Managing product labels, procedures, and compliance documentation can quickly become overwhelming as organizations grow and processes and systems evolve. Over time, legacy content and inconsistent updates can make even routine document maintenance a challenge.
One of Glemser’s large pharmaceutical clients faced this exact issue. Over many years, they had developed global labeling documentation across multiple systems and processes, resulting in a fragmented and inconsistent set of standard operating procedures (SOPs).
To address this, Glemser helped the client optimize how they generate and maintain global labeling SOPs. Collaboratively, their existing documents were ingested into a generative AI tool, producing harmonized first drafts of SOPs, all without disrupting the organization’s day-to-day operations.
Analyzing and Structuring Legacy Content
Using automated data analysis and AI-assisted classification, we first grouped and tagged the client’s content by topic, process step, and evidence source. This analysis revealed opportunities to combine, template, and reference content in ways that achieved the consistency the client wanted. Importantly, the process was fully transparent, documenting every decision point. Once similar documents were grouped, we trained the generative AI tool on this organized content, enabling it to assemble SOP drafts efficiently. At every stage, a human-in-the-loop could review and refine the AI-generated drafts, ensuring accuracy and compliance while accelerating the overall process.
Training Generative AI to Follow Client-Specific SOP Processes
Early in the project, we worked closely with the client to understand their SOP assembly process. Once they provided a process flow diagram, we were able to train the generative AI tool to assemble documents automatically. The model learned the client’s templates, approval language, and required citations. We also embedded additional rules and prompts so outputs consistently met the client’s expectations. To ensure full traceability, the AI tool linked content back to source documents, supporting end-to-end verification of the generative authoring.
Human-in-the-Loop Authoring with Familiar Tools
With the model trained, the next step was to set up a simple authoring interface, allowing users to request fully assembled documents in client-approved Microsoft Word templates for each identified topic. The generative AI tool broke down the templates by section and handled multiple output structures, producing first drafts assembled from reusable content modules.
Client subject matter experts (SMEs) then reviewed and marked up the drafts in Word as they normally would. For sections that required correction, SMEs could “lock down” approved content and adjust prompts for further AI iterations. This human-in-the-loop approach preserved editorial control, shortened review cycles, and improved consistency, all without altering the company’s existing governance or approval workflows.
Results: Faster SOP Harmonization at Scale
Within weeks, we used AI to categorize legacy documents and delivered a working generative AI tool that produced harmonized first drafts, audit logging, version control, and a validation checklist. What would normally have taken months of manual rework was reduced to just days of configuration and review by a small team.
Because the templates and assembly logic are modular, the solution can be quickly reconfigured for different document sets or departments, cutting overall processing time from weeks to days. After rollout, one to two people could maintain a division’s document set, and the same approach can be applied to other departmental collections with minimal reconfiguration. The benefits were clear: dramatic reductions in authoring time, fewer review cycles, improved traceability, and consistent, company-aligned language across all SOPs.
A Practical Model for Safe, Scalable Gen AI in Regulated Content
This project demonstrated how gen AI, when trained on company-specific data and constrained by auditable processes, can harmonize complex legacy documentation quickly and safely. The approach preserves existing governance, transforms authors into high-value reviewers, and scales across an enterprise with minimal friction. For large organizations wrestling with fragmented documentation, this model represents a practical path to faster, cleaner, and more consistent SOPs delivered faster than ever before.
See How Glemser Can Modernize Your SOP Authoring
If your organization is struggling with fragmented SOPs, slow review cycles, or manual harmonization across regions, Glemser can help. Our approach applies generative AI in a controlled, auditable way, accelerating document creation while preserving governance, traceability, and compliance. Contact us today to get started.