Artificial intelligence (AI) is a transformative force that is reshaping how many people work. During a recent fireside chat at the Philadelphia Alliance for Capital & Technologies (PACT), Chris Denver, Director of Accounting and Reporting Advisory at Stout, and Heath Durrans, Director at nimbl Consulting LLC, discussed five aspects of preparing for corporate AI implementation.

1. AI is not novel.

The integration of AI into core platforms like Microsoft Office underscores the inevitability of its adoption. But AI is not a novel concept; AI concepts date back to the 1950s, shortly after the invention of the computer. Certainly, many are aware of the 1997 defeat of chess superstar Gary Kasparov by IBM’s Deep Blue, a made-for-purpose chess AI. Humanity has encountered disruptive technologies throughout history, from the invention of the wheel to the digital revolution. AI represents only the most recent disruptive evolution.

This should be comforting for leadership. There’s no need for panic related to this technology; rather, a rational and objective perspective is essential. While AI may bring challenges, it also promises significant benefits.

2. Humans are needed for proper AI use. 

Humans will play a crucial role in the AI-driven world as a check-valve on data quality and reasonableness. Quality control is imperative, as initial AI-generated content often requires human editing and refinement, or the prompt may return invalid results (often called AI “hallucinations”).

As such, organizations need to consider talent acquisition strategies. What roles will be needed to use AI properly? Should a company build AI capabilities internally, or is it more beneficial to seek external expertise through consulting firms? The decision will depend on the company’s culture and goals, as well as the speed at which this technology develops.

3. Proper AI implementation requires an understanding of its intricacies.

Many companies that want to implement AI may not fully comprehend its complexity – and limitations. AI is not a magic solution but a tool that does what it is instructed to do, relying on human expertise for its effectiveness.

Roles like AI managers and prompt engineers are becoming increasingly crucial, as these people possess the skills needed to safely harness the power of AI tools. Defining a vision of a successful AI implementation will aid company leadership in making the proper investments in talent, software, and governance when planning an AI strategy. This vision will likely include AI integrated into daily operations, working in the background to enhance productivity without requiring active user engagement.

4. AI presents unique cybersecurity concerns.

While AI offers immense potential, it can also pose risks if not handled with care. Concerns about AI’s ability to compromise cybersecurity have been raised, with claims that it can crack Advanced Encryption Standard 256-bit encryption quickly. However, cyber defenders are also using AI to advance their efforts and thwart hackers.

In the realm of fintech, where security is paramount, companies must be proactive in communicating their security posture. Transparency is key, and customers need assurance that their financial data is protected. Fintech companies must articulate their AI adoption strategy and convey the value, benefits, and risks of AI in their operations.

Preparing for potential risks and having mitigation strategies in place is essential, as no system is entirely immune to hacking attempts. Companies must work to minimize those risks and make it as difficult as possible for malicious actors to exploit vulnerabilities in all systems, including AI.

5. AI success should be defined and measured.

Defining and measuring success in AI initiatives requires a nuanced approach. In fields like robotic process automation (RPA), success isn’t necessarily achieving 100% automation. Instead, it involves finding the right balance between automation and cost-effectiveness.

For example, a bot that efficiently processes 93% of transactions but routes the remaining 7% to humans for handling can be a more cost-effective solution than striving for full automation, which may come at a significantly higher cost.

AI initiatives might be considered in the same way: getting the greatest return for the lowest risk and expense. AI will not solve every problem, but it can significantly enhance operational efficiency and accuracy, allowing human resources to be deployed to a higher and greater use.

AI in the Future

In a rapidly evolving landscape, human oversight, realistic expectations, and robust cybersecurity measures are pivotal to harnessing the potential of AI while mitigating its risks. Third parties that are skilled in IT governance and controls can deploy AI solutions experts who are at the forefront of this technology to ensure that company assets are protected and opportunities are maximized.