Enterprise Risk Strategy and Emerging Technologies to address critical needs

Posted By on Aug 14, 2019 | 0 comments


On July 31, 2019 I had the opportunity to listen into a PricewaterhouseCoopers (PwC) webinar hosted by SDRAN discussing how to build a risk strategy utilizing emerging technologies such as Artificial Intelligence, Blockchain and IOT to improve quality, strengthen compliance and controls such as adverse event controls, reporting and reduce non-conformances.

As technology has advanced, the amount of data being produced by corporations is enormous and at best only a small fraction (PwC estimates .5%) is currently analyzed leaving a huge opportunity for anyone willing to invest in these emerging technologies. A more data driven focused approach can assist in accuracy, completeness and timeliness of reporting and compliance.

Underlying the point of adoption of emerging technologies in large firms, PwC conducted a survey of 7,300 respondents in 123 territories with the majority of respondents being senior executives. The results showed that continuous monitoring of network and email for security is showing the greatest adoption at 40% of respondents utilizing the tech. Anomaly detection, proactive detection of threats and data dashboards all have around 30% adoption and are a critical component of an analytics focused company. The leading edge is using data scientists and AI (with 17% and 11% adoption respectively) to use big data to solve ranges of issues to combat fraud, advance automation projects and streamline operations and workflows.

PwC’s 2018 Global Economic Crime and Fraud Survey was completed by 7,228 respondents from 123 territories. Of the total number of respondents, 52% were senior executives of their respective organizations, 42% represented publicly-listed companies and 55% represented organizations with more than 1,000 employees.

The three main areas of focus that are ready for enterprise adoption now are: Artificial Intelligence (defined as a collection of “smart” technologies and algorithms that are aware of and can learn from their environment to assist /augment human decision making), Blockchain (immutable, publicly distributed ledger) and IOT or Internet of Things (“Devices utilize embedded technology to communicate, record, and interact with the external environment using the internet as a means of communication”), with a synchronicity or convergence of these technologies allowing for benefits to rise while negating some of the downsides as well as the potential for true disruption.

Side note: In the mid-term (3-5 yrs.) 3D printing and robotics are coming up for wide scale enterprise use, followed by nanotechnology and quantum computing in the next 5-10+ yrs.

By demonstrating some case studies and best practices, it made these topics more relevant and realistic for those looking to adopt these strategies. Some examples of how adoption of these three emerging technologies is and could play out around the areas of Risk and Regulatory Affairs:

Registration and License Tracking can benefit from securing data sharing to guarantee patient data privacy through securing the credentialing process using an immutable anonymous ledger or blockchain.

Regulatory Intelligence and Complaint processing could benefit from Unstructured data mining using NLP (natural language processing) and using risk algorithms to determine fraud.

Regulatory Submissions and Clinical Trial Data Analysis could benefit from machine learning algorithms, using NLP as well as well using automated scripts in the cloud or on servers to reduce manual inputs and better manipulate, pull, label and organize data across an IT system.

Case Study 1 [Blockchain Application]: Blockchain can transform and vastly simplify the gathering and capturing of transaction detail across systems, allowing for a complete record to be stored on the blockchain and automating appropriate access and audits of the data to appropriate entities (i.e. Regulatory entities and payers)

Post Marker Regulatory Change Management could become more efficient and benefit in reduced manual processes using AI algorithms and cloud synchronization across IOT devices.

Adverse Event Reporting has many different sources (Mobile texts, Email, information in excels, dashboard or internal tools, etc.) and can all be merged using cloud hosted data source to ingest the data into a separate database for processing. There the data is read through NLP nodes to identify and label key entities. Finally, the data is interpreted with machine learning and rules-based models perform interpretation and then sent back to the appropriate workflow across the IT system. All the while, the entire process is being tracked and monitored by humans through a detailed user interface.

Case Study 2 [AI Application]: Example of funneling data sources to a centralized Machine learning algorithm to enable more automated adverse event reporting

Centralizing and streamlining Health Data to create virtual consultations and analysis. One further application which currently is being led by IOT devices such as the Apple Watch, is tracking, monitoring and packaging health data progression over time to be sent to physicians. A case study provided was where PwC partnered with iBData to create a single process for capturing and transmitting IBD progress in a way that is easy for the patient and informative for the clinician. This resulted in Clinicians being able to compile a comprehensive patient profile and develop a targeted treatment plan with standardized updates on a patient’s symptoms.

Case Study 3 [IOT Application]: Using Apple Watch to track IBD and then turn the data into useful dashboard for both clinicians and patients

The speakers did caveat this with the fact that biases are inherent in AI designed systems by the programmers and inputs, so best to be as objective with data as possible and conduct data audits and set clear baselines and review all model outputs. Other suggestions include setting clear validation tests for the model with real and created inputs. Furthermore, to really make these technologies work cohesively they need to constantly be a work culture of testing and looking at data, defines clear objectives of models and improve the model to get closer to answers that meet the objective as well as ensure it is reverse engineerable and explainable/auditable.

Background art source: https://www.lightfarmbrasil.com/

Submit a Comment

Your email address will not be published. Required fields are marked *