Insights: Know your process

If you are responsible for process improvement, if you are new to the manufacturing team, if you experienced but want to confirm intuitions on objective grounds, or if you are developing a new process, this module is for you. The Insights module preforms very useful visualization as well as semi-automated statistics on every measures from your process.

Model : Build digital twins

This module uses advanced Machine Learning routines, but very well grounded into manufacturing business. It will discover what are the key process drivers (initially flagged as CPP or not) of any outcome of interest (such as yield, production lead time, or any possible CQA, ...). And it will do so taking automatically into account regulatory constraints. The produced models are thoroughly evaluated, based on data sciences metrics, as well as business-related metrics.

Optimize : Put models at work, get recommendations for improvements.

This module will put models at work. It makes recommendations about how to reconfigure production lines to improve with respect to any specified endpoint. It will also measure the originality and potential gain of each recommendation. Finally, it will suggest the next round of experiments you could run to improve your processes.

Continuous Process Verification

As a production manager, you want to ensure that operations goes smoothly, to speed up batch release, to avoid unplanned maintenance, to stop production of a batch that is doomed to fail. You also need to produce batch release records as fast as possible. All these questions can be addressed by the Continuous Process Verification of Hercule. It gathers all available data in almost real-time and produces useful insights and call-to-actions for every single batch. It also allows you to track data management progress. If a part of the data is not yet available to the system, you will know, and may take corrective actions with the team involved. Insights generated by this module can be exported into editable reports, in order to support batch release, with up to weeks gained in the process!

Datasets: future-proof !

When it comes to AI and modelling, data comes first. AI Act, Good Machine Learning Practice (GMLP), GAMP5 V2... they all require traceability, transparency and thorough evaluation of your datasets. This module allows for complete management of datasets. Select what batch should be included or not, what measures should be involved or not, exclude outliers, some specific values, etc. It also enables validation and storage of datasets as well as datasets audit trails.

Datasets can also be conveniently exported in various formats, so that they can be used within your (second) favorite data science environment (R or Python homemade scripts, JMP, SAS,...).

Data Warehouse: Relevant data model.

QC data are in your LIMS, Raw Material data and batch meta-data are in your ERP, Automation data are in an Historian database (and are even not aware of batch production start and stop), Environment data are in an EMS. Microbiology data are yet in a separate system. How are you going to make these data silos talk to each other? How are you going to consolidate these data for joint analyses? The Data warehouse module of Hercule interfaces all these different systems, to reconstitute a unified electronic batch record. And it does so in a gentle way: your data silos are left untouched, but push new data in the data lake through a standardized API. This data lake can then be used by other modules of Hercule to perform all their data analyses and visualizations. The API can also be used to export datasets for other uses, for example to feed your favorite internal statistical software.

Data Entry Module: digital forms.

Either because you have an already long legacy, or because you start transferring a new process from R&D to production, there might be part of your production activities that are still generating unstructured data (paper form or free-text word reports) or structured but uncaptured data (stand-alone xls files). How can the information contained in these files be part of your data-driven strategy? The Data Entry module of Hercule allows you to progressively (at your pace) replace these legacy elements by structured data that can feed your data analyses just as any other piece of data.

Form Editor

This module is the companion of the data entry module allowing you to work autonomously. Your procedures are tailored for paper-based form at some points and you do not want to start modifying them? No worry, Data Entry module have workarounds to fit into this frequently encountered situation. A second module, the Form Editor will allow to you create, edit, customize your forms.

User management.

Configure roles and access per user and/or per process. Authentication within Hercule can talk to your own authentication server for single sign-on.