Is AI creating a new generation of “Legacy Trap” solutions?

By Tom King

This has exploded with AI tools that can help non-technical people generate very sophisticated solutions to satisfy both simple and complex requirements, all from simple text descriptions of the functionality desired. While this is creating amazing capabilities in a variety of fields, it also creates a significant amount of risk for organizations. One of the primary risks is falling into a new variant of the “Legacy Trap” that insurance has only recently been avoiding fairly successfully.

First, let's define what a “Legacy Trap” is.

Legacy Trap: When an insurer invests in a tech solution (built or bought) that can’t adapt to advances in technology or changes in business strategy—eventually requiring costly rework or replacement.

Insurers fall into the Legacy Trap in one of two ways.

1: Self Build - insurance is famous for building its own solutions, especially up through the end of the 1990’s. These solutions may have met all of the business requirements at the time they were put into use but insurers usually lacked the appetite for investment to upgrade and reinvent these solutions as technologies emerged and evolved. That is why there are still many mainframe and COBOL solutions out there today.

2: Products without upgrade paths - this doesn’t happen as much anymore, but it was not unusual for some commercial products not to have a path to upgrade technology. Or, in the case of on-prem solutions, insurers may not take an upgrade due to the amount of cost associated with the upgrade process or licensing issues. Many insurers are still running legacy systems out there where they have no relationship with the company they bought the solution from, if the company even still exists any longer.

In the past, the biggest competitor to a vendor’s software sale opportunity was the self build, not the other competitors in the field. This has shifted tremendously. Most insurers today see the value of choosing a software partner that not only supports their business requirements but has an underlying platform that addresses security, scalability and, most importantly from a Legacy Trap perspective, invests in keeping the underlying platform up-to-date on the most relevant technologies. By investing in the right partner, insurers push the need to keep technology fresh onto the shoulders of the vendor, reducing the insurers overall operational risk and cost. SAAS solutions make this process of staying technically relevant even easier - the system just refreshes with the latest technology underneath it without the client having to do anything. Salesforce, for example, has three major releases a year, constantly upgrading functionality and underlying technology without a long, laborious and expensive upgrade cycle. Insurers don't have to worry about whether the technology is there, they just have to decide how to use it.

Now, with AI system generators, there’s a temptation for tech teams to build proprietary AI solutions from scratch. While innovation is good, this can create a new form of Legacy Trap for insurers to fall into. An AI solution that works today could become obsolete quickly if it isn’t built with adaptability in mind, if it relies on niche technology that isn’t widely supported or if there isn't a systemic investment and infrastructure plan in place to keep the solution relevant. In addition, IT departments have always been fighting with “shadow IT organizations” developing point solutions that aren’t in line with existing IT development and security guardrails. Insurers have always struggled to create solution environments that can work seamlessly across the organization. A spate of homegrown AI developed point solutions that are developed independent of the company’s IT strategy might seem like a good idea at the time, but it puts that effort to create a more seamless organization at risk. In essence, just because you can do something doesn’t mean you should.

Commercial-grade software platforms are already incorporating AI features or enable working with AI toolsets to speed development and broaden functionality. At PS Advisory we routinely use AI to generate code as part of client implementations. This speeds development, and enables us to create a quality product for clients at a more reasonable cost WITHOUT sacrificing quality, security or skipping important longevity/safety issues.

Insurers can leverage AI for development without falling into the new AI Legacy Trap. Keeping in line with the trend to invest in industrial strength, commercial products instead of developing from scratch, Insurers should focus on partners that will help insurers take advantage of AI development within the products being considered for implementation. The idea is to stay on the technology curve with the help of partners, instead of creating an AI application that could quickly become the next legacy system that the team doesn’t have resources, funds or time to update.

PS Advisory focuses only on insurance and specializes in Salesforce implementations. If you would like to understand how you can avoid the AI Legacy Trap but still effectively leverage AI in both your operations and your development processes, let’s talk.

Email: Info@psadvisory.com

About the author:

Tom King has 25+ years of experience in the insurance industry, both as an insurer and as a consultant to the industry. Much of his career has been dedicated to helping insurers leverage technology in conjunction with defining business strategy and addressing issues facing the industry. You can connect with Tom on LinkedIn.