The insurer’s AI program exists to grow the business, not only to cut cost. That puts pressure on every function to do more with the people it has (IT included).
A company-wide OS migration at this scale is a multi-quarter campaign. Every wave of users brings its own ticket spike: questions before the move, issues on the day and fixes afterward. The insurer needed that surge handled without burying its IT team.
High volume, low complexity. Most migration tickets fell in a narrow band (eligibility, scheduling, application compatibility, post-cutover troubleshooting), where a technician adds the least value per minute.
Senior engineers on routine work. Routine questions pulled experienced engineers away from complex incidents.
Secure, instant help without a new login. IT support touches real accounts, so employees had to be identified reliably inside the conversation, with no separate sign-in.
No rip-and-replace. ServiceNow was the core of IT service management and had to stay the system of record. Automation had to run through it, not beside it.
A workforce that isn’t always at a desk. Employees needed help in the office during work hours and away from it, on the road or after hours.
During the migration, the agent handled the high-volume questions that would otherwise have landed on IT: eligibility, scheduling and common issues after the cutover. It pulled the list of migration candidates directly from ServiceNow. Employees were identified by their corporate email address, so the agent could act on a real account without a separate login. When a case needed a person, the agent created the ServiceNow ticket itself, which kept ServiceNow as the system of record. Anything outside its scope went to a human through live chat.
The insurer chose two channels on purpose. The corporate intranet covers employees in the office during work hours, and WhatsApp covers them on mobile, on the road and after hours. The agent also answers a broad range of general IT questions beyond the migration, so when the migration ended it was transitioned into the insurer’s everyday IT helpdesk.
With the agent established in IT support, the insurer and Druid are focusing on underwriting as the next step. An AI agent will process the data provided by a policyholder, get it into the system, and flag what is still needed and the best next step for the underwriter. At a higher level, the insurer is looking to expand the scope of its AI roadmap to include external-facing agents. The opportunity is large: about 60% of inbound contact to the UK business is still non-digital, and its leadership estimates half of all contact is unnecessary from the customer’s point of view.