It is difficult to imagine a world now without AI. The advances in AI and the impact it is having on the economy are quite real. AI is here to stay it appears, and there is a reasonable fear that many jobs could be replaced by AI.

What about jobs related to financial operations, analytics, and audit? How will these professions and the work they do be impacted by AI?

Summary

I believe that AI will be a strong complement and will be able to automate many relatively easy tasks. However, there will still be many complex tasks that AI may not be able to solve for a long time, or possibly ever at all.

So what do you do in the meantime? Ideally you want a solution that solves today’s problems, while laying groundwork for the potential adoption of AI in the future.

That brings us to my predictions of the future related to Data Preparation applications and strategy

Prediction

• Data Preparation applications will continue to be a useful and relevant to financial professionals, and analysts, and they will have AI embedded/integrated to accelerate business processes.

Key Reasons why AI for Complex Financial Operations will take time.

Financial professionals still need to learn the skills of working with data to accomplish those tasks that AI can’t yet, to then teach AI how to accomplish them, and to validate that what AI performed is right. Additionally many financial operations that require audit reviews or the creation of financial statements require higher scrutiny and human oversight. I think we are a long way from AI submitting a call report.

What are some of the reasons why AI adoption and success for complex business processes may take a long time, or may never come about?

AI is adding to the complexity of Enterprise Technical Debt as multiple AI platforms are adopted to improve specific processes

Many organizations already have complicated technology stacks, and with the introduction of AI, we are going to see a disparity of AI software adding to the complexity of technical debt Enterprise.

For example, a financial institution has a core banking system, a financial general ledger, loan servicing system, a customer relationship module, multiple databases including an enterprise data warehouse. Challenges already exists in integrating data from all these systems, and many of these systems may now have their own AI, which means that you then have the challenge of integrating AI and constantly cross training and information sharing one AI to another.

We are seeing this now at a personal level. We have copilot AI from Microsoft, ChatGPT, Claude, Gemini, which one do I use? Use them all you say? Use one to address specific challenges?…this growing list of AI tools and what they are uniquely good at is adding to the stress and complexity of technical debt.

AI is bound by the same rules of data quality — especially the GIGO problem; garbage in and garbage out

All business process, especially financial operations must be based on good clean quality data. You need to perform a journal entry? You need to reconcile your system of record against the financial general ledger, you need to compile and create your financial statements, you need to file a quarterly call report with regulators, or you need to file your business taxes. There is no room for error with such financial operations, the negative impact and repercussions are substantial.

“We had AI reconcile the accounts, generate the financial statements and file it with the regulators”. –  says “He who does not want to be named” 😀

How do you trust that AI got access to all the data it needed to perform the entire process? If AI is provided with data that is not accurate and or not provided with the right context, and the right series of steps or prompts to perform the task appropriately you may have a process with bad data.

Complexity and exceptions processing of many use cases

Some business processes are more complicated than others, not only in the amount of work required to prepare the data, but also the amount of exclusions, exceptions, and the number of data sources and systems that are required to perform the task. These complex processes may take a long time to teach AI to do the process.

Audit/Governance and Controls

Last and certainly not least is the topic of Audit, Governance and Controls. How do you put controls and audits around an AI financial operations process? How do you prove that the data wasn’t manipulated? That it was acquired from a trusted source?

Do the right resources have access to the AI tool or can anyone access it?

Do we trust the AI tool and are we comfortable with it having access to our systems?

What are the worst case scenarios if AI has access to our systems and something terrible happens?

If AI got something wrong in an Audit, who is held accountable?

Conclusion

AI is not the enemy of financial professionals — it’s a powerful tool that, when paired with clean data and human expertise, can unlock tremendous value. Data Preparation applications are well-positioned to continue bridging the gap by optimizing operations in the present while laying the foundation for future AI adoption, that will be safe, reliable, and more effective for financial operations.

The organizations that will thrive are those that invest now in data quality, governance, and process documentation — not to resist AI, but to be ready for it. Our expertise in Data Preparation Strategy for finance and analytics teams can help accomplish this goal.