Author: Liz Peuster, CAE
This article originally appeared in the Winter 2025 issue of the South Carolina CPA Report
As many of you know, I’m passionate about helping CPAs embrace the power of AI to transform their work. CPAs can leverage AI to streamline operations, enhance productivity, and drive success.
Navigating the AI Landscape: Key Considerations
Integrating AI effectively requires careful consideration of several key factors:
- Data Privacy & Security: It’s crucial to understand how AI models use data and to adopt an acceptable use policy to protect your business.
- Responsible Use: Remember to always infuse a human touch, verify AI outputs, and remain aware of potential biases.
- Understanding Features & Limitations: Clearly define your goals for AI integration and be mindful of each model’s limitations. Commercially available models, like ChatGPT, Claude, and Gemini, can perform a multitude of tasks in record time. But, they are not all built the same – they all have features that surpass their competitors but also drawbacks. For instance, Gemini generates better images than ChatGPT’s Dall-E,* but ChatGPT can export data and content into Word and Excel files.
- Building Trust: Never blindly accept AI outputs; human review is essential. Nothing should leave your building or your inbox without a gut check. AI produces fantastic results, but it’s rarely perfect.
- Time Allocation: Dedicate time to explore AI tools and stay adaptable in this rapidly evolving
landscape. To start, block 20 minutes twice or three times a week. Use this time to tinker, listen to a podcast, or scour LinkedIn. Once you build this habit, you’ll identify quick wins and start stacking time.
*This is my opinion. Art is subjective.
The Art of Prompt Engineering
Have you ever had a conversation with a peer or an intern where you assigned a task or project? Think about the context you provided; the background, the detail, the specifications. My guess is the details that person had directly impacted the quality of the work or end product. Prompting an AI is no different.
Effective prompt engineering is essential for maximizing AI’s potential. Remember, prompts are more than just commands; they’re the foundation of a conversation with the AI. Remember—they are called Large Language Models. These models thrive on words. The more context they have, the better.
Here are the four key components:
- Role: Assign the AI a specific role to focus its expertise.
- Context: Provide ample background information. If you don’t feel like typing it out, most of the models have a voice component. You can simply “talk” to your AI. The latency is low, and the conversation will feel more human.
- Task at Hand: Clearly state the tasks you want the AI to complete.
- Questions: Think of that wide-eyed intern sitting across from you. They have questions, but they won’t ask outright. AI can be the same way.
Unless prompted, the AI will perform its assigned actions. But, as you are a human, you may have believed you provided full context but forgot a key component. To avoid rework, ask your AI if you provided enough context, what else it needs to know, or (my favorite) what you forgot to tell it. Facilitate conversation, and you will maximize your results.
Practical Applications & Quick Wins
- AI offers a wealth of practical applications for CPAs, including: 26 South Carolina CPA Report | Winter 2025 www.scacpa.org
- Effortless information scanning and organization (think business cards, notes you scribbled during a meeting, the layers of sticky notes on your desk).
- Brainstorming and idea generation for projects and reports.
- Content regeneration across various mediums. Are you tired of writing engagement letters? Let AI help you with that lift.
- IT support and troubleshooting. Show AI screenshots of what you’re seeing and ask for solutions. Describe your issues.
- Transcription and summarization of meeting notes. Were you so preoccupied with taking notes during a meeting that you forgot to engage? It happens! But you can leverage AI to help you! For instance, Microsoft Teams and Zoom will use an AI to record meetings, provide transcripts, and recommend follow-ups and action items.
Measuring AI Effectiveness
To gauge the effectiveness of your AI integration, track these key metrics:
- Efficiency & Resources: Monitor time and cost savings.
- Accuracy & Reliability: Benchmark AI outputs against human performance and implement feedback loops for continuous improvement.
While these metrics are allegorical, you will understand very quickly what is and what isn’t working. These metrics will help motivate you to experiment with AI and encourage your team to lean in. I would wager that you’ll discover efficiencies and time savings that ultimately compound and save hours.
The Future is AI-Powered
The AI landscape is constantly evolving, and even seemingly simple applications can lead to significant improvements in your workflow and productivity.
By understanding AI’s capabilities and limitations, employing effective prompt engineering, and prioritizing data privacy and responsible use, CPAs can unlock new avenues for growth and efficiency. I believe that AI has the potential to revolutionize the accounting profession, and I’m excited to be a part of this journey with you.
