Aderant builds intelligent ticket triage with Amazon Nova
Amazon Web Services Angela Mapes
Aderant used Amazon Nova Lite to triage support tickets automatically. It’s routing about 96% right, and the math says it saves engineers hours each week.
Based on reporting by Amazon Web Services, Angela Mapes — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Aderant has pushed its support workflow a step beyond search and into scheduled automation. The legal-software vendor built an Intelligent Ticket Analyzer with Amazon Nova Lite through Amazon Bedrock, and it now helps sort, enrich, and route support tickets for the company’s 38-person SierraOps team across 268 client environments.
The system runs on weekday hourly cycles. It pulls unassigned tickets from Jira, gathers context from Jira, Confluence, Amazon Athena, and Microsoft SharePoint, then asks Nova Lite for a structured classification and a suggested next step. If confidence is high enough, it can reroute the ticket and send out the right notifications on its own. If not, it hands the decision back to a human.
That guardrail matters. Before automation, each ticket took 15–25 minutes of manual investigation before real work even began, and the team handled about 34–40 tickets a week. Aderant says the new setup reviewed 109 tickets in its first 2.5 weeks in production, from June 30 through July 17, 2026, and got routing right about 96% of the time. Based on the old baseline, the company estimates a gain of 8–14 engineering hours a week, with total operating cost under $30 a month and Bedrock inference cost under $1.
This was not a blind leap. Aderant says it started with Jira automation, then used its earlier Amazon Quick CloudOps Helper for AI-assisted validation, then built the production workflow with AWS Lambda, Amazon EventBridge, Amazon DynamoDB, CloudWatch, and Kiro support for scaffolding and tests. The team also ran the analyzer against live tickets for about two weeks before launch, comparing its output with manual decisions and tightening the prompts and routing logic.
The payoff is less about replacing engineers than clearing the sludge in front of them. The analyzer can post procedures for after-hours work, surface commands and troubleshooting steps for newer engineers, and track recurring patterns so future tickets can be linked to a known fix. Aderant says that leaves more time for complex troubleshooting, infrastructure improvements, and the kind of judgment call software still can’t fake very well.
My take — AI-written commentary, not fact-checked reporting
This is the sort of AI deployment that actually makes sense: boring, narrow, measurable, and boxed in by human review. The industry keeps selling grand autonomy, but the real win is shaving 15 minutes off a ticket without turning support into a casino. Also, under $30 a month is a much better pitch than another glossy “platform vision” deck.
Read more about this at: Amazon Web Services
