Amid the rapid transformation that artificial intelligence is bringing to various sectors, Cynthia Merhej, a specialist in AI for accounting, auditing, and finance, is at the forefront of efforts to develop this field and empower professionals with the tools they need.
In this interview, we discuss the future of the profession, the skills professionals need, the key challenges of integrating AI into organizations, as well as the opportunities available to the next generation in the job market. So, how will accounting and auditing change in the coming years? And what will the role of humans be in this transformation?
– You have launched initiatives to train accounting and auditing professionals on AI tools. What skills should they acquire today to remain competitive in the future job market?
At AIKIT LB, we do not build standalone courses; we build an empowerment ecosystem for accounting and auditing professionals in Lebanon and the region. The first thing we tell participants in every program is: the skill is not “learning the tool.” Tools change every few months, and anyone who builds their future around a single tool will find themselves starting from scratch each time.
The real skill is knowing how to frame the problem. Most of the poor results I see are caused by a bad question, not a bad model.
Then come three layers: understanding the data—where it came from, how it was organized, and where its gaps are; critical thinking to review and reject the output when necessary; and governance. AI can give you a completely confident answer even when it is wrong, and anyone who does not understand the subject matter will fail to spot the error.
Finally, there is governance: protecting client data and documenting what you used and how you used it, because the working papers must remain auditable.
And accounting fundamentals remain an essential requirement with no substitute. You cannot review work you do not understand.
– What is the biggest mistake companies make when introducing AI tools into their financial operations, and what advice would you give them to avoid these mistakes?
Starting with the tool instead of the process. They buy a subscription, distribute it among the team, and then wait for results. Three months later, someone says, “AI did not work for us,” when what actually failed was the lack of proper design.
The second mistake, and the one with greater financial and legal risks, is entering sensitive data into tools without examining their data-retention and data-protection policies. This concerns me more than anything else in the organizations I work with; data-protection policies in many of them are still developing.
My advice is simple and not particularly exciting: choose one recurring, high-volume process. Define what a “correct result” looks like before you begin. Have a human being responsible for reviewing and signing off on the output. Document everything. Then scale.
Companies that treat AI as a change-management project succeed. Those that treat it as a purchasing exercise do not.
– If you could go back to the beginning of your career, would you choose the same path given the current AI revolution? Why?
Yes, and with even greater confidence than I had back then. In fact, I made that decision myself: I moved into IT auditing and analytics in 2014, at a time when artificial intelligence was neither a major market topic nor a headline at conferences. For some, it seemed like I was moving away from the “safe” career path in the profession.
What I learned over those years is that technology alone was not enough. What made it valuable was staying within the profession—understanding the working papers and understanding what would convince an audit committee.
An auditor who understands data is completely different from a data expert trying to understand auditing, and the difference becomes clear at the moment when professional judgment is required.
What would I have changed? One thing: I would have turned that knowledge into a scalable model much earlier. I spent years applying it within closed teams before thinking about building it into a system that could reach the profession as a whole. That is precisely what I am working on today with AIKIT LB.
– What message would you give accounting and finance students who are worried that AI might take away their job opportunities?
Your concern is not naïve, and I will not reassure you with comforting but inaccurate words. The routine tasks traditionally assigned to entry-level professionals are indeed the most vulnerable to automation, which means the ladder that we climbed is changing shape.
But AI does not bear responsibility. No one signs an audit report in the name of a language model. Professional judgment, accountability to an audit committee, and understanding the client’s context remain human responsibilities.
Practically speaking: master the accounting fundamentals first, then add a layer of data and AI on top of them.
The advantage today is not being the fastest at completing a task; it is reaching the stage of professional judgment earlier than those who came before you.
– Looking at the next five years, how do you expect the accounting and auditing profession to change, and who will be best positioned to benefit from this transformation?
I expect three transformations, and I say this as a forecast, not as certainty.
From sampling to the full population: testing all transactions instead of a sample is no longer a technical challenge; it has become a matter of readiness.
From periodic to continuous: controls operating throughout the year rather than during an intensive audit season.
And from preparation to interpretation: professionals’ time will shift from compiling numbers to explaining what they mean.
As for those best positioned to benefit, they are not necessarily the largest firms. They are those who combine genuine professional expertise with fluency in working with data. I have noticed that smaller firms are sometimes faster to adopt these technologies, simply because their decision-making process is shorter.
And the group most at risk? Those who define themselves by the tasks they perform rather than by the judgment they provide.