We’re debating the best approach for routing 500+ daily support cases in our ServiceNow case management implementation. Currently using manual triage by L1 team, but considering full automation with assignment rules. Our cases span IT, HR, and Facilities with varying complexity levels.
The automation advocates argue for faster routing and reduced L1 workload. Manual routing supporters emphasize accuracy and context understanding that automated rules might miss. We’re seeing 15-minute average routing time manually with 92% accuracy. Pilot automation showed 30-second routing but 78% accuracy requiring reassignments.
Key concerns: SLA clock starts immediately, so routing errors impact metrics significantly. Complex cases need judgment calls that rules struggle with. However, manual routing creates bottlenecks during high-volume periods and depends heavily on experienced triagers.
What’s your experience with automated vs manual case routing at scale? How do you balance speed with accuracy? Are hybrid approaches effective?
Consider your staffing model too. Manual routing requires experienced triagers who understand all service areas - that’s expensive and risky if they leave. Automation requires upfront investment in rules and ongoing maintenance. We use automated routing with a 15-minute review window: cases route automatically but a senior analyst reviews the queue and can reassign within 15 minutes before SLA clock truly starts. This catches obvious errors without losing the speed benefit. After six months, we reduced manual interventions to under 5% as rules improved.
The misroutes are mainly cross-functional cases that touch multiple departments or cases with insufficient initial information. For example, a laptop issue that turns out to be a VPN problem requiring network team involvement. Our categorization at intake is limited by what end users provide. How do you handle cases that don’t fit clean categories? Do you route to a default queue or force better intake categorization?
This is where your intake form design matters enormously. Implement conditional questions that appear based on initial selections - this improves categorization quality before routing even happens. For cross-functional cases, create an ‘Assessment Team’ queue that handles initial investigation for ambiguous cases, then routes to appropriate teams. This hybrid approach gives you automation speed for clear cases while maintaining accuracy for complex ones. Your SLA metrics should account for assessment time separately from resolution time.
The 78% accuracy you’re seeing suggests your assignment rules need refinement, not that automation is wrong. I’d recommend analyzing the 22% misroutes - are they specific categories? Invest time in building better categorization logic and using machine learning classification if available in your ServiceNow version. We achieved 94% automated routing accuracy by spending three months tuning rules with historical case data. The initial accuracy hit is normal but shouldn’t be accepted as final state.