Skip to content
Product data & engineeringFood & Beverage

Set shelf life from storage tests

demi puts the microbiology, pH and sensory results of your storage tests beside earlier studies and proposes a justified date.

Today

Every new recipe, new pack and change of raw material needs a shelf life that holds. Storage tests take weeks, and their results are scattered across the lab system, spreadsheets and reports from external labs. Earlier studies of similar products are hard to find, so the new product gets a shorter date to be safe, which costs remaining shelf life at retail. And since July 2026 the rules for Listeria are stricter: without proof of the shelf life, any positive finding at retail can lead to a recall.

An example

Tarvina cuts sugar and vinegar in its SC 250 g herb sauce, and the pH rises from 4.2 to 4.6 at a water activity of 0.97. The agent shows that Listeria could now grow in the sauce, so proof covering the whole shelf life is needed; no challenge test exists for it yet, and the agent drafts the lab order. From two storage tests of similar sauces and the running test at 7 °C it reads that yeasts reach their limit around day 34, and proposes 28 days instead of the previous 42. The head of quality sets 28 days and releases the lab order.
Fictional company, typical workflow.

Your data, wherever it lives

Today the results of your storage tests sit in the lab system, in a spreadsheet per study and in PDF reports from external labs, while recipes and packs live in recipe management and the HACCP plan and specifications in the QM system. demi reads, understands and links them per product and recipe version, with your access rights. The AI then proposes a justified date in minutes, where someone today spends days searching old studies, and your head of quality decides.

Where the data lives today

  • Databases
  • Office files
  • PLM
  • QM system
  • PDFs
Data LayerOne view of everything, with your rights

How demi solves it

Three layers in this order: connect the data first, then the rights, then the AI.

  1. 1 · ConnectData Layer

    Data sources in this case: the lab system, recipe management, the QM system, spreadsheets and PDF reports from external labs. demi replaces nothing and links every storage test to its recipe version, pack, heat treatment and storage temperature, and to complaints from the market. So everything known about a product's shelf life sits side by side.

  2. 2 · PermitPermission Layer

    Recipes and lab results are visible to development and quality; sales sees only the date that has been set. Only the head of quality sets the shelf life; every decision is recorded with name, date and the studies used.

  3. 3 · ActIntelligence Layer

    The agent ranks earlier storage tests by similarity, reads results from PDF lab reports as well, and shows for each parameter the day it reaches its limit. Limits, safety margin and the classification by pH and water activity are fixed, versioned rules; the language model only writes the justification. Where proof is missing, such as a challenge test, the agent drafts the order to the lab.

How it runs

  1. 1

    For a new recipe, a new pack or a raw material change, development sets up a storage test.

  2. 2

    The agent finds earlier storage tests of similar products by pH, water activity, preservation, pack and storage temperature.

  3. 3

    For each test day it collects microbiology, pH and sensory scores, from external lab reports too.

  4. 4

    It shows when each value reaches its limit and proposes a date with a safety margin, every figure with its source.

  5. 5

    The head of quality reviews the proposal and sets the shelf life.

What changes

  • Shelf lives backed by data, not cautious guesses

  • Earlier storage tests reused instead of repeated

  • Proof for retailers, audits and authorities on demand

What does it bring you?

Check it with your own numbers

Every starting value is an assumption you can change. The calculation sits right below, step by step.

Assumption, editable

Your numbers

Result

+86 hfreed per year

Assumptions and calculation

demi sets storage tests beside earlier studies and proposes a justified date; you set it.

4 h → 2.8 h per case

How does it run today?

What is already digital saves less: the factor lowers the share demi brings.

How we calculate

6 shelf-life studies × 4 h × 30% × 1.0 (manual) × 12 months = 86 h

Value per year, before demi costs

Freed per year+86 h

Assumption: labor cost of $48.62 per hour in US manufacturing in June 2026 (BLS), plus workplace, IT and overhead.

= Value per year, before demi costs

+$5,590

gained per year

An estimate, not a promise. The starting values are assumptions for an industrial company with 150 to 500 people.

Combine with more cases and get the package

Questions about this use case

Which values does demi base a shelf-life proposal on?

On the results of your storage tests: total viable count, yeasts, molds, lactic acid bacteria and, where needed, Listeria, plus pH, water activity, sensory scores and, for modified-atmosphere packs, the gas values. Your quality team sets limits and safety margin; they are stored and versioned. The figures come from the measurements; the language model only explains.

How does demi find a comparable storage test?

By features such as recipe, pH and water activity, preservation, heat treatment, pack and storage temperature. Each match shows where it resembles the new product and where it does not. You decide which studies count as proof.

What do the Listeria rules in force since July 2026 change?

For ready-to-eat foods in which Listeria can grow, "not detected in 25 g" now applies throughout the shelf life, unless you can prove that 100 cfu/g will not be exceeded by its end. demi classifies every product by pH and water activity and shows where such proof is missing. Which study you order is your decision.

Does demi set the shelf life itself?

No. demi proposes a date with a safety margin and the evidence; your head of quality sets and signs it. Every decision is recorded with name, date and the studies used, so you have it at hand in the IFS audit.

See this case with your data

In the demo we show you the workflow on your own systems.

Your configurator

No use cases yet

Your configurator is still empty

Add a use case from any card or case page and see what it brings per year, with the calculation. Free, no sign-up, no commitment: try it out and change any number.

Browse use cases