Key Discussion Points
- Generative AI (GenAI) can simulate various realities, but it shouldn’t be the go-to tool for every business task focused on reality.
- Combining AI with specialized tools and human skills enhances AI’s accuracy and minimizes errors.
- Survey findings indicate that while AI usage is climbing, the pace of transformation is uneven, and agentic AI shows potential for increased value if supported by effective governance.
As a partner in Audit & Assurance (A&A) and a leader in digital products, my focus is on transforming auditing and accounting through the integration of artificial intelligence (AI) to address real-world challenges. Yet at home, I’m an enthusiastic hobbyist—sometimes a bit too optimistic. My experiments have led to flooded kitchen floors, boat oil caps sinking in lakes, and the odd minor electrical issue. I believe in trying to fix things myself, and while I often learn from my mistakes, I inevitably relearn the same lesson: always use the right tool for the job.
This principle is particularly relevant in today’s landscape of Generative AI (GenAI). Many AI solutions are marketed as universal interfaces, suggesting a single tool can handle every task. This is problematic because a generative model acts primarily as a simulation engine. It doesn’t directly reflect objective reality; rather, it imitates reality by reconstructing learned data patterns. While this simulation simplifies complex systems and can be quite beneficial, its detachment from factual reality can also lead to inaccuracies.
AI and Specialized Tools: A Comparison
This brings us to the essential topic of this article: just because a powerful AI model is available doesn’t mean it’s the best tool for every task. Not all activities require a digital simulation. In many scenarios, more straightforward technologies—like calculators for arithmetic, structured databases for data management, or deterministic workflows for verification—can be more dependable.
For professionals navigating complex technological landscapes, the real challenge lies in discerning which tools are best suited for different tasks, rather than automatically resorting to a simulation interface for everything. Reducing inaccuracies involves strategically employing generative models where they can add the most value, while relying on specialized systems for their intended purposes. This approach fosters precision, consistency, and reliable outcomes.
Strategies for Companies to Better Leverage AI
In line with this, a recent survey by Deloitte, released in January as part of our report, State of AI in the Enterprise, sheds light on the successes and challenges experienced by companies in their AI endeavors. The survey gathered insights from over 3,200 business and IT leaders globally, all of whom are directly involved in their organizations’ AI initiatives.
What does the survey reveal about the swiftly evolving AI landscape? It indicates that many organizations are still at the beginning stages of extensive AI-driven transformation. Over one-third (37%) report using AI merely on a superficial level, with minimal alterations to current operations.1 This suggests that they are merely adding AI onto existing processes rather than reimagining workflows to harness the combined strengths of humans, AI, and other technologies. In essence, they are applying AI in a one-size-fits-all manner instead of selecting the appropriate tool for each specific task.
In conclusion, understanding the strengths and limitations of both AI and specialized tools is crucial for organizations aiming to transform their operations. By strategically selecting the right tools for the right tasks, companies can maximize the benefits of AI, leading to more effective outcomes and enhanced value.
