You don't have an inventory problem. You have a visibility problem.

Somewhere in your facility right now, there's a scientist looking for a reagent that's sitting on a shelf fifty feet away. There's a production batch waiting on a material that expired last week. There's a purchase order going through for something you already have three of. This isn't chaos. It's the natural result of disconnected systems.

Your ERP knows what you ordered. Your LIMS knows what samples exist. Your freezer monitors know what's cold. Your warehouse system knows what shipped. But none of them talk to each other, and none of them know what's actually on the shelf right now. That knowledge lives in spreadsheets, paper logs, and the heads of people who might be on vacation when you need them.

Every organization starts the same way. Someone creates a spreadsheet to track inventory. It works for a while. Then someone forgets to update it. Then someone creates their own version because the original is "always wrong." Soon you have Katie's sheet, the Lab 4 sheet, the Master Sheet (old), and the Master Sheet (new) that nobody trusts either. The fundamental problem is that spreadsheets track what people remember to enter, not what actually happens. A scientist grabs a vial and forgets to log it. Someone moves a box to a different shelf and doesn't tell anyone. A shipment arrives and sits in receiving for three days before anyone updates the system. The spreadsheet diverges from reality within hours of being created.

Container-level tracking in physical space

Most inventory systems track SKUs. They know you have "Taq Polymerase" and they might even know you have "47 units." But they don't know that 12 of those units are expired, 8 are in a freezer that's been alarming for two days, and 3 were already consumed last week by someone who forgot to log it.

Seal tracks individual containers. Not just "Taq Polymerase" but "LOT-2024-0892, 250µL aliquot, currently in ULT-03 on Shelf 2 in Box C, received January 15th, expires March 2025, 180µL remaining, last used March 12th by J. Smith for Batch B-2024-041." This granularity matters because GMP requires it. When a lot is recalled, you need to know exactly which containers you have, where they are, and what they touched. When an auditor asks about chain of custody, you need to show them a complete history. When a scientist needs material, they need to find a specific container that's released, unexpired, and actually present—not a theoretical inventory count that may or may not reflect reality.

Your facility has structure. Buildings contain rooms. Rooms contain storage units. Storage units contain shelves or racks. Shelves contain boxes or positions. Every container lives at a specific address in this hierarchy. Seal models your physical world exactly as it exists. You define your hierarchy once—however many levels make sense for your operation—and then every container has a precise location. Not "somewhere in Lab 4" but "Lab 4, Freezer ULT-03, Shelf 2, Position C-7." This structure enables everything else: seeing what's in a specific freezer, finding the closest unexpired container of what you need, tracking how containers moved through your facility over time, identifying everything that was in a room when an environmental excursion occurred. The hierarchy is flexible because facilities are different. A small lab might have Site → Room → Shelf. A large manufacturing plant might have Site → Building → Floor → Room → Zone → Storage Unit → Rack → Shelf → Position. You model what you have.

Status controls and enforcement

Not all inventory is equal. Raw materials arrive and need to be tested before use. Some materials are released for R&D only, not GMP. Some lots get quarantined pending investigation. Some containers are reserved for specific projects. Seal enforces status controls at the point of use, not just in reports. A container that's "Quarantined" cannot be selected for any operation. A container that's "Released - R&D Only" cannot be selected for GMP batches. A container that's past its expiry date cannot be selected at all—it doesn't even appear as an option.

This enforcement happens through the scan-to-use workflow. When an operator needs material, they scan the barcode. The system checks status, expiry, and whether this is the right material for this step. If everything passes, the operator proceeds and the system automatically deducts the quantity used. If anything fails, the system blocks them with a clear explanation of why. There's no checkbox to override. There's no "continue anyway." The control is architectural. This is what makes the difference between a tracking system and an enforcement system.

Expired materials in GMP operations cause deviations. Deviations cause investigations. Investigations cause delays, rework, and sometimes rejected batches. Most organizations manage expiry through reports—someone runs a weekly report of materials expiring in the next 30 days, emails the list to relevant groups, and hopes people review it. Sometimes they do. Often they don't, or they do it late, or the material gets used before anyone notices. Seal manages expiry through enforcement. Materials approaching expiry generate alerts—configurable by material type, so critical raw materials might alert at 90 days while common reagents alert at 14 days. When material actually expires, it's automatically blocked. The container still exists in the system, but it cannot be selected for use. This happens at midnight on the expiry date, automatically, without anyone needing to run a report or update a status. For operations that use FEFO (First Expired, First Out), the system can require operators to use the container with the earliest expiry date, or at least warn them if they're not.

Stock levels reflect reality rather than theoretical inventory because Seal tracks actual quantities at the container level. You define min/max thresholds for each material, and the system monitors continuously. When stock drops below minimum, the system alerts—or, if integrated with your ERP, can trigger a purchase requisition automatically. Stock level calculations include pending consumption: if three batches are scheduled to use a material next week, the system factors that into available stock. You see not just what you have, but what you'll have after committed work completes.

From receiving through operations

Materials enter your inventory through receiving. Seal captures this from the start, creating a container record when material arrives and tracking it through quarantine, testing, release, storage, use, and eventual disposition. When a shipment arrives, receiving logs the containers with their lot numbers, quantities, and storage conditions. The system assigns locations, prints labels if needed, and sets initial status (typically "Received - Pending QC"). Materials stay in this status until testing completes and someone with appropriate permissions changes status to "Released." Every status change, location move, and quantity adjustment is logged with timestamp, user, and reason. This creates an unbroken chain of custody from receiving dock to final use. When an auditor asks "show me the history of this lot," you show them—in seconds, not hours.

Inventory doesn't exist in isolation. Materials are consumed by processes. Seal integrates inventory directly with lab and manufacturing operations so that consumption happens automatically and traceability is built in. When a scientist runs a protocol in the ELN, the system prompts them to scan the materials they're using. The lot numbers get recorded with the experiment, and quantities are deducted from inventory. When manufacturing executes a batch record, operators scan materials at each step. The batch record captures exactly which lots were used, and inventory reflects exactly what was consumed. This integration eliminates the reconciliation problem—there's no gap between what the batch record says was used and what inventory says is left. They're the same data, captured once at the point of use. The integration also enables impact analysis: when a lot is recalled or an OOS result is found, you can instantly identify every batch, experiment, and sample that used that lot. This is the kind of question that used to require days of investigation across multiple systems. Now it's a query.

Scanning only works if everything has a barcode. Seal includes label printing for materials, locations, and containers. You define label templates—what information to include, what barcode format to use, what size label to print—and the system generates them as part of receiving or at any point in the material lifecycle. Standard 1D and 2D barcodes are supported; most organizations use a combination of human-readable lot numbers plus a 2D barcode that encodes the full container identifier for fast scanning. Labels can include expiry dates, storage conditions, hazard warnings, or whatever else your operation requires. Mobile scanning is supported for operations that happen away from fixed workstations.

Physical inventory should match system inventory. In practice, discrepancies accumulate—containers get moved without logging, quantities get estimated instead of measured, things get lost or mislabeled. Seal supports cycle counting to identify and resolve discrepancies. You can count by location (everything in Freezer ULT-03), by material (all Taq Polymerase containers), or by random sample. The system generates count lists, captures counted quantities, and flags discrepancies for review. When discrepancies are found, they're resolved with adjustments that require reason codes and approvals. The adjustment history becomes part of the audit trail, documenting not just what the inventory is but how it got that way.

AI and the business case

Setting up material definitions shouldn't mean re-typing from supplier datasheets. Drop your supplier documentation—datasheets, specifications, safety data sheets—and AI extracts what matters. AI parses supplier documentation and extracts material properties, storage conditions, shelf life, and handling requirements. A stack of supplier datasheets becomes structured material definitions in hours, not days. When materials arrive, drop the supplier's Certificate of Analysis and AI extracts lot numbers, test results, expiry dates, and manufacturing dates into structured records. Receiving becomes verification, not data entry. AI matches supplier specifications to your incoming inspection requirements, flagging any gaps—if the supplier tests for potency but you also need endotoxin, the gap is visible before materials arrive. The pattern is consistent: AI extracts, you review the changeset, approved data enters the system. Material setup becomes verification instead of transcription.

Inventory visibility failures have real costs. An expired material used in a GMP batch triggers a deviation that costs $5,000-15,000 to investigate. A stockout stops a $50,000/day manufacturing line. A failed audit observation around inventory control can cost $100,000+ in remediation. Beyond avoiding failures, good inventory management improves capital efficiency. Organizations typically carry 20-30% excess inventory as buffer against uncertainty. Better visibility means less buffer needed. Better expiry management means less waste. Better consumption tracking means more accurate forecasting. Most organizations see payback within months—from avoided deviations, reduced waste, and optimized stock levels. The ongoing benefit is operational confidence: knowing what you have, where it is, and that it's safe to use.