How AI Changed the Math on Technology Investment

For decades, buying software was the practical default for most banks. Developing it internally required specialized engineering talent, significant investment and timelines that were difficult to justify when established vendors already offered proven systems. Banks bought technology to meet new needs and when necessary, adjusted their operations to fit what those systems could support.

After years of that approach, many institutions now operate across a patchwork of specialized applications and integrations. Each investment may have solved a legitimate problem, but the cumulative effect can be expensive and difficult to change. A relatively straightforward business decision can trigger vendor requests, customization work or changes across several connected systems.

AI-assisted development is beginning to alter that equation. As the resources required to create and modify software decline, banks have more realistic choices about how they address technology needs. That gives boards an opportunity to reconsider where buying remains the best use of capital and where having more control over the technology could be worth the investment.

The Investment Equation Changes
The economics of a technology purchase extend beyond its initial price. Implementation, integration and customization add costs, while future changes may depend on the vendor’s development priorities and product roadmap. Those dependencies matter most when the system supports a process the bank expects to change frequently or considers important to its strategy.

A commercial system may be an obvious choice for a standardized function that differs little from one institution to another. The calculation changes when technology supports an operating model the bank has deliberately designed, such as how it serves a particular customer segment or manages a specialized business line. In those cases, the ability to change the technology can have value of its own.

Historically, that value often was not enough to overcome the expense of internal development. AI-assisted development is lowering some of that barrier by helping teams create prototypes and applications more quickly and refine them through shorter development cycles. A use case that could not justify a traditional software project several years ago may deserve a fresh financial analysis today.

Consider Today’s Options
The choice is also broader than building an entire system internally or purchasing one from a vendor. Banks can configure existing platforms, develop targeted applications around core systems or use a combination of internal and commercial technology.

Each path carries costs. Commercial software can bring licensing fees, integrations and ongoing customization. Internally created technology requires appropriate governance, security, maintenance and support. Boards should expect management to account for those differences and explain why the recommended approach provides the strongest economics over the expected life of the investment.

This also creates a reason to examine the bank’s broader technology architecture. An institution that can create a targeted application without adding another stand-alone vendor may avoid some of the complexity that has accumulated through years of point-solution purchasing. Over time, that capability can give management more options when a new business requirement emerges.

Challenge the Default
The board’s role remains one of oversight and capital allocation. Directors do not need to determine how software is developed, but they should understand the assumptions behind a major technology recommendation.

Management should be prepared to explain what the bank will spend to implement and maintain the technology, how easily it can be changed and what long-term dependencies the investment creates. For capabilities closely tied to the institution’s strategy, the analysis should also address whether the bank has a practical alternative to purchasing another system.

Buying will continue to make sense for a large share of bank technology. AI-assisted development simply means the alternatives deserve a more serious look than they did in the past.

Banks have spent decades making technology decisions in an environment where creating software was expensive and slow. As those economics change, capital allocation should change with them. The opportunity for boards is to make sure the next technology investment reflects the choices banks have now, rather than the constraints they had when their technology strategies were first established.