AI in EQMS: Three Things Executive Heads of Quality Need to Know

My colleague Niels Andersen recently wrote a Research Spotlight and blog post on Artificial Intelligence (AI) in manufacturing. However, there has been some development in the AI space since Niels' research was published, namely that OpenAI made significant announcements; most important is the ability to build your own GPT (they call them GPTs) without coding. Of course, Open AI has had some recent public drama as well. Generative AI has been adopted into Enterprise Quality Management Software (EQMS) at a very rapid rate. Virtually every EQMS vendor has rolled out some capability for Generative AI on their platforms.

Generative AI is the category of AI associated with language processing known as Large Language Models (LLM). These tools are adept at consuming spoken or written language and digitizing the result. Where Generative AI is different is that it turns out results of queries across a large amount of language data sources and produces concise results. For the detailed inner workings of Generative AI, refer to Niels' work mentioned above.

AI Saves Significant Time and Reduces Errors

Generative AI typically has several potential use cases across Quality work processes. The theme for AI use cases in EQMS is efficiency. Traditional Quality work processes sometimes require searching, gathering, and evaluating information from disparate sources or from many weeks, months, or years to assemble a complete record for review and decisions. These types of processes can be laborious, time-consuming, and prone to error.

This is where Generative AI can save Quality team members significant time. For those in regulated industries, such as Life Sciences, gathering information for a regulatory submittal or event investigation is daunting.

While not every EQMS vendor is approaching Generative AI exactly the same way, they all are addressing time-consuming information-gathering processes with Generative AI in one way or another.

You Have Many Choices in AI Solutions

Let's highlight a few EQMS vendors and their approach to utilizing Generative AI in their applications.

1. 1. 1. Veeva Systems

         Veeva Systems had its annual Quality and R&D conference in Boston this past September, where it announced the rollout of several new applications and some new features, including an AI Application Bot for key processes within the Veeva platform.

      2. ComplianceQuest

         ComplianceQuest had its inaugural User Conference in Tampa in April of this year. At this event, CEO Prasanth Rajendran also announced a Large Language Model enhancement for the ComplianceQuest platform to facilitate what they call the "informed user" persona. The concept of the Informed User enabled by AI is a user that can quickly gather relevant information from across the ComplianceQuest platform to speed up users' abilities to provide summaries and reports through AI-assisted discovery. Since then, ComplianceQuest has spent a lot of effort infusing several different areas of its platform with AI-supported functionality, including:

         1. 1. 1. 1. Assisted Decision Making

                  2. Intelligent Automation

                  3. Guided Workflow

                  4. Intelligent Recommendations

                  5. Predictive Visibility

      3. Honeywell Connected Life Sciences (Formerly Sparta Systems)

         Honeywell Connected Life Sciences was an early adopter of an LLM integration. In June of 2022, at its conference in Orlando, they announced Qualitywise.ai as the first use case around risk management using AI. The AI use case in QualityWise.ai categorizes complaints and narrows possible root causes. A company with large amounts of data generated over a long period would benefit from this application of AI. Over time, growing the use cases for AI will add more value to users of the Trackwise and Trackwise Digital processes. An example of this is "automated summarization' which pulls information from various sources to generate quick, accurate, and complete summaries. Users can tweak these, allowing them to spend time on value-adding activities rather than repetitive and time-consuming tasks.

      4. Intellect
         
         

         Intellect held its user conference, Innovate 2023, in November, where they rolled out their approach to using AI in three areas:

         1. 1. 1. 1. Regulatory Support

                  2. Document Integration

                  3. Audit Assistant

         The first half of 2024 will see the next iteration of Intellect AI™, which uses LLMs to derive insights from company data, enabling quality programs to adopt a more proactive and prescriptive approach.

      5. MasterControl

         MasterControl has a Natural Language Processing component called "Gen.IQ." MasterControl recently launched the first beta program for Gen.IQ capabilities to enable the auto-generation of comprehension exams for a training module as a unique, time-saving angle for their customers. MasterControl will continue to build out the Gen.IQ portfolio with additional releases in 2024 for the Annual Product Quality Report (APQR), auto-generation of SOPs, and auto-translation.

         In addition, MasterControl has AI and ML features as part of their integrated analytics package – Premium Insights. This includes ML-supported trending analysis, process controls, and more, as well as AI-supported natural language search (e.g., ask, "What is my biggest Quality problem today?").

You Should Build on What You Have

With EQMS implementations over 30% across industry, there is still plenty of room for homegrown solutions. Our research on Unified Performance Excellence (UPX) shows a strong value in building on what you have. This value doesn't stop at Performance Excellence; many are architecting homegrown solutions for various applications on existing platforms. Microsoft, Google, and Amazon are the big players here. Microsoft's efforts on Generative AI have been in the news recently. These platforms' efforts to build Generative AI into their ecosystems make this new tool available to those taking the "build on what you have" approach.

From this sampling across some EQMS vendors, we see that most recognize the potential for AI to improve information-gathering tasks associated with typical quality processes. The potential for growth in the applicability of Generative AI in support of Quality roles is significant. Several potential use cases that we can see are:

1. 1. 1. 1. Alerting on condition changes impacting the quality of the product

         2. Batch release record gathering

         3. Knowledge capture and transfer associated with new product development or discovery

         4. Next best action suggestion for CAPA and/or Customer Complaints

Recommendations for Executive Heads of Quality

• • • Leverage Generative AI to reduce administrative burdens. Most EQMS vendors are working furiously to integrate Generative AI into their applications. The potential for this is to reduce administrative burdens associated with gathering information and classifying for various purposes but be cautious about assuming that there are cost savings to be had immediately. AI still needs to be supervised for accuracy by knowledgeable resources.

    • Utilize the newfound freedom from administrative burdens to upskill employee resources for more creative and impactful work. AI is not a cost savings play; it is an efficiency play, and the best way to take advantage of freedom from administrative burdens is to elevate expectations and enable broader, more creative tasks that AI cannot perform.

    • Work with IT teams and partner functions if you don’t have AI enablement across your enterprise software. They can help you develop compelling AI-enabled personas that IT can clearly communicate as requirements to vendors or search for standalone capabilities. Microsoft is very involved in generative AI and offers an AI copilot solution that would be a simple add-on if the company's email and business software is from the Microsoft suite, as most are.

Every new technology experiences a hype cycle. It seems the hype cycle for Generative AI is shorter than typical. Vendors and manufacturers are finding ways to explore the potential of this new tool in the toolbox.