Automating IT Support with Generative AI
CLIENT:
Our client, a top-tier financial services firm with over 2,000 employees and operations in 5 countries, came to us in 2025: the new, digital-focused team was overwhelmed by outdated IT support.
With an average of 1,700 support tickets hitting their desk monthly primarily from remote and hybrid workers. Response times struggled to meet internal SLAs (e.g. to fix high-priority issues within 4 hours), and support costs were rising with each new branch.
CHALLENGE:
The support team was underwater.
Routine requests — password resets, VPN issues, software installs, onboarding another new analysts — ate up entire days.
Roughly 40% of tickets just sat in the backlog longer than 48 hours, especially after-hours and over weekends.
Hiring more people wasn’t on the table: budgets were locked, and even if it was, onboarding new hires in banking compliance isn’t quick. Outsourcing was not an option because of strict data rules.
SOLUTION:
REW Technology partnered with the client in the late 2024 -early 2025 to deliver a practical, proven, and deeply integrated solution centered around generative AI for IT support — a technology we’ve seen deliver measurable ROI for mid-market leaders.
We started by mapping out how requests actually played out (e.g. every Monday there was a password reset surge and it jammed things up until midday).
Our team built and customized a chatbot inside Teams, powered by Azure’s GPT-4 (running entirely in their cloud, no data ever left their infrastructure). A chatbot allowed employees to handle common IT issues — like unlocking accounts, setting up VPNs, or requesting licences — without waiting for human intervention. Most requests were resolved in under two minutes, any time of day. We hooked the bot up to ServiceNow for ticket logging, so nothing slipped through the statistics, and used Power Automate for workflows like account unlocks or device re-provisioning.
The system kept learning after every ticket — it didn’t make the same mistake twice. Also, every action taken by the system was logged and auditable. The deployment met CIS benchmarks and complied with internal and external audit standards — critical for a finance organization
- Microsoft Teams: the employee interface for all IT interactions.
- ServiceNow: the source of truth for ticket management.
- Azure OpenAI Service: hosted securely in the client’s cloud, running GPT-4 for natural language understanding and access to internal support documentation.
- Power Automate: behind-the-scenes workflows, like access provisioning and escalation.
- Azure Active Directory: user verification and secure execution of automated actions.
RESULT:
Within three months, the bot handled more than half the tickets end-to-end — about 1,000 cases a month — with zero manual intervention. Password resets, software reinstalls, and VPN/config issues disappeared off the IT radar.
58% of tickets were resolved by the assistant, end-to-end. That’s around 1,000 cases per month.
Response times for simple issues dropped from nearly 5 hours to under 10 minutes — even in the middle of the night.
Support costs fell by 37%, allowing IT to freeze new headcount requests and reinvest savings into security initiatives.
Employee satisfaction with IT support increased by 42%, according to internal surveys.
Security incidents? Zero compliance incidents. Every interaction was logged and available for audit in ServiceNow.
All results come from real ServiceNow data and user surveys (Q2 2025). Curious how we built it? Reach out through our website, and we’ll show you how we can create the same for your company.
Share:
