Is Artificial Intelligence the Future of Data Analytics? (article)

Is Artificial Intelligence the Future of Data Analytics? (article)

Home /  Insights / Articles / Article Details

Illustration of Artifical Intelligence functions

Artificial intelligence (AI) often evokes thoughts of cyborg assassins or humanoid robots. While Hollywood has extravagantly portrayed the concept, AI in the real world is much more basic. It entails electronic processes that mimic human cognitive functions, like learning or problem solving, and we use it for everything from ride-sharing apps to Google. As AI becomes more prevalent, organizations are looking for ways to capitalize on it. Recent technological advances, like IBM’s Watson or Google’s Virtual Assistant, have placed AI at the forefront of business strategy. But just because the technology is new and carries the potential to create efficiencies in the workplace, doesn’t mean it’s flawless. AI comes with risks, from system configuration to information security concerns. Organizations need to understand the prerequisites required to be successful in AI to protect themselves from those risks and ensure the AI is working effectively to meet their business objectives.

How Does AI Work?

AI works by combining large amounts of data with fast, iterative processing and intelligent algorithms, allowing the software to recognize patterns or trends in data and learn automatically from them. It runs on neural networks or networks set up to think like a person. Imagine teaching a child to throw a ball, learn a new language, or complete a puzzle. Children experiment with their surroundings, and their brains analyze this information to adapt to changes. Creating AI isn’t really much different. You provide it with data, and it learns from patterns and trends to make predictions on how to behave in a variety of situations.

What Do I Want AI to Do?

Once you have an understanding of what AI is, you need to determine exactly what you want it to do for your organization. Often executives have a goal in mind, such as improving customer service, but they haven’t identified the specific function for AI to perform to achieve that goal. For example, consider how you would teach the child to throw the ball. You wouldn’t simply tell them to “play”. That instruction is too broad, which opens you to a wider variety of potential outcomes. Rather, you’d provide the child with the specific instruction to throw the ball as along with examples of what a ball is, how an arm throws an object, and how you would throw a ball yourself. AI should be treated the same way. Clearly defining what you want it to do will help you better understand how to build the best program.

Do I Have the Right Data?

Knowing what you want AI to do isn’t enough to fully build a program. You need to understand both the data that you already have and the data that you’ll need for AI to execute the specific functions you’d like it to perform. Think back to the child and ball. Demonstrating the act of throwing with a baseball could help the child learn how to throw. However, if you fail to explain to the child what a ball is, he or she may not be able to apply his or her newly learned throwing capabilities when he or she encounters another type of ball, like a football with a different shape or a tennis ball with a different texture and color. Teaching with incomplete or bad information could lead to incorrect decision making. This concept applies to AI as well.

Before you decide to upgrade a system to AI-based technology, spend some time creating an inventory of the system’s data. Understand how it is managed, and document what you possess, where and how it is stored, and who has access to it. Create internal controls to help you facilitate complete, accurate, and efficient data use. Do not forget controls for protecting any sensitive or personally identifiable information (PII) that will be stored or used by the AI program. As regulatory environments continue to evolve, most notably with the recently passed General Data Protection Regulation (GDPR), security and privacy needs to remain top of mind for all organizations.

How Will I Maintain Data Integrity?

Before launching an AI program, it’s important to consider how you’ll ensure your data’s maintenance, accuracy, and consistency throughout the program’s lifecycle. Imagine if someone taught the child that heavy rocks and balls were interchangeable without your knowledge. Now, the next time you attempt to play catch, you find yourself on the receiving end of fast pitched skipping stone. Data integrity is incredibly important for those continuously learning. Altered or deleted data can have major ramifications on business decisions driven by your AI program. Review your data queries to ensure the results are retrieving the desired information. Make sure that none of your data has been altered throughout any transfers. Consider the input and processing controls associated with your systems to ensure data is complete and accurate.

Should I Build or Should I Buy?

Once you have a grasp of your data and have mapped your program objectives, you’ll need to determine whether to build out yourself or use a vendor. Many organizations jump right into procuring a tool with the anticipation that licensing a software solution will be enough. The reality is that AI needs to learn through information, trial and error, and feedback. Without your data and review, the AI can’t produce your desired results. That means that regardless of whether you build or buy, someone from your organization will still need to be involved. Identify what resources you currently have in-house that can be devoted to your AI program. Consider the buy-in you have from stakeholders and the potential value AI will provide to you down the road. This analysis will help you decide how robust of an investment to make now and where there is potential to scale up resources in the future.

How Will I Keep My Information Secure Once My Program is Implemented?

Suddenly you’re collecting and processing high volumes of data, and automation is allowing you to do so with little to no direct human involvement. Let’s face it – implementing AI makes your organization a prime target for hackers. You’re placing a high reliance on the output of your AI program, so it is critical to protect the confidentiality and integrity of your data. Start with standard security and privacy measures. These may include multi-factor authentication or up-to-date security patching. Implement an Identity Access Management program to protect the privacy of data by ensuring that only those with a business need have access to it. In instances where your AI program interacts directly with your customers, like a chatbot, consider encrypting the conversations to make it more difficult for hackers to access the information if it were to be released into the wrong hands. Vulnerability management programs should be in place to detect any potential weaknesses surrounding data storage and transmission, and a security operations center could be utilized to monitor information security. Periodic penetration tests can test your environment to help you continuously improve data security controls.

Consider Your Situation before You Begin

There are clear examples of how organizations can capitalize on AI, but that doesn’t mean what works well for one will work well for all. Those that excel at analytics take the time to understand their goals and limitations and tailor their programs to their unique criteria. They also understand that further automating data processing opens them up to security risks, so they put the proper cybersecurity measures in place to ensure their data integrity and confidentiality remains intact.

Success with AI is obtainable, but not overnight. It will take proper planning, resources, and evolution over time. For more information about how AI can help with your data analytics program, contact us.

Cyber risk assessment screenshot


Copyright © 2018, CBIZ, Inc. All rights reserved. Contents of this publication may not be reproduced without the express written consent of CBIZ. This publication is distributed with the understanding that CBIZ is not rendering legal, accounting or other professional advice. The reader is advised to contact a tax professional prior to taking any action based upon this information. CBIZ assumes no liability whatsoever in connection with the use of this information and assumes no obligation to inform the reader of any changes in tax laws or other factors that could affect the information contained herein.

CBIZ MHM is the brand name for CBIZ MHM, LLC, a national professional services company providing tax, financial advisory and consulting services to individuals, tax-exempt organizations and a wide range of publicly-traded and privately-held companies. CBIZ MHM, LLC is a fully owned subsidiary of CBIZ, Inc. (NYSE: CBZ).

Is Artificial Intelligence the Future of Data Analytics? (article)~/Portals/0/PackFlashItemImages/WebReady/AIThumb.jpghttps://www.cbiz.com/Portals/0/liquidImages/WebReady/AIThumb.jpgArtifical intelligence is rapidly being incorporated into business functions, but in order to use it effectively, you'll need controls in place to protect your data....2018-10-29T12:17:07-05:00

Artifical intelligence is rapidly being incorporated into business functions, but in order to use it effectively, you'll need controls in place to protect your data.