The power of AI: how quiet victories in IT support drive big impact, too

Scanning the artificial intelligence landscape, you can find some pretty revolutionary developments. In biotechnology, for instance, researchers are using AI to predict protein structures, accelerating the design of newer, more effective medicines.

There are big insurance advances, too. At Swiss Re, we've created Gen AI tools to help identify claims fraud and third-party recovery opportunities as well as to streamline life insurance underwriting to help our clients accelerate complex coverage decisions.

While I'm fascinated by big tech-enabled breakthroughs like these, some of AI's most meaningful impacts will also come when this technology is applied to help us manage routine tasks that consume our time and resources every day. In other words, not everything will be an AI-powered revolution. There will be many smaller AI victories, too, to lift productivity, make us more efficient, and reduce operational costs.

One project within Swiss Re's Group Digital & Technology Organisation fits well into this category: we built and deployed a new, Gen AI-powered analysis tool to help better organise IT support requests from Swiss Re employees who use hundreds of business applications across the company to do their jobs and serve our clients.

Swiss Re's business applications range from internal learning tools to the critical platforms we rely on to process claims, underwrite risks, and complete billion-dollar transactions - in short, our core processes. Much like at any company, our employees who use these applications occasionally have questions or encounter glitches. It's just part of doing business.

When such a situation arises, employees can raise a ticket with Swiss Re's internal business application support teams to get help, so they can quickly get back to work. Their tickets are assigned to a support team based on their complexity. Sometimes, those submitting requests can solve an issue on their own with just a little guidance; in other instances, expert IT help is needed.

About a year ago, however, as we were developing a new IT support framework, we discovered potential for improvement in how we assign tickets to make the most efficient use of our experts' valuable time. This has real business implications, because the more expert support a ticket gets, the more expensive it becomes. Managing tickets effectively, and routing them to the correct support tier, helps control costs and boost operational efficiency.

For our analysis, there was a lot of data to crunch. In 2024 alone, we had 267,000 support tickets across Swiss Re, each with structured and unstructured data relevant to understanding the ticket and determining how it should be resolved. A human looking for meaningful patterns to help fine-tune our support processes going forward would need a full year to review just 10,000 tickets, making manual analysis impractical.

Instead, our tech team turned to Gen AI technology to build a faster, more efficient analysis tool. Incidentally, we did it at a fraction of the cost and half the time that an external vendor told us would be required for such a project. Critically, our tool was also scalable, equally at home analysing support tickets regardless of whether they had originated in Swiss Re Corporate Solutions, our P&C Re and L&H Re Business Units, Group functions or elsewhere across the company.

Ultimately, what we discovered was eye-opening: in some instances, up to a quarter of the IT tickets potentially could have been "shifted left" – in other words, moved from higher-level support tiers to lower tiers where they could be resolved either by the employees who had submitted them or by less-technical support teams.

Our AI tool is quite sophisticated, as well, employing multiple Gen AI agents, each with its own distinct objective. One agent evaluated each ticket to make a quick determination about the appropriate support tier; another clustered tickets into groups of similar requests, easing comparison and analysis.

Meanwhile, a third agent extracted key ticket information to be used for developing new support materials. These can be put to work to train IT teams to handle challenging future tickets or to create intuitive, easy-to-understand texts aimed at enabling more employees to independently resolve their own tickets, whenever possible.

Additionally, an interactive dashboard allows users to visualise how past tickets had been categorised, which helped them better understand opportunities to boost efficiency. For our business application IT support teams, the resulting insights are now enabling them to focus on higher-value tasks without losing valuable time on simpler tickets.

The bottom line? Our IT support teams can now deliver faster, more effective, less expensive solutions, helping Swiss Re accomplish its business objectives: providing risk protection and risk knowledge to our clients around the world. We are now leveraging the results of our analysis to drive continuous support improvements across all of our IT domains.

For me, this is a great example of deploying technology not for technology's sake, but to enhance business-critical internal processes and improve how we work in ways that add up to real operational impact. It also shows that in a world dazzled by AI’s revolutionary breakthroughs, its quieter wins are proving to be very powerful, too.