Enterprises are turning to AI to solve their efficiency problems, but is just a short-term fix?

As companies grow and merge over time, they accumulate technical debt. Technologies, capabilities, and processes that in a perfect world would be retooled and improved every time the company grows or merges with another company, are instead cobbled together and strung together through short-term fixes due to a combination of resource constraints and simple inertia. Technical debt can accumulate for years or decades – for example, a surprising number of government and payment systems still run on 1970s COBOL code, despite a decreasing number of programmers who are fluent in it – and is often only resolved when a critical system fails and short-term fixes are no longer enough to fix it.

But while it’s easy to blame organizations for accumulating technical debt instead of simply keeping their systems continuously updated, the reality is that some technical debt is inevitable in almost every organization. Particularly in situations where companies merge, integrating one company’s systems with those of another means resolving countless conflicts between technology, processes and workflows. And with a limited number of people and time to address each issue, businesses must triage to ensure that the most important issues receive the highest priority – with the inevitable result being that issues that can be more easily resolved with a short-term fix than with a long-term solution, or which do not pose a threat to any critical systems, are sidelined until they become a problem. In other words, as in the world of finance, not all debt is bad, and there are sometimes prudent uses of a little (technical) debt leverage!

However, when a system or process becomes so outdated that it significantly impacts productivity, it may be worth investing in ways to update it. The question then becomes: should we demolish the whole thing and rebuild it as it would have been if it had been built from scratch, or apply another quick fix to make things better (which inevitably leads to the problem resurfacing in a few years, repeating the scenario)? The answer invariably depends on the cost in time and resources of both options, as well as how close the problem is to the critical systems that keep the business running smoothly.

It seems that in the post-AI world, the calculus has changed somewhat as to which problems require a complete rebuild and which can be solved with a short-term stopgap. For example, LPL Financial recently announced a new AI-powered tool this would allegedly make it easier for advisors to navigate their complex payment schedules and understand “how to earn more.” Which, in a vacuum, seems like a good example of investing in AI, since AI can be great for bringing together complex information into a simple, easily understandable format, and helping advisors increase their pay can at a minimum make advisors happier, more likely to stick with LPL, and perhaps can also generate additional revenue for LPL. But at the same time, the AI ​​tool does not solve the underlying problem: LPL’s payment system is so complex that it requires an AI tool just to understand it. Simplifying the payment structure so that advisors can clearly understand their own compensation would be more beneficial in the long run for advisors and their clients, but for now, AI can serve as a workaround that gives advisors a little more clarity without requiring a complete revamp of employment contracts and compensation schedules.

In other cases, companies use AI as an overlay to improve existing processes to the point where they are at least closer to what they would be if rebuilt with modern technology without AI. For example, Cambridge Investment Research announced its own AI-powered tool this would reduce the time needed to open an account from nine days until “only” 17 minutes. Which again represents a significant improvement in efficiency given the time savings per account on thousands of accounts opened per year in Cambridge… but in a world where custodians from Betterment to Altruist to incumbents like Charles Schwab enable almost instantaneous paperless account opening. without using AI, Cambridge’s announcement is more of a comment on the technical debt of their systems. rather than as a revolutionary new use of technology. Again, although the optimal path, in hindsight, would have been to make technological improvements to integrate and simplify the systems, rather than stacking so many existing systems on top of each other that AI is needed to tie them all together, the existence of AI allows Cambridge to defer investment in more fundamental updates to their systems while implementing rapid improvement that actually improves their short-term productivity and support for their advisors.

The broader question, in cases where large companies rely on AI tools to “solve” problems created by technical debt, is whether AI itself is the long-term solution, or whether it represents just another short-term stopgap that will inevitably need to be resolved in a few years. It still takes a lot of resources to rebuild existing technology, or to rip it out and replace it from scratch, and with AI a potentially cheaper solution to an increasingly complex set of problems, it will be tempting for companies to use AI as a band-aid to provide short-term relief. But in reality, this is exactly how technical debt is created and accumulated: by prioritizing timely, short-term solutions over making substantial investments to solve the core problem. Which suggests that in the future, AI will be used even more as a way to repair existing systems rather than rebuild them (including systems that are already fixed with AI!), as a way to extend technical debt rather than as a way to pay it off.

So, amid the pervasive hype around AI and claims that AI will replace large swathes of workers across the economy, it is worth recognizing that, at least among some of the consulting industry’s largest employer firms, one of the most concrete uses of AI has been simply to move systems and processes closer to the point where they already would have been if companies could have allocated the time and resources to keep them up to date throughout the process. This can certainly represent material improvements for these companies, given that even incremental improvements in efficiency can have a substantial impact on a mega-company scale. But we’re a long way from the promise of AI as the productivity revolution of the future when, at present, it’s more useful as a short-term loan to pay off long-term technical debt.

Ben Henry-Moreland is a Senior Financial Planning Specialist atKitces.comwhere he specializes in writing and presenting on financial planning topics including tax, practice management and technology. He is also co-author of the monthly Kitces #AdvisorTech column. With his experience as a financial planner and solo owner of an advisory firm, Ben is passionate about the site’s mission to make financial advisors better and more successful.

Michael Kitces is the Chief Financial Planning Specialist atKitces.comdedicated to advancing financial planning knowledge and helping to make financial advisors better and more successful. Additionally, he is responsible for planning strategy at Focus Partners Wealth, co-founder ofXY Planning Network,AdvicePay,Recruitment of new planners,fpPathfinderAndFA Bean Countersthe former editor-in-chief of the Journal of Financial Planning, the host ofFinancial Advisor Successpodcastand the editor ofthe popular financial planning industry blogNerd’s View.

This article is part of our Monthly Spotlight series, focusing in January on AI in Wealth. Full coverage can be found here.