Where to start: Frequently asked questions about generative AI

Generative AI remains a hot topic in the business world, as organizations across all industries are attempting to utilize the technology for greater efficiency, productivity and business insight. But as companies begin to understand the vast potential of the technology and as solutions rapidly evolve, they can struggle with determining where it could make the most impact and properly aligning strategies to overall objectives.

Leaders from RSM’s Acceleration Center for Innovation Research Lab have compiled answers to frequently asked questions about generative AI to help you understand the technology and how to best utilize solutions within your business.

How does generative AI work?

Several companies have created generative AI models trained on large datasets known as large language models. Users prompt the application with tasks or questions processed by the neural network or model. The application then provides the most probable results based on the request, and the user can provide positive or negative feedback, which ultimately trains the model to be even more effective.

I don’t know if I need a chatbot. Do I need to care about generative AI?

Embracing generative AI has become an important element in enhancing productivity for many businesses. Regardless of the size of your business, your competitors are likely evaluating or already using generative AI solutions, and you may be missing opportunities for growth by not keeping pace. In addition, generative AI solutions are much more expansive than just chatbots (see below).  

I have heard about ChatGPT and Google Bard. Are those the only solutions?

No. Generative AI tools are evolving every day, and new solutions are rapidly emerging on the market. While chatbots from major technology companies are among the most popular generative AI applications, many other solutions can perform key tasks, from writing copy to generating music, images and videos. With the rapid expansion of available solutions, it’s important to understand the market and choose the best tool for your needs.

What are some common use cases for using generative AI tools?

Generative AI can increase efficiency and insight in many ways, including summarizing large amounts of text in articles and reports and brainstorming new ideas, concepts and outlines. Available tools can also provide coding for scripts and simple programs, reformat text into tables and bullet points, and analyze the sentiment of text-based datasets such as reviews and survey results.

What is an effective prompt?

Focusing on how you structure your prompts is important to getting better outputs from generative AI tools. Important considerations when building a prompt include:

• Context: Provide a role, description, persona, etc.
• Task: Detail what you want the tool to do.
• Data input: Provide the necessary data (note: do not enter sensitive data into publicly available large language models).
• Format: Specify how you’d like the output formatted.
• Tone: Suggest a tone, such as professional, conversational or enthusiastic

For example, if you are a middle market clothing retailer specializing in sustainable materials and practices with an earthy, minimalistic modern brand, some effective prompts may include:

• “We’re hiring a full-time store manager. Write a job description and include details about my company.”
• “I want to reduce my operating costs by 15% this quarter. Give me a list of cost-cutting strategies to consider.”
• Come up with 10 potential slogans for the company that match our aesthetic and would resonate with eco-conscious customers.  

My organization is not using any AI applications today. What governance do I need in place to manage it?

Like other innovations, generative AI requires an effective governance plan to guide usage. Generative AI requires oversight because it evolves rapidly and uses probabilistic analysis to determine what to “create.” Several AI governance frameworks are currently available, including platforms from the National Institute of Standards and Technology (NIST), Google and Microsoft, as well as guidelines from several countries and organizations. In addition, RSM’s AI Governance Framework combines elements from several leading standards as well as best practices from other countries and organizations.

Many of these technologies are new. Is the data that goes into generative AI applications truly secure?

While generative AI offers the power to create content rapidly, it also poses risks like cybersecurity vulnerabilities and privacy concerns. To ensure secure usage, adopting an AI governance framework that ensures transparency, works to prevent bias and protects sensitive data is crucial, supported by active involvement from leadership. This approach can help align AI adoption with strategic goals, mitigate risk and enable the technology to deliver rapid results for your business. When using public generative AI tools, you must consider that the data provided to the open-source model is being used to train the large language model continuously. However, companies can take steps to create their own more secure versions of these tools that allow users across the business to use more sensitive data when interacting with the model.

How do I know if my IT infrastructure is ready for generative AI?

Assessing the readiness of your information technology (IT) infrastructure for AI adoption is crucial. You can start by evaluating your current infrastructure’s scalability, performance and security. Consider factors like data storage capacity, the quality of your data sources and network capabilities. Additionally, ensure that your IT team has the necessary skills to support AI integration and that you have a clear governance plan in place. Using AI tools or any other type of machine learning model can be extremely beneficial. However, the results are only as good as the data that they are trained on. Your company’s data architecture must be evaluated when leveraging AI tools with company data. Additionally, training your people to understand the technology can be beneficial to ensure tools are being used appropriately and securely.

What is my first step in taking advantage of generative AI?

The number of generative AI solutions available now is very extensive, and the number of potential uses is expanding daily. Leveraging a strategy for your business is important. But you must consider if any solutions are specifically designed for your industry, whether you must follow any regulatory guidelines and what models can best align with your business to strengthen your operations. Working with an experienced advisor can be essential to helping you understand the market and integrating generative AI solutions to stimulate greater efficiency and growth.