The art of the practical: Adopting AI in a responsible way

Focus on what’s practical to begin your AI journey and become data-driven

January 12, 2024
#
Machine learning
Data analytics Artificial intelligence Digital transformation Predictive analytics

The amount of information that is currently circulated about artificial intelligence (AI) can feel overwhelming. AI tools and applications have truly transformative potential for many key business processes, and much of the conversation has focused on the art of the possible. But in many cases, focusing on the art of the practical is likely a better approach for companies just beginning their AI journey and determining where to start to become data-driven.

RSM US LLP Advanced Analytics Practice Leader George Casey and National AI Risk Leader Dave Mahoney recently hosted a webcast titled Art of the practical: Expanding what’s possible to responsibly start adopting AI today. In it, they provided insights and best practices on how to navigate the intricate AI environment and drive enhanced growth and efficiency.

Understanding the potential of advanced analytics applications

Before deploying an AI strategy, companies must have an effective data analytics foundation in place. A data analytics maturity model can provide a road map for collecting raw data to leverage advanced analytics capabilities, such as AI. The steppingstones in the model, which can show what you can get from your data as maturity increases, are:

  • Descriptive: What happened?
  • Diagnostic: Why did it happen?
  • Predictive: What will happen?
  • Prescriptive: What should we do next?

The prescriptive category represents the most mature level of analytics, and this is where AI resides.

Generative AI

Generative AI takes AI to another level, with the potential to affect many tasks workers are currently performing, often by significantly increasing the speed of their completion with the same level of quality. And the influence of generative AI spans all levels of the organization, with higher-income jobs facing the biggest impact.

Generative AI can be used to enhance several business functions by providing summaries of any text or document and generating text, images and code in multiple programming languages. In addition, businesses commonly use generative AI tools for chatbots that engage in conversations based on learned knowledge and plug-ins that can unlock powerful capabilities within applications.

The following advanced analytics tools have practical applications that companies can leverage today:

  • ChatGPT is an AI chatbot powered by an advanced large language model that can effectively process and generate natural language text. It is specifically fine-tuned to generate human-like responses in conversational contexts.
  • AutoML automates the process of developing and deploying machine learning models without requiring prior knowledge of programming or data science. It uses algorithms to carry out effective feature engineering, algorithm selection, hyperparameter tuning and model deployment, resulting in accurate and scalable predictions.
  • Microsoft AzureML supports a wide variety of automated machine learning tasks, increasing productivity with predictive models that provide easy data exploration and intelligent feature engineering.
  • Prevedere delivers cloud-based automated and advanced forecasting solutions with no new hardware or software to install. It has extensive global data sets that enable AI-powered predictive analytics and machine learning capabilities.

These and other emerging AI and advanced analytics solutions can deliver optimized business solutions in many ways, including:

  • Creating hyper personalized campaign outreach programs
  • Understanding the drivers behind, and timing of, customer churn
  • Generating responses to questions in proposal requests with built-in governance controls
  • Detecting potential outliers or anomalies in financials

Considering the inherent risks of AI and advanced analytics

However, while the increased access to AI tools and applications has ushered in a new wave of options for productivity, insight and efficiency, the new technology comes with risks. Companies need to educate employees on what AI and advanced analytics tools are, how they can be used, and what the risks associated with them are. Because the technology will continue to grow in popularity, there will be data security and privacy concerns that companies need to consider.

If you are looking at how you want to transform your business, consume your data and enable your business and your workers to do different things, start to define what that is going to look like,” says Mahoney. “Where are you going to put some technical innovation and investment? And then you can strategize and manage risk around that strategy and what makes sense for you.

Understanding multidimensional AI-related risks—regulatory, privacy, legal, ethical, operational and financial—is vital for smooth implementation, responsible use, trusted results and compliance. These risks are most prevalent with the end user, but also extend to the product, provider and model. It’s important to understand that pre-trained generative models that are available to us, unless specified otherwise, are not purpose-built. Therefore, independent validation and verification of information from AI applications is necessary.

Ultimately, successfully implementing AI requires a strategic combination of automated safeguards, human oversight and layered defense to ensure reliable, trustworthy operation. The risk profile and risk ramifications will change, depending on the scope of AI strategies within the business but responsible deployment should be accurate, transparent and explainable, reliable, robust, fair, resilient, secure and private, with extensive user control.

To help guide your advanced analytics and AI strategy, you must integrate a governance model, ideally incorporating ethical and social considerations, transparency, accountability, data governance, human control, and legal oversight. Several structures exist, but RSM has developed our own comprehensive responsible governance framework that takes best practices from internal frameworks and other regulatory guidelines to govern AI from inception to operations.

In terms of practical applications, people are not being replaced by AI, but they may be replaced by people that know how to use AI,” says Casey. “AI can empower workers and make them more productive and change the definition of a full-time equivalent and what they can accomplish.

The takeaway

AI and advanced analytics are rapidly changing the world by disrupting critical processes throughout organizations. While your company can experience seismic change from AI and advanced analytics applications, starting with a practical approach that considers potential risks and how solutions align with current goals and business objectives can result in significant rewards.

To learn more, view the recordings or register for upcoming sessions in the data analytics and AI webcast series. In addition, contact our AI solutions team to understand how your business can successfully leverage emerging AI tools and applications.   

RSM contributors

Data analytics and AI webcast series

Register for our webcast series that explores the challenges and opportunities presented to middle market leaders as they progress on their data analytics journey. 

Data analytics learning portal

RSM has developed a comprehensive collection of webcasts from our experienced data analytics advisors to help you capitalize on your data. Topics include: 

  • Data management and governance
  • Data architure and data federation
  • Demos of several leading data analytics tools

Related solutions