Build the operation your new facility needs.

1A greenfield programme is not an MES implementation.

A greenfield pharmaceutical facility has one opportunity to make digital operation the original operating model. Once paper records, spreadsheet trackers and departmental applications become the way the site opens, replacing them becomes a second migration while producing. Seal treats the digital facility as part of the facility project: before go-live, the site must be able to receive a material, manufacture and test a representative product, handle exceptions and release the batch without reconstructing the record between systems.

An MES controls manufacturing execution. It should own approved master records, executable batch records, material consumption, process parameters, signatures, exceptions and review by exception.

A facility launch has a larger job. Incoming materials need identity, status, locations and sampling. QC needs specifications, methods, instruments, raw-data capture, stability and Certificates of Analysis. Quality needs documents, deviations, CAPA, change and disposition. Equipment needs qualification, calibration, maintenance and cleaning state. People need roles, training and practical qualification. Those records have to agree before the first GMP batch begins.

This blueprint defines the architecture, dependencies, validation evidence, opening state and proving path that let those capabilities go live as one facility.

1.1Why teams choose Seal for a new facility

Build manufacturing, QC, quality and training together on Seal, starting with one complete receipt-to-release workflow. Neil helps turn the site’s procedures into working steps, fields and checks, which the team can test against its requirements. After go-live, the same platform supports investigations and revised workflows. Each change carries its verification and approval, so the operation can keep improving as the site gains experience.

2Readiness is a single controlled state.

Construction schedules naturally divide work into disciplines. Digital readiness cannot end as six independent completion percentages. A laboratory can be configured while its instrument interfaces remain unqualified. Production can have an approved master record while the material lots needed to execute it do not exist. Training can be assigned while the final controlled procedures are still changing.

Seal makes those dependencies visible as facility gates. In Figure 1, each workstream moves through the same sequence: operating design, configured records, qualified use and reconciled cutover state. Go-live becomes possible only when a complete value stream crosses all four gates, so the opening decision rests on shared decisions proven end to end rather than on departmental completion percentages.

3Design around decisions, not application names.

The target architecture starts with the questions the site must answer:

• May this material be used? Identity, supplier approval, receipt condition, sampling, laboratory disposition, storage and retest state contribute to one answer.
• May this person perform this work? The active procedure, role, training, practical qualification and expiry state determine eligibility.
• May this equipment be selected? Qualification, calibration, maintenance, cleaning, allocation and product-contact history determine availability.
• May this process continue? The approved recipe, captured parameters, in-process results, hold times and unresolved exceptions determine the next action.
• May this batch be released? Manufacturing, QC, environmental context, reconciliations, deviations and approvals resolve into one disposition.

System boundaries should support those decisions rather than split them. Seal can run the warehouse, MES, pharmaceutical QC, QMS, equipment, training and batch release workflows natively. Where an ERP, building system, control system, chromatography platform or other specialist application remains authoritative, the boundary is explicit: which object it owns, what event crosses the boundary, how failure is detected, and which system records the accountable decision.

4Build shared master data with the physical facility.

Digitising forms first creates cleaner silos, so the shared operating model comes before individual screens. Products connect to specifications, master manufacturing records, packaging, stability commitments and release requirements. Materials connect to suppliers, sampling plans, storage rules, tests, recipes and hazards. Equipment connects to classes, locations, qualification, calibration, maintenance, cleaning and eligible operations. People connect to roles, curricula, practical qualifications and signature authority.

This master data removes choices from execution that should already have been controlled. A receiver does not invent a sampling plan. An operator does not select any available scale. An analyst does not search for the likely specification. A reviewer does not decide which evidence ought to exist after the batch ends.

The first configuration deliverable is a governed object model with accountable owners, identifiers, lifecycle states, approval rules and source documents.

That model is built alongside the physical facility. Facility design produces information the operating system needs: room and material flows become locations and allowed movements, the equipment list becomes the asset register, critical environments become monitoring points, process flows become master records and personnel flows become roles and qualification requirements.

The relationship also runs in the other direction. Configuring the operation exposes unresolved design decisions: a sample has no defined handoff, a material has no allowed quarantine location, an equipment class lacks a cleaning state, or a release requirement has no authoritative result. Finding those gaps while layouts, procedures and commissioning plans can still change costs far less than discovering them during performance qualification or the first production campaign.

Seal keeps facility, equipment, process, laboratory and quality configuration connected to the controlled sources and approval decisions that created it. Commissioning evidence can establish initial equipment state without becoming a detached archive.

5Prove one complete receipt-to-release value stream.

The safest implementation unit is not a department or an application module. It is one representative product journey.

Start with receipt of its API, excipients and components, and exercise quarantine, sampling, QC approval, dispensing, production, in-process and finished-product testing, deviations, reconciliation, CoA generation and disposition, with the rooms, equipment, methods, roles and interfaces that path requires.

Then execute exception scenarios deliberately: damaged receipt, failed identity test, expired calibration, wrong material scan, out-of-range process value, delayed laboratory result, OOS investigation, reconciliation difference and blocked release. Readiness means that normal and abnormal work both reach controlled, reviewable states without side spreadsheets or verbal reconciliation.

The resulting value stream becomes the approved pattern for additional products. Shared steps, methods, specifications, roles and controls are reused; reviewers focus on what differs.

QC often determines the real critical path. Specifications and methods must be approved. Instruments must be inventoried, qualified, calibrated and connected at the appropriate data boundary. Standards, reagents, columns, media and consumables need controlled identities and states. Calculations and reporting rules need verification. Analysts need method and instrument qualification.

Incoming materials, in-process controls, finished-product release, environmental monitoring and stability all depend on that foundation. If the laboratory goes live later, the site creates its first permanent interface: manufacturing generates samples one way, QC receives them another way, and release depends on reconciliation between them.

Configure the pharmaceutical QC laboratory alongside receiving and the first manufacturing value stream, with instrument connectivity, source-data review, OOS handling and CoA generation in the proving path.

6Cut over current state and validate for intended use.

Day-one operation requires an accountable opening state. That includes physical inventory and container locations; material status and retest dates; equipment qualification, calibration, maintenance and cleaning state; current training and qualification; active product, material, specification, method and recipe versions; approved suppliers; open quality events; stability commitments; and outstanding work.

Each population needs a source, transformation rule, owner, reconciliation, exception process and approval. The objective is the minimum trustworthy history and current state required to make the next GMP decision correctly, rather than an import of every historical document. Opening inventory reconciles to the physical site, equipment due dates to approved certificates, training to current role requirements and master-data versions to effective controlled documents. Anything that cannot be established confidently enters a visible exception state rather than being loaded as assumed truth.

The FDA describes drug CGMP as minimum requirements for the methods, facilities and controls used to manufacture, process and pack drug products.¹ A digital system supports those controls only when its intended use, configuration, data flows, permissions, records and failure behaviour are understood and shown fit for the operation.

Seal structures validation around the configured value stream. Requirements connect to the decisions and risks they control, tests cover normal paths, exceptions, access, signatures, audit trails, interfaces and recovery, and executed evidence links back to the configuration and requirement it supports.

This avoids two opposite errors: testing every screen with equal effort, or treating platform assurance as proof that the site’s configured process works. FDA describes a comparable risk-based approach, computer software assurance, in CDRH and CBER guidance for production and quality system software used in device manufacturing; it is not a drug CGMP requirement, but the same reasoning applies here.² The quality unit approves intended use, risk decisions, acceptance criteria, deviations and release of the configured system for use.

7Run the programme on its dependencies.

The duration of a facility programme is set by its dependencies, not by a fixed timeline, and it cannot be shortened by running six disconnected workstreams faster. It needs an explicit critical path.

1. Operating design: the first product family, areas, flows, quality decisions, system boundaries, owners and intended use.
2. Configuration foundation: identifiers, roles, locations, materials, products, equipment, specifications, methods, recipes and lifecycle states.
3. Value-stream qualification: receipt through release, including interfaces and exception paths, tested with the people who will do and review the work.
4. Cutover and rehearsal: reconciled opening states, role qualification, an end-to-end rehearsal, closed blocking defects and approved readiness gates.

Work can overlap, but dependencies remain real. A method cannot be qualified before its calculation and instrument boundary are defined. A batch record cannot be proven before eligible materials, equipment, users and samples exist. A cutover cannot be approved before its source populations reconcile.

The facility foundation is shared; manufacturing controls are specific to the modality. Sterile injectables add aseptic interventions, environmental and personnel monitoring and sterilisation evidence; oral solid dose adds dispensing, transformations, packaging and reconciliation; API manufacturing adds process phases, intermediate genealogy, yields and retest management. Those blueprints define how each operation behaves. This one brings them into a common laboratory, quality, material, equipment, training and release architecture.

8The go-live gate is operational evidence.

Before digital day one, the programme should be able to answer yes to concrete questions:

• Can a real user receive, label, locate, sample, test and disposition a representative material?
• Can only qualified people and eligible equipment execute the approved process?
• Can instrument and manual data be reconstructed from source through reported result?
• Do exceptions begin with their operational context attached?
• Can QA see every unresolved release requirement without assembling a packet?
• Do opening inventory, equipment, training and master-data states reconcile to accountable sources?
• Can the site continue safely through an interface failure, correction, rejected transaction or unavailable instrument?

The launch date is a project milestone. The readiness decision is a quality decision supported by evidence.