A free AI Skill Bundle from the ControlUp Innovation Guild is now available to everyone, whether you’re a ControlUp customer or not.
Your service desk already contains a huge amount of information about what is slowing users down, consuming IT time, and repeatedly going wrong. The challenge is finding it.
Thousands of incidents might be sitting in ServiceNow, Jira Service Management, Zendesk, Freshservice, or another ITSM platform. Traditional reporting can tell you how many tickets were opened and how quickly they were closed.
But the more interesting questions are harder to answer:
- What are users really struggling with?
- Which problems keep coming back?
- What could have been prevented?
- And where should IT automate first?
That is what the new AI Ticket Analyzer from the ControlUp Innovation Guild (CIG) is designed to uncover.
Meet the AI Ticket Analyzer
The AI Ticket Analyzer is an AI Skill Bundle that teaches AI assistants such as ChatGPT and Claude how to perform a structured analysis of your service desk data.
Instead of trying to create the perfect prompt yourself, the skill provides the AI with a repeatable methodology for preparing, analyzing, and presenting ticket data.
Simply export your incidents as CSV, point your AI assistant at the file, and ask it to analyze your tickets.
Within minutes, you get a professional, self-contained HTML report designed for both technical teams and IT leadership.
And importantly, no ControlUp deployment is required.
Getting Started
All you need is the AI Skill Bundle and an export of your service desk tickets.
Step 1: Install the Skill
Claude Desktop
Open:
Settings → Capabilities → Skills
Upload the downloaded Skill Bundle ZIP and enable code execution if prompted.
Claude Code
Extract the skill into:
~/.claude/skills/
The bundle can also be used with supported ChatGPT Skill environments.
Step 2: Export Your Tickets
Export the incidents you want to analyze as CSV from:
- ServiceNow
- Jira Service Management
- Zendesk
- Freshservice
- Or another service desk platform
We recommend starting with a useful analysis window, such as the previous 90 days.
Make sure the export includes a narrative field such as short_description or description. This allows the AI to understand what the tickets are actually about.
If your platform exports .xlsx, save the file as CSV UTF-8 first.
Step 3: Run the Analysis
Point your AI assistant at the export.
For example:
Analyze ./incidents-last-90-days.csv with the ticket-insights skill.
For best results, we recommend Claude Opus 5 with Medium effort as a minimum. Higher effort may provide an even deeper analysis, particularly with larger or more complex datasets.
The skill handles the rest.
More Than an AI Summary
You could upload a CSV file to an AI assistant today and simply ask, “What do you see?” But the results can vary dramatically depending on your prompt.
The AI Ticket Analyzer gives the model a specific methodology and looks for:
- Where ticket volume is concentrated
- Recurring problems hidden across different categories
- Issues that could potentially have been prevented
- Resolution, reassignment, and intake patterns
- High-impact automation opportunities
- Practical recommendations for where IT should focus next
It deliberately looks beyond application names and ticket categories to identify the underlying problem.
Password resets, account lockouts, and “can’t sign in” incidents might actually represent one larger identity issue. Similarly, slow devices, memory pressure, full disks, crashes, and overheating could reveal a broader endpoint-health problem.
That matters because the goal isn’t just to understand your ticket volume. It’s to understand what is driving it.
Find Work That Shouldn’t Be a Ticket
Resolving tickets faster is useful. Preventing them from being created is better.
The analyzer identifies incident types where monitoring, self-service, configuration changes, policy improvements, or automation could potentially remove demand from the service desk entirely. It then turns those findings into a ranked list of automation opportunities.
Each recommendation includes an estimated level of effort: Quick win · Project · Program
And a confidence level: High · Medium · Low
This gives IT teams a starting point for an evidence-based automation backlog, based on their own ticket history, which is way more useful than a generic list of AI suggestions.
What Does the Report Look Like?
The final output is a standalone HTML report that can be opened in any browser or printed directly to PDF.
It includes ticket statistics, top categories, recurring themes, preventable issues, automation opportunities, and ticket intake channels.
In our example dataset of 1,500 tickets, the analyzer identified nine recurring themes and a 37.9% reassignment rate. But the interesting part wasn’t the number…
Instead of simply reporting that endpoint-performance incidents represented a large share of ticket volume, the analyzer connected slow machines, freezes, memory issues, full disks, application crashes, and overheating into a broader problem: Endpoints degrading in place.
That creates a very different conversation between the service desk, endpoint engineering, automation teams, and IT leadership.
What Happens Behind the Scenes?
The AI Ticket Analyzer does more than feed a giant CSV directly into the model. A bundled preparation script automatically handles common service desk export quirks such as alternative delimiters and encodings and turns the source data into a compact dataset for analysis. The AI then works through the ticket patterns in a structured pass and generates the analysis. Finally, the bundle renders everything into a polished, self-contained report.html.
No external reporting platform is required to view the result.
For administrators running the scripts directly, Node.js 16 or newer is required and there are no npm dependencies.
What About My Ticket Data?
Your ticket export is not sent to ControlUp. The Skill Bundle operates on the files you provide, and ControlUp does not receive your ticket data.
Your chosen AI assistant will process the information according to that provider’s account, security, and data-processing configuration, so organizations should follow their normal policies for using AI with service desk information.
The current version also does not perform PII scrubbing. If your ticket descriptions contain usernames, email addresses, hostnames, customer details, or other sensitive information, anonymize the export first if required by your organization’s policies.
Built by the ControlUp Innovation Guild
The ControlUp Innovation Guild explores new ideas and technologies that can help IT teams solve real operational problems.
AI Skills give us an exciting new way to do that.
Instead of building every idea into a traditional application, we can package knowledge, analysis methods, scripts, and workflows into something your AI assistant can immediately understand and use.
The AI Ticket Analyzer is one of our first releases using this approach, and we’re making it available to everyone, whether you’re a ControlUp customer or not. Because you don’t need another monitoring agent to discover something useful in the tickets you already have.
You already own the data. Now you can ask better questions of it.
Try the AI Ticket Analyzer
Download the ControlUp AI Ticket Analyzer Skill Bundle, export your service desk incidents, and give it a try.
One CSV. One AI Skill. A completely new view of your service desk.
Download the AI Ticket Analyzer →
