Learn / Industry
Lot tracking and food safety audits — what the auditor actually checks
SFDA, ETA, EFSA, and global food safety auditors look at the same five things: receipt traceability, production records, finished-goods identification, where-used queries, and recall response time. Here's what each one requires and how AION's lot tracking handles it.
A food safety audit isn’t a paperwork exercise. The auditor is asking one question in five different ways: if something went wrong, would you know fast enough to act?
For an F&B factory, “something went wrong” means contamination in a raw material lot, a deviation from the recipe, an under-temperature pasteurisation, a packaging defect that lets oxygen in, or a customer complaint that suggests a quality issue. The audit checks whether you’d catch the issue, identify the affected products, and recall them before harm spreads.
This article walks through what auditors actually check, what evidence they want to see, and how integrated lot tracking changes the conversation.
The five things an auditor checks
1. Receipt traceability. When a shipment of mango concentrate arrived from Cairo Citrus Co. last quarter, can you show: the supplier’s lot reference, the manufacturing date, expiry, certificate (halal, food safety, organic if applicable), and the temperature on receipt? Yes or no.
2. Production records. When you produced job JO-2026-00125 of Oasis Mango Juice 1L, can you show: which raw material lots were issued, the routing operations completed (pasteurisation temperatures, hold times, blending parameters), the QA test results, the output lot identifier, and the operators involved? Yes or no.
3. Finished-goods identification. A bottle of Oasis Mango Juice 1L on the shelf has lot number FG-SA-MNG-1L-2026-0512-A on the label. Can you take that label and reconstruct the production history — which job, which day, which raw material lots, which operators, which QA tests? Yes or no.
4. Where-used queries. Given a raw material lot you’ve identified as problematic (supplier announces a recall, your own QA finds an issue), can you list every finished-goods lot that contains material from that raw lot, and every customer who received any of those finished lots? Yes or no.
5. Recall response time. Stopwatch test: from the moment you’re informed of an issue to the moment you can issue specific recall instructions (these lots, these customers, these quantities) — how many hours? Targets typically range from 4-24 hours depending on certification.
Auditors check these one by one. Each “yes” is documentation backing it up. Each “no” is a finding.
What integrated lot tracking gives you
In AION, all five answers exist as data, not as paper trails:
Receipt: Every PO receipt creates lots with supplier lot reference, dates, certificate IDs, and configurable additional fields (temperature, condition, photographs). Audit query: filter receipts in the period, all data exportable.
Production: Every job order’s operations are logged with timestamps, operator IDs, and parameter readings where instrumented. Material issue transactions record exactly which lots went in. QA hold and release events are timestamped. Audit query: open the job’s batch record.
Finished goods: Every output from production creates a lot with a unique identifier. The lot label includes the lot ID, production date, expiry, and a verification code that the QA team can validate. Audit query: scan the label, see the full chain.
Where-used: Bidirectional queries from any lot. Given FG lot X, show all consumed raw material lots. Given raw lot Y, show all FG lots and downstream customers. Audit query: under 60 seconds for a typical factory.
Recall: From the moment a problem lot is identified, the data side of the recall is minutes, not hours. Identify affected FG lots, identify shipped customers, compile quantities, draft customer notifications. The physical recall (logistics, customer contact) takes longer; the data side stops being the bottleneck.
A worked example — auditor visiting Oasis Fresh
The certification auditor arrives Monday morning. They want to see:
10:00am — Receipt traceability check. They pick a recent supplier receipt at random. The CFO opens AION → Procurement → Receipts. The receipt shows: Cairo Citrus Co., supplier lot CC-2026-0432, mango concentrate 500kg, manufactured 2026-03-15, expiry 2026-09-15, halal certificate ID 2026-CC-0234, food safety certificate attached, received 2026-04-02 at 4°C. The auditor checks the certificate dates. Pass.
10:30am — Production records check. They pick a recent job order. AION → Manufacturing → Jobs → JO-2026-00125. The batch record shows: produced 10,000 bottles of Oasis Mango Juice 1L on 2026-05-12 across the 6am-2pm shift. Operators: Mahmoud El-Sayed (production lead), Abdul Rahman (operator), Ramesh Sharma (operator). Materials issued: concentrate lot CC-2026-0432 (130kg), sugar lot SG-2026-0089 (95kg), citric acid lot CA-2026-0042 (0.5kg). QA tests: pH, Brix, microbiological — all pass. Pass.
11:00am — FG identification check. They walk to the warehouse, pick a bottle off a pallet. Label shows lot FG-SA-MNG-1L-2026-0512-A. AION → Inventory → Lots → search FG-SA-MNG-1L-2026-0512-A. Shows: produced in JO-2026-00125 (same job as above), in stock 8,400 bottles, shipped to Panda Retail Co. (1,200 bottles 2026-05-13), shipped to Carrefour Saudi Arabia (400 bottles 2026-05-14). Pass.
11:30am — Where-used query. The auditor poses a hypothetical: “Supplier announces mango concentrate lot CC-2026-0432 is recalled. What would you do?” The CFO opens AION → Inventory → Where Used (Backward) → input CC-2026-0432. Results in 15 seconds: lot was consumed in 3 jobs (JO-2026-00125, JO-2026-00127, JO-2026-00134), producing 5 FG lots. Cross-reference: those FG lots went to Panda, Carrefour, Lulu, BinDawood, Movenpick Hotels, and Riyadh Sugar Refineries — 23,200 bottles total still potentially in circulation. Pass.
12:00pm — Recall response simulation. Auditor asks: “If you decided to recall, what would your next 60 minutes look like?” CFO walks through: generate customer notifications from AION (5 min), assemble logistics plan with 6 customers (30 min), notify SFDA via the regulatory portal (15 min), update internal stakeholders (10 min). Total: 60 minutes from decision to action plan. Pass.
The audit takes 2-3 hours instead of 2 days. Findings: zero. Certification renewed.
What this looks like without integration
The same audit at a factory running on Excel and paper:
- Receipt traceability: physical supplier folder. The lot reference is in the supplier’s paperwork. Certificate dates require checking the binder. 30 minutes per receipt.
- Production records: production binder for the period. Each job has a paper batch record. Cross-referencing material lots against the receiving paperwork is manual. 1-2 hours per job sampled.
- FG identification: the lot on the label has to be looked up in a separate FG lot register. The link to production is via the date-on-label and the production log book. Manual reconstruction.
- Where-used query: cannot be done quickly. Factory has to compile the answer over hours by chasing through three or four physical records.
- Recall response: 8-16 hours typical to assemble the answer, then the physical recall starts.
The audit takes 2 days. Findings: typically several gaps — lots that couldn’t be traced cleanly, certificate dates that were unclear, batch records with missing operator signatures. The certification body issues a corrective action plan with deadlines.
What “integrated” really means
It’s not about having lot data somewhere. Most factories have it somewhere. Integrated means:
- One lot ID, shared across receipts, production, inventory, sales, and accounting
- One source of truth — no parallel spreadsheets that drift from the system
- Real-time updates — when a job order issues material, the lot is consumed in the system at that moment, not at end-of-day data entry
- Bidirectional queries — from any lot, walk in either direction
- Audit trail per transaction — who, when, what changed
AION provides this because lot tracking is at the core of the inventory module — not bolted onto a financial-first system.
What auditors are increasingly looking for
The frontier of food safety auditing is digital evidence. Specifically:
- Temperature logs from cold chain (auditors want timestamped sensor data, not “the warehouse manager attested it was cold”)
- Photo evidence at receipt (auditors want photos timestamped to receipt, not “we took photos sometimes”)
- Real-time clearance integration (where-used queries done in seconds, not hours)
A factory with integrated lot tracking is naturally positioned for this evolution. A paper-driven factory has to build new workflows for each new requirement.
For the costing side of the lot story — how lot tracking interacts with average cost computation — see WIP valuation — the hidden number on your F&B balance sheet. For halal certification specifically (which is a parallel certification regime), see Halal traceability built into your ERP.
See this in the Oasis Fresh demo
Log into the Oasis Fresh (Saudi) BG as cfo.saudi
Common questions
What's the difference between batch and lot in food manufacturing?
Often used interchangeably. Strictly, a lot is an inventory unit (a quantity of material received or produced as a unit); a batch is a production unit (a job order's output). One job order produces one finished-goods batch, which becomes one or more inventory lots. AION uses 'lot' for the inventory unit and 'job' or 'batch' for the production unit; they're related but distinct.
How long do food safety regulations require lot records to be kept?
Varies by country and product category. SFDA in Saudi requires minimum 2 years for non-perishable, longer for high-risk categories. Egypt's regulations follow similar lines. Many factories keep 5+ years as a safety margin and for export-market requirements. Electronic records in an ERP make this trivially compliant — physical paper records become a fire/water/loss risk after a few years.
What's a typical recall response time for a well-instrumented factory?
Under 4 hours from incident notification to identification of affected lots, customers, and quantities. Some certifications (BRC, IFS) require formal recall drills annually with specific response-time targets. AION's where-used queries make the data side of recall response trivial; physical recovery (truck routes, customer notification) is the actual time-consuming part.