predictive Intelligence helps
Predictive Intelligence can create a model to predict the likelihood of an incident breaching its SLA. This prediction can be used to drive proactive actions, such as alerting managers or escalating priority before the breach occurs.
Trigger
Task created or updated
Expected benefit
Improved SLA performance, reduced risk of critical service failures, and increased customer satisfaction by focusing attention on the incidents that need it most.
Owner: Incident Manager, Service Delivery Manager
Config: Configure a regression model in Predictive Intelligence to predict 'Time to SLA breach', then use Flow Designer to trigger notifications or escalations when the prediction falls below a defined threshold.
Evidence
67% of sampled items breached an SLA
Top SLAs breached include Priority 1, 2, and 3 response and resolution times
Opportunity 3
Accelerate Resolution with AI-Powered Incident Summarization
HighImpact: MediumEffort: Low✓ High confidenceSummarisation
Now Assist
Problem signal
The process has a high number of variants (142) and an average of 6.9 transitions per incident. This complexity forces agents to spend valuable time reading through long activity streams and work notes to understand the incident context, especially after a reassignment.
How Now Assist helps
Now Assist automatically generates concise summaries of incident history, including user interactions and previous actions. This allows agents to quickly understand the current situation without manually reviewing the entire record.
Trigger
Agent opens an Incident form
Expected benefit
Reduced agent ramp-up time on complex or reassigned incidents, leading to faster diagnosis, fewer errors from missed context, and improved resolution times.
Owner: IT Service Desk Manager
Config: Enable the 'Incident Summarization' capability for Now Assist within the Now Assist Admin console and add the summary component to the Incident form layout.
Evidence
142 process variants discovered
Average of 6.9 transitions per incident
High volume of rework transitions like 'Assigned - Active' (712 occurrences)
Opportunity 4
Reduce Information-Gathering Delays with AI-Assisted Communication
MediumImpact: MediumEffort: Low✓ High confidenceKnowledge Assist
Now Assist
Problem signal
The most frequent rework loop is 'Work in Progress -> Pending User' (860 occurrences), indicating agents regularly need more information from users. This back-and-forth introduces significant delays and extends resolution times.
How Now Assist helps
Now Assist can help agents draft clearer and more comprehensive requests for information from users. It can also generate resolution notes, ensuring that the solution is clearly communicated and documented, reducing the likelihood of reopening.
Trigger
Agent composes a work note or user communication
Expected benefit
Fewer clarification cycles with users, reduced time spent in 'Pending' states, and higher quality of resolution notes, improving overall process efficiency and customer satisfaction.
Owner: IT Service Desk Manager, Knowledge Manager
Config: Activate the work notes and resolution notes generation skills for Now Assist in the Incident Management application.
Evidence
Top rework transition is 'Work in Progress -> Pending User' with 860 occurrences
High overall rework rate of 59.13%
Opportunity 5
Improve Intake Quality and Deflect Common Incidents with a Virtual Agent
MediumImpact: HighEffort: High✓ High confidenceTriage
Virtual Agent
Problem signal
A significant portion of incidents are created via manual channels like Phone (12.9%) and Email (12.3%). These channels often lead to incomplete information, triggering the 'Work in Progress -> Pending User' rework loop (860 occurrences) as agents seek clarification.
How Virtual Agent helps
Virtual Agent provides a guided, conversational intake process that ensures all necessary information is collected upfront. It can also deflect common, simple incidents by providing users with knowledge articles or automated solutions, freeing up agent capacity.
Trigger
User initiates a chat from the service portal or other channel
Expected benefit
Better quality incident data from the start, reduced rework cycles, higher agent productivity due to lower volume of simple incidents, and an improved user experience.
Owner: Digital Experience Owner, Service Portal Manager
Config: Design and publish Virtual Agent conversation topics in VA Designer for the most common incident types, integrating with knowledge bases and fulfillment flows.
Evidence
Phone contact type accounts for 12.9% of incidents
Email contact type accounts for 12.3% of incidents
Top rework transition 'Work in Progress -> Pending User' indicates poor initial data capture
⋯
Opportunity Themes
Intelligent Triage and Risk Management
These opportunities focus on using Predictive Intelligence to automate routing and proactively identify incidents at risk of breaching SLAs, addressing the core drivers of delay and inefficiency.
Automate Incident Routing to Reduce Reassignment and SLA BreachesProactively Manage and Escalate Incidents at Risk of SLA Breach
Agent Productivity and Experience
This theme leverages Now Assist to reduce the manual effort for agents in understanding incident history and communicating with users, directly targeting process complexity and rework loops.
Accelerate Resolution with AI-Powered Incident SummarizationReduce Information-Gathering Delays with AI-Assisted Communication
Enhanced User Self-Service and Intake
This theme focuses on improving the front-end of the process by using Virtual Agent to guide users, ensure complete information is captured at intake, and deflect common requests.
Improve Intake Quality and Deflect Common Incidents with a Virtual Agent
→
Implementation Sequence
1
Phase 1: Foundational Agent Assistance No build required
Empower agents with immediate, low-effort AI tools to cope with existing process complexity and establish better performance visibility.
Now Assist
SLA/Performance Analytics
Deploy Now Assist for incident summarization to reduce agent research time.
Enable Now Assist for drafting user communications and resolution notes to tackle rework.
Build Performance Analytics dashboards to monitor rework and SLA breach trends.
2
Phase 2: Predictive Triage and Routing
Automate core triage and assignment decisions to fix the root cause of reassignments and SLA breaches.
Predictive Intelligence
Flow Designer
Train and activate Predictive Intelligence for automated assignment group routing.
Implement the Predictive Intelligence model for SLA breach prediction.
Create a Flow Designer flow to alert teams about incidents with a high risk of breaching SLA.
3
Phase 3: Conversational Self-Service
Shift demand to automated, self-service channels to improve intake data quality and deflect routine incidents.
Virtual Agent
Develop and launch Virtual Agent conversations for the top 5-10 incident types currently handled via phone and email.
Integrate Virtual Agent with the knowledge base to increase incident deflection.
◎
Target Workflow State
The future-state workflow uses AI at the point of creation to ensure incidents are correctly classified and routed. Agents are supported by AI-powered summaries and communication tools, while proactive monitoring prevents SLA breaches. The result is a faster, more accurate, and less manual incident resolution process.
Preferred flow
New
→
In Progress
→
Resolved
→
Closed
New
Open
Capture, classify, and route the new incident.
Enter: Incident is created.
Exit: Incident is assigned to a group and resolver.
Predictive Intelligence: Predicts and sets the optimal Assignment Group based on incident details. Virtual Agent ensures complete data capture if created via chat.
In Progress
Active
The assigned agent investigates and works on resolving the incident.
Enter: An agent accepts or is assigned the incident.
Exit: A resolution is identified and applied.
Now Assist: Provides an on-demand summary of the incident. Predictive Intelligence continually assesses SLA breach risk in the background.
Awaiting Information
Waiting
Awaiting input from the user or another system.
Enter: Agent determines external information is required.
Exit: The required information is provided.
Now Assist / Flow Designer: Now Assist helps the agent draft a clear request for information. Flow Designer can send automated reminders to the user after a set period.
Resolved
Review
Resolution has been applied and is awaiting confirmation from the user.
Enter: Agent applies a fix and documents the resolution.
Exit: User confirms resolution or a timer elapses.
Now Assist: Generates proposed resolution notes for the agent to review and save.