Engineering Amnesia

Last week I came across an article published by ISPE (International Society for Pharma Engineering) about knowledge management which I felt was a must share.

ISPE identifies Knowledge Management (KM) and Quality Risk Management (QRM) as two of the fundamental enablers of a modern Pharmaceutical Quality System. In other words, knowledge management is no longer viewed as an administrative function or an IT initiative. It is considered a strategic capability that directly contributes to product quality, compliance and continuous improvement throughout the product lifecycle.

While ISPE's discussion is centred on pharmaceutical manufacturers, I couldn't help but think about the companies that design and build the sophisticated equipment that makes modern pharmaceutical manufacturing possible.

That observation led me to a simple question.

If knowledge is so fundamental to pharmaceutical quality, what makes engineering knowledge so difficult to preserve, connect and reuse across the lifecycle of pharmaceutical equipment?

Pharmaceutical equipment manufacturers generate an extraordinary amount of engineering knowledge every single day. Every customer requirement influences a design decision. Every engineering review results in calculations, drawings and specifications. Every Factory Acceptance Test (FAT) uncovers insights that improve future projects. Every Installation Qualification (IQ), Operational Qualification (OQ) and commissioning activity adds another layer of experience. Every customer installation, service visit, deviation, CAPA and troubleshooting exercise contributes new knowledge to the organization.

Over the years, these activities produce an extensive repository of engineering drawings, validation protocols, calibration records, material certificates, service reports, firmware release notes, application notes and customer-specific configurations. Collectively, they represent one of the organization's most valuable assets.

Yet this raises an important question.

How easily can engineers access this accumulated knowledge when they need it?

Consider a typical service scenario. A customer reports an unexpected issue on an HPLC system that was installed three years earlier.

The service engineer doesn't necessarily need new knowledge. What they need are answers to questions such as:

• Has this issue occurred before?

• Which firmware version was installed on that system?

• Which corrective action resolved the problem previously?

• Were any related CAPAs or engineering change requests raised?

In most organizations, the answers already exist.

However, they may be scattered across document management systems, PLM platforms, ERP systems, shared network drives, email conversations, validation reports or, quite often, in the experience of a senior engineer who solved the problem years ago.

The challenge, therefore, is rarely a lack of knowledge. It is the ability to discover, connect and reuse the knowledge that already exists within the organization.

This challenge is becoming even more significant as pharmaceutical manufacturing embraces Pharma 4.0. In their 2025 paper, Information Model for Pharmaceutical Smart Factory Equipment Design, Wölfle and colleagues describe how modern pharmaceutical equipment has evolved into highly interconnected systems where mechanical, electrical, automation, software and compliance disciplines are tightly linked. They argue that traditional document-centric approaches are no longer sufficient to manage these complex relationships and advocate richer information models that capture dependencies and connections across the engineering lifecycle.

Taken together, the message from ISPE and the engineering research is remarkably consistent.

The future of pharmaceutical equipment manufacturing is not simply about creating more documentation. It is about ensuring that engineering knowledge is connected, accessible and reusable throughout the organization.

This is where an Engineering Knowledge Assistant can play an important role.

Engineering knowledge is often distributed across CAD repositories, PLM or product data management systems, ERP applications, quality management systems, document repositories and service management platforms. These systems excel at storing engineering information, maintaining version control and ensuring regulatory compliance. However, they were designed primarily to manage information, not to make it easily discoverable across multiple repositories.

As engineering organizations grow, knowledge becomes increasingly fragmented. Drawings may reside in a PLM system, validation documents in a document repository, service reports in a CRM application, change requests in a quality system and technical discussions in emails or collaboration platforms. Finding the right answer often requires knowing where to look before the search even begins.

An Engineering Knowledge Assistant addresses this challenge by providing a unified, natural language interface to an organization's engineering knowledge. Instead of searching through folders, document numbers or multiple enterprise applications, engineers can ask questions in the same way they would ask a colleague.

For example, a service engineer might ask:

• Has this failure mode been encountered on this equipment before?

• Which firmware version was installed on this customer's system?

• Were any engineering change requests raised after this machine was commissioned?

• Which validation protocol applies to this equipment configuration?

The assistant retrieves information from approved engineering documents, validation records, service reports, quality records and other authorized sources, while maintaining references back to the original documents. This enables engineers to verify the source of every answer rather than relying solely on AI-generated responses.

The value extends beyond faster information retrieval. By making existing knowledge easier to access and reuse, organizations can reduce troubleshooting time, improve consistency across engineering teams, shorten onboarding for new engineers and preserve valuable expertise that might otherwise remain siloed within individual projects or experienced employees.

The objective is not to replace existing enterprise systems. Those systems continue to serve as the systems of record for engineering and quality information. The role of an Engineering Knowledge Assistant is to make the collective knowledge stored within those systems readily accessible, allowing engineers to spend less time searching for information and more time applying it.

In many ways, this represents the next stage in the evolution of Knowledge Management. For years, organizations have focused on capturing and storing knowledge. Generative AI now provides a practical way to make that knowledge available through conversation, enabling engineers to interact with decades of accumulated organizational experience as naturally as they would consult a trusted colleague.

References:

1. ISPE Good Practice Guide: Knowledge Management in the Pharmaceutical Industry (2021)

   ISPE Good Practice Guide: Knowledge Management in the Pharmaceutical Industry

2. Wölfle, R., Saur-Amaral, I., & Teixeira, L. (2025). Information Model for Pharmaceutical Smart Factory Equipment Design. Information, 16(5), 412.

   Information Model for Pharmaceutical Smart Factory Equipment Design (2025)