Modernizing Turnover Package Reviews in Pharmaceutical Projects: A Step Towards Digital Project Delivery

As pharmaceutical projects continue to embrace digital transformation, the TOP review process presents a significant opportunity for digitalization. This blog post explores a potential future state for TOP review using artificial intelligence (AI) and outlines how organizations can transform a labor-intensive, time-consuming review process into a streamlined, data-driven, and highly efficient operation.

Introduction

In pharmaceutical capital projects, the successful turnover of systems and equipment is essential for ensuring project completeness, thereafter for ongoing operations and regulatory compliance. The TOP serves as a consolidated collection of documents that demonstrate that a system has been designed, installed, tested, and documented according to project requirements and applicable standards.

A non-exhaustive list of documents that are typically contained within the TOP binder.

• Piping and instrumentation diagrams
• Bill of materials
• Datasheets of component, equipment, instrument, etc.
• Manufacturing documents (E.g., weld isometrics, weld log)
• Material of construction certificates
• Mill test reports
• Certificate of compliance for elastomers
• Surface finish reports
• Construction test and inspection records (e.g., pressure test, cleaning and passivation reports)
• Loop calibration certificates
• As-built drawings

Before turnover to the end user, these documents must be reviewed and approved to verify their completeness, accuracy, and compliance. Despite the advancements in digital technology and availability of various turnover software, review of TOP remains largely manual, creating inefficiencies and increasing project risk. The aim is to have a fully digital and automated TOP review workflow leveraging AI capabilities with human oversight for decision making and TOP approval.

Current State

At present, TOP reviews are performed using either of the following approaches,

• Paper-Based Review: TOP exists physically as hard copies and a reviewer manually examines hard-copy documents for compliance.
• Paper-in-Glass Review: Documents are managed digitally within the document management system (DMS) or turnover software. Though the document is electronically available, reviews are performed manually.

In both the cases, the review is conducted manually, by comparing against the specifications and requirements. Any non-conformities or comments identified are provided using review forms or within the document or via spreadsheets or as per agreed project practices and then shared with the equipment supplier/service provider for corrective action. Upon addressing the non-conformities, the documents are re-verified, and correct versions are placed in the turnover package binder. While numerous DMS and turnover software platforms are available on the market, at present they are used primarily for managing and collating the documents. Turnover package review remains a manual-person driven activity. The review is iterative in nature as the review cycle is repeated each time the document is revised.

Though these current practices seem like a viable approach, they present several challenges:

• The review approach, level of technicality, and depth of review vary from person to person, leading to inconsistency and variation in outcomes.
• Reviews can be a time-consuming activity, as the turnover package contains numerous pages. Reviewing every page/document consumes more hours and resources.
• Reviews may have limited traceability due to manual review process; also, tracking the document review status and reason for acceptance or rejection of comments can be challenging when comments are scattered across emails, spreadsheets, and review forms.
• The quality of a review depends on a person’s experience and expertise.
• Reviews may be prone to human error; reviewers may overlook discrepancies due to timeline pressure, document complexity, or fatigue.
• There may be limited scalability for large size facilities; as facility size increases, the TOP quantity also increases making manual reviews inefficient and difficult; delays may also occur in system turnover and handover.

Advances in technology provide an opportunity to modernize this process. Rather than manually processing bulk volume of documents for review, automated workflow can be deployed to perform the review. The intention is not just to carry out the review in digital platform, but to transform the way in which turnover package review is approached.

Future State

Moving towards an automated review process requires a clear and deep understanding of system lifecycle requirements, technical intricacies, document dependencies, review methodologies and workflows.

One potential solution is the adoption of intelligent workflow automation, where a trained AI agent reviews the TOP against the specification document(s) and generates a compliance report highlighting whether the TOP meets the review criteria or not, with human in the loop for final decision making. Depending on the organizational need, the workflow can be developed as a standalone application or can be integrated into existing digital solutions.

The process begins with the ingestion of technical documentation(s) in a document processing module, where the turnover documents are uploaded. Similarly, requirements and specification documents such as user requirement specifications, design specifications, project standards, turnover, and good documentation practice criteria are ingested in the requirements management module for extracting the requirements. The extracted requirements become the benchmark against which the turnover documents are evaluated. The module(s) are designed to process multiple file formats such as PDF, Word, Excel, AutoCAD drawing files, and handwritten records. Optical character recognition and document intelligence technologies shall be used in converting these structured and unstructured documents into machine readable format. These extracted and organized data are used for subsequent compliance evaluation.

Reviews are performed within the compliance assessment module, where an AI agent extracts text from all the uploaded files. It then analyses and structures the checklist by organizing each requirement with unique IDs, categories, and metadata. Finally, the compliance assessment modules evaluate extracted documentation against each checklist item, determine the compliance status for every requirement, and identify the supporting evidence used to justify an assessment.

The TOP content varies across equipment, systems, and document types. Each vendor document has its own structure and format. For example, the contents, document layout, measuring range, and technical specifications in calibration certificates differ depending on the instrument. Similarly, some certificates are manually signed, while others use digital signatures. Given this variability, an AI model must be trained on diverse, high-quality data that captures these different document formats and characteristics. Therefore, the functioning and reliability of the compliance assessment depends heavily on the quality of the training data and the robustness of the model training process.

Upon completion of the review process, an AI agent will generate a structured, traceable and detailed review summary highlighting the compliance status with recommended actions and routed for human verification and approval. A system-generated review summary is reviewed by an expert for verification and approval. Based on the review, an outcome turnover decision and final TOP review report are issued. If the reviewer rejects the review summary, the workflow captures the feedback, records with the audit trail and notifies the package owner for corrective action and resubmission.

Each module performs dedicated functions while an AI agent orchestrates these modules to execute the overall TOP review process. Depending on the project requirements, organizations can redefine the module names, functions, and workflows that align with their project and quality management systems.

This transformation shifts the turnover review process from a person-dependent activity to an automated methodical review, which has the following benefits:

• Standardized and methodical approach improving the consistency across the system/projects
• Reduces the review time from weeks/days to less than a few hours
• Ability to analyze large volume of documentation without fatigue, thereby reducing the rate of human errors
• Better traceability with an audit trail
• Shorter review cycles translate to faster phase handover and project delivery
• Improved resource allocation and utilization
• Data-driven insights and metrics for turnover quality, recurring issues, and performance.

Every system turnover package has its own characteristics, and the acceptance criteria may differ from one piece of equipment to another. For this reason, workflows should be developed as a configurable framework that can accommodate different turnover packages and system specific acceptance criteria, without requiring changes to the underlying process layer.

Considerations for Implementation

Though the TOP review may look simple to automate, it is not. Turnover requirements, documents, and contents vary drastically from piece of equipment to piece of equipment, from project to project, and based on client requirements. Therefore, an automation workflow should be carefully planned and developed considering these variations with the flexibility to adapt to emerging needs. A few points to consider are:

• Data quality is paramount. Performance of an AI agent is directly proportionate to the quality of the data provided. Therefore, complete and accurate source documentation is essential.
• Standardization of documents, templates, naming conventions, and meta data structures can significantly improve the accuracy and reliability of automated reviews.
• Define clear and unambiguous user requirements, turnover review criteria, and acceptance criteria early in the project as these forms the baseline against which the turnover package will be reviewed.
• Vendor engagement is crucial. Engage early in the project and brief the document expectations and digitalization objectives. The effectiveness of a solution depends on the vendor support, documentation standards, and data quality.
• Train and validate AI models using historical project turnover data to learn about the structure, content, and patterns of documentation including material certificates, calibration records, inspection reports, and execution documents.
• Establish data governance policies, access management, and auditability to ensure data integrity and compliance objectives are met.

Conclusion

TOP reviews are among the most resource-intensive activities in pharmaceutical projects. Advances in digitalization can provide an opportunity to improve these traditional manual processes. By leveraging technology to automate repetitive document verification and compliance tasks, project teams can improve project delivery, while enabling engineers to focus on higher value assignments. Human oversight will continue to play a crucial role as the final decision making regarding system acceptance and turnover will be taken by human. With the right combination of high-quality data, standardized practices, effective workflow design, stakeholder engagement and appropriate human oversight, the proposed solution could enhance the turnover readiness and project delivery across pharmaceutical projects.