I am exploring the use of artificial intelligence (AI) to make compliance more robust in the three prongs of prevent, detect and remediate. This series is based upon an article in the Harvard Business Review (HBR), entitled “Artificial Intelligence for the Real World” by Thomas H. Davenport and Rajeev Ronanki, which laid out a structure that every Chief Compliance Officer (CCO) and compliance practitioner can use to think through how AI could be applied in your organization. Today I want to explore overcoming some of the challenges when implementing an AI initiative.

The benefits of implementing AI initiatives in compliance can be as varied as the organizations that implement them. These include enhancing the features, functions and performance of the compliance department and allowing both CCOs and compliance practitioners to make better and more informed decisions going forward. It can assist companies to more robustly operationalize compliance directly into the business operations by using the business inputs to better inform compliance decisions. It can free up compliance professionals from mundane and routine tasks to allow for not only greater creativity but also higher productivity. Finally, a key component can be to not only capture but also apply scarce knowledge throughout an organization.

Conversely there are several factors which can challenge your implementation of an AI initiative. These include both the technology and expertise to initiate and implement a program, as both can be too difficult and too expensive. Obviously, most compliance professionals do not have experience with AI so the learning curve will be steep and outside experts must be relied upon. AI technologies are not that mature and so do not have a long track record. This has led to some AI technologies having been oversold in the marketplace. Finally, as most companies have many, often disparate sets of ERP and other systems, it is often hard to integrate AI technologies with such processes and systems. To overcome these difficulties the authors laid out a framework for thinking through these challenges and some others.

Understanding the Technologies

In Part I, I discussed the types of AI a compliance professional could employ in a compliance program. This is significant as each type of AI initiative has both strengths and limitations. Both process automation and cognitive insight are “transparent in how they do their work, but neither is capable of learning and improving.” Conversely cognitive engagement “is great at learning from large volumes of labeled data, but it’s almost impossible to understand how it creates the models to do it.” This final point can be important where documentation is critical not only for processes but also to demonstrate how and why decisions are made.

Through an understanding of the different types of technologies, it allows you to become better positioned to determine the specific AI system which will enhance your compliance regime. It will also give you insight into how quickly you can implement such a system and even which types of vendors to work with going forward. This will require a CCO or compliance practitioner charged with such innovation to leverage other corporate disciplines such as IT and within IT a data scientist who can assist you in data-skills and also innovation. If such skills and talent are not available to you internally, you will need such subject matter experts from outside your organization. The bottom line is that having the right expertise available to you, whether in-house of external, is critical for your success.

Different Types of AI Initiatives

You need to evaluate your compliance needs with the capabilities of AI systems to help create “a portfolio of projects.” Interestingly, this is essentially a risk assessment but with the focus not on risk but opportunities. The authors found that “typically, there are parts of the company were “knowledge”—insight derived from data analysis or a collection of text—is at a premium but for some reason is not available.” That is a precise description of many aspects of a corporate compliance function.

This can include areas of “bottlenecks” where decisions are slowed down because of a “bottleneck of information”. This is where the information is available in an organization but siloed in departments or disparate corporate functions, clearly the bane of the compliance practitioner. A second area is “scaling challenges” where again the information exists within an organization, but it is too costly or inefficient to scale manually. A final area identified by the authors is “inadequate firepower” where there is simply an overabundance of information.

The next area is to assess the business case for the implementation of the AI solution for compliance. You need to determine not only additional value but what would bring the biggest bang for your buck through the greatest contribution to business success. Consider both short and long-term value, as well as projects that might expand into a broader “suite of cognitive capabilities to create competitive advantages.” Some of the questions you should consider are:

  • How critical to your overall compliance strategy is addressing the targeted issue?
  • How difficult will it be to implement the proposed AI solution?
  • How would the benefits merit the efforts and are there other uses for the AI solution?

Finally, you will need to assess and select the right technology for your organization. This means differentiating tools which can or cannot match live-person problem solving “beyond simple scripted cases. You will need to understand the differences between tools which can streamline simple process but actually slow down more complex problem solving.” The authors end this section by cautioning, “it’s wiser to take incremental steps with the currently available technology while planning for transformational change in the not-too-distant future.”

The process laid out by the authors provides a CCO or compliance professional a framework to think through the challenges they will face in an AI implementation. Yet by using this, or a similar framework, the compliance function can work to garner many of the benefits the authors found for companies which used AI implementation in areas such as customer facing issues, large data issues, including both structured and unstructured data and overly siloed knowledge basis.

This publication contains general information only and is based on the experiences and research of the author. The author is not, by means of this publication, rendering business, legal advice, or other professional advice or services. This publication is not a substitute for such legal advice or services, nor should it be used as a basis for any decision or action that may affect your business. Before making any decision or taking any action that may affect your business, you should consult a qualified legal advisor. The author, his affiliates, and related entities shall not be responsible for any loss sustained by any person or entity that relies on this publication. The Author gives his permission to link, post, distribute, or reference this article for any lawful purpose, provided attribution is made to the author. The author can be reached at tfox@tfoxlaw.com.

© Thomas R. Fox, 2018

0 comments