Automating Compliance Operations Using Motor, Sensor, & Decider Bots
By Rahmat Muhammad, Deloitte
Life sciences businesses need to move quickly and innovate. Compliance is often seen as a brake that slows businesses down, and in some cases it may become a barrier to innovation.
In life sciences, compliance organizations are under pressure to meet several objectives, including:
- Shift from after-the-fact checks for compliance to real-time detection and resolution of noncompliant events
- Provide the business with predictive risk insights
- Increase visibility into non-U.S. activities and quality events
- Empower markets outside the U.S. to take ownership over compliance
At the same time, compliance organizations face specific challenges when managing compliance risks (see sidebar). Underlying most, if not all of these challenges, are restrictions of time, money, and available talent. In light of these restrictions and in order to address the challenges that hinder their objectives, many compliance organizations are turning to analytics and process automation tools.
In life sciences, compliance organizations have traditionally focused on ensuring quality and compliance after the fact. The business produces a product or engages in an activity, and the compliance organization checks after the fact, identifying problems after failures have already occurred — and then wasting precious time and resources on remediation. This approach is costly and no longer sustainable.
Introducing The Robots
Compliance “bots” (software robotic process automation “robots”) are being embedded into business processes so that issues are detected in real time and mitigated before they can continue downstream. A bot is a software application that performs a specific task — for example, a compliance bot can be programmed or trained to perform specific compliance tasks. Bots can be built with “motor,” “sensor,” and/or “decider” configurations, depending on task complexity.
Motor Bots
Relatively simple and repetitive tasks that occur in a stable environment can be automated using the motor bot configuration. Motor bots, which rely on rule-based algorithms, are used for executing predetermined actions (e.g., export an output to a folder). This configuration does not decide what action to execute nor does it recognize when the action is no longer appropriate.
Sensor Bots
When a task tends to vary in unpredictable ways or when it occurs in an unstable environment, a sensor bot configuration can learn the appropriate responses to these variations. The sensor bot, using machine-learning algorithms, will recognize when a previous action is no longer appropriate, but it does not decide which action, among a set of options, is appropriate.
Decider Bots
Complex tasks that involve decisions and balancing risks can apply the decider bot configuration. For example, a typical action for a motor bot might be to perform specific operation “X” regardless of conditions. In the same situation, the sensor bot could detect a change and identify a set of alternative actions “Y” and “Z” to take as a result of the specific conditions. However, the decider bot, using augmented intelligence algorithms, can weigh the risks associated with options X, Y, and Z and trigger the most appropriate action based on the internal decision logic.
These three bot configurations can be combined and applied in many ways to help execute various types of compliance operations.
Case Study: Computer System Validation (CSV)
In life sciences, the CSV process is an unavoidable part of doing business that is mandated by regulations and encouraged by industry-leading practices. In the typical CSV process, validation documentation is typically manually reviewed by a quality assurance (QA) team.
A global pharmaceutical company struggled to implement a standard CSV process across business lines. QA reviews were often subjective and driven by the knowledge and experience of the individual reviewer. As a result, technology teams did not implement consistent documentation practices and efficiencies were lost across the board.
The company implemented a new compliance operating model with a single systems development life cycle (SDLC) and common validation templates. It empowered a central CSV group to execute QA reviews against standardized criteria. To further encourage consistency, QA reviews were supported by automation bots that perform initial reviews of the validation documents using the standardized criteria:
- Sensor bots “read” through the validation documents and learn to identify content that is relevant to the review criteria
- Decider bots decide if the content sufficiently meets the review criteria
- Decider bots then trigger motor bots to generate an output that clearly and consistently identifies which parts of the document require additional work and the reasons for the required changes
The output generated by the motor bots is produced in a structured format that is aggregated over time across all validation documents. As a result, the company is able to analyze the information for insights and trends related to process quality.
Case Study: Product Labeling
Product labeling in life sciences is a highly regulated and complex process. On one hand, a company must meet country- and region-specific regulatory requirements, while on the other hand it must ensure consistency in the information across various labels for the same product. In addition to the regulatory complexity, many organizations’ processes for approving product labeling involve multiple stakeholder groups, including medical/clinical, legal, regulatory, supply chain, manufacturing, and quality assurance.
A global pharmaceutical company’s product labeling approval process was time-consuming and often delayed a new product’s time to launch.
The company implemented an operating model that empowered a global label approval group to review and approve product labeling against a standard set of review criteria. It implemented a cloud-based content management system to streamline the workflow process. However, to drive consistency in messaging, improve compliance with regulatory requirements, and cut down on actual review time, the company deployed automation bots as follows:
- Label approval sensor bots that “read” through labeling materials to extract content
- Decider bots then decide if the content meets the company’s review criteria
- Decider bots trigger a motor bot to generate an output that clearly and consistently identifies which parts of the product label require additional work and the reasons for the required changes
Similar to the CSV case, the output generated by the motor bot can be produced in a structured format that is aggregated over time across all product labels. This enables the company to analyze the information for insights and trends related to process quality and labeling consistency and accuracy.
How To Get Started
A first step to implement real-time compliance operations using robotic automation is to identify and prioritize compliance processes to be targeted for automation (e.g., continuous monitoring, audits, regulatory reporting, copy approval, pharmacovigilance, process rationalization). When prioritizing, focus on processes that, if automated, would add significant value to the business. At this stage, process complexity should not be a factor.
High-priority processes are then broken down into steps and their associated standard criteria. Each criteria is further broken down into logic, and each logic block is then aligned with an appropriate bot configuration. This level of detail is crucial to developing a realistic plan and timeline to build and deploy the bots.
The Value Of Compliance Automation
Standardization and automation effectively allow compliance organizations to help the business achieve “right first time” without wasting time and resources.
When compliance operations are standardized and automated, the compliance organization can shift unused employee creativity to tackle more challenging and strategic risks posed by emerging innovations such as mobile cloud computing and additive manufacturing, among others.
Standardized processes and robotic automation can also be applied to markets outside the U.S. – markets that have even fewer resources to devote to compliance – where they can improve process efficiency and generate a wealth of data that can be mined with advanced analytics for greater visibility and insight into these markets.
About The Author
Rahmat Muhammad is a consultant in Deloitte Advisory’s Life Science practice. She has 14 years of experience in the design and delivery of operating model, process, and analytic solutions to address complex scientific and business operations problems in the life sciences. At Deloitte, she uses a hypothesis-driven approach along with qualitative and quantitative data to assess and generate insights into areas of business risks and operational inefficiencies. Rahmat is currently focused on using analytic tools to develop automation technologies for near real-time risk detection and compliance remediation.
About Deloitte
Deloitte refers to one or more of Deloitte Touche Tohmatsu Limited, a U.K. private company limited by guarantee (“DTTL”), its network of member firms, and their related entities. DTTL and each of its member firms are legally separate and independent entities. DTTL (also referred to as “Deloitte Global”) does not provide services to clients. In the United States, Deloitte refers to one or more of the U.S. member firms of DTTL, their related entities that operate using the “Deloitte” name in the United States and their respective affiliates. Certain services may not be available to attest clients under the rules and regulations of public accounting. Please see www.deloitte.com/about to learn more about our global network of member firms.
This publication contains general information only and Deloitte is not, by means of this publication, rendering accounting, business, financial, investment, legal, tax, or other professional advice or services. This publication is not a substitute for such professional 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 professional advisor. Deloitte shall not be responsible for any loss sustained by any person who relies on this publication.