Fixing the fundamentals in Financial Services operations before getting full leverage of AI

Quite a few years ago, we explored mid-office optimisation with an investment bank — even bringing in Porsche for their Lean excellence perspective.

I have done many other such projects, helping teams identify and automate. Automating the flow from trade execution to settlement is not new. Financial institutions have been digitalising operations for decades.

AI may accelerate some capabilities, particularly around handling unstructured data and helping resolve breaks, exceptions, and reconciliation issues, but it is not a magic wand.

If banks build AI on top of fragmented processes, poor data, duplicated controls, and years of operational workarounds, they are simply layering on top of very shaky foundations. The results are likely to be random, which is not what you want in regulated businesses.

Cut your waste before you have to cut your costs

Before removing the people who understand where the problems actually are, there is critical work to do: simplify processes, reduce variation and variability, rationalise exceptions, improve data quality, and eliminate operational waste.

It is easy and dangerous to place AI on top of such broken fundamentals, because it will make mistakes, and those mistakes can be costly in Financial Services.

For the last couple of decades of global outsourcing, Banks have chosen to ride the waste, only with cheaper people to handle it, rather than eliminate it. The culture of Kaizen & Continuous Improvement is often drastically lacking.

Technical foundations still rely heavily on people's best judgment and informal communication networks. If you doubt that, look at how much still runs on spreadsheets!

If you think that you have a User-Created-Application headache with spreadsheets, you are heading for severe corporate migraine by encouraging the adoption of AI

The institutions that will succeed with AI will not be the ones cutting fastest, but the ones combining AI with continuous improvement, operational discipline, and teams empowered to improve the system itself.

That is Jidoka (translated to autonomation or automation with a human touch) applied to Services and Tech businesses.

AI is not the magic wand of fixing decades of legacy, but it might be the opportunity to finally address it

AI will not fix broken banking operations overnight, and the banks that start by offloading the people who understand the operational realities may learn that at their peril. Financial Services has pursued digitalisation for decades, yet many institutions remain far from clean data and true Straight-Through Processing (STP). Rather than hoping AI will perform miracles on weak foundations, organisations should challenge the fundamentals: simplify processes, improve data quality, eliminate waste, and combine AI with Continuous Improvement and operational discipline.

The uncomfortable reality is that operational excellence has rarely been the favourite child of Financial Services organisations. For years, many institutions chose to ride the waste through outsourcing to cheaper locations rather than eliminate it at source — which, in Lean terms, is fundamentally the wrong optimisation. You should not optimise waste; you should remove it.

Perhaps AI finally creates the forcing function to revisit those neglected foundations: data quality, process discipline, technology architecture, Continuous Improvement capability, and the leadership behaviours required to sustain them. Building robust AI-enabled operations is far more complex than simply buying tokens from OpenAI or Anthropic. It requires building operational systems capable of learning, adapting, and improving continuously.

Get in touch if this challenge resonates and you are looking to develop the foundations of Continuous Improvement and waste elimination as part of your AI transformation.

---

Note: This blog post was in reaction to another Investment Bank announcing mass lay-offs due to AI. Please see article at: https://www.finextra.com/newsarticle/47776/standard-chartered-to-cut-7800-jobs-as-ai-takes-over-the-back-office