Taiga Intelligence
Products contain data.
We build understanding.
An evidence-led intelligence architecture. It is designed to understand the connections between products, materials, manufacturing, scientific evidence and legal requirements, right down to the level of individual products across the supply chain.
The information exists.
Transparency doesn’t.
Product information is scattered across manufacturers, material specifications, certification bodies, scientific databases and changing regulations.
Most shop systems store attributes. They do not inherently understand how those attributes relate to chemical restrictions, manufacturing conditions, the validity of evidence or the requirements of individual markets. TAIGA is designed to close that gap.
Watch a product
become a knowledge model.
Choose a reasoning mode. The network highlights the relationships the model would need to examine in order to draw a product-specific conclusion.
Not a chatbot.
A reasoning architecture.
The intended model does more than retrieve documents. It organises product knowledge, determines relevance, checks what evidence depends on and identifies which conclusions are well-founded.
Semantic product representation
Creates typed relationships between products, materials, substances, processes, suppliers, documents and legal jurisdictions. This makes product context computable, not merely searchable.
Contextual applicability
Checks which datasets, certificates and regulations may be relevant to a specific product and target market, including effective dates and scope.
Evidence-based reasoning
Links every proposed conclusion to its evidence, searches for conflicting information and prevents missing data from being presented as certainty.
Reassessment across dependencies
Records which product conclusions depend on which underlying facts. If evidence or rules change, affected statements can be flagged for reassessment.
When a fact changes,
the model knows what matters.
Imagine a substance is reclassified. A static product database does not know which statements made to customers could be affected.
The TAIGA architecture is designed to trace the dependency: substance → material → affected products → applicable rules → supported claims. What we are exploring is not the alert itself, but the reasoning needed to determine its relevance.
Illustrative representationScale is measurable.
So is trust.
A claim such as “over 1,000 data points” only has meaning if it specifies whether it counts attributes, observations, sources or relationships.
An intelligence engine.
Four dimensions of understanding.
We are starting with the new Transparency module for TAIGA customers. The same evidence-led product model opens up two directions: richer shopping experiences and environmental intelligence for manufacturers, brands and retailers.
Evidence first.
Understanding always.
The next milestone is the new TAIGA Transparency module: product-specific information, 12 transparency criteria, traceable sources and uncertainty shown openly. No overall score and no unverified impact figures.
PRIORITY 0 · CUSTOMER MODULEFrom individual decisions
to entire supply chains.
See how the same product intelligence could appear to a customer at checkout and to a business overseeing thousands of products.
Stainless steel saucepan
See the material, how it was made and its life cycle.
Know what’s in
your purchase.
Material composition, manufacturing information and disposal guidance. Shown only where sources support the claim.
The intelligence layer connects product-specific evidence with the claim it can support. Research-based estimates would also need methods, system boundaries and uncertainty information.
Environmental overview
Illustrative view. Primary measurements, secondary data and model assumptions remain separate.
What if material became more efficient?
Try an assumed sensitivity scenario for the selected portfolio.
Illustrative view
Don’t just measure what is known.
Explore what can be inferred.
Our frontier research tackles the difficult problem beneath the surface: reading and reconciling scattered evidence, mapping product relationships, estimating resource consumption and CO₂ from incomplete information, and testing how valid and robust those estimates are.
Understanding the evidence
Knowledge graph, different sources, extraction and standardisation.
Infer & explain
Multi-criteria metrics, explainability and model integration. The former weighted score is not the current customer product.
Deriving resources
Probability models for material, energy and water consumption, benchmarks and validation.
Model & test CO₂
CO₂ derivation, scenarios, sensitivity analysis, robustness and uncertainty testing.
Example range: 5,0 to 14,0 L
This interactive demonstration changes assumed inputs, not real product measurements. The research aims to establish whether such an inference can be tested against reliable benchmarks.
Missing data, domain shifts, missing ground truth and changing regulations are open research challenges. The roadmap describes planned research capabilities, not confirmed results or certification of legal compliance.
Ambitious by design.
Rigorous by necessity.
The real task is not to connect interfaces. It is to develop a robust method that brings together incomplete product information, inconsistent evidence and changing product-related requirements.
This page shows the architecture and product direction. The research roadmap covers evidence integration, explainable inference, resource models and validation.
TAIGA INTELLIGENCE
Understanding always.The next milestone is the new TAIGA Transparency module: product-related information, 12 transparency criteria, traceable sources and uncertainty shown openly. No overall score and no unverified impact figures.PRIORITY 0 · CUSTOMER MODULEHow we checkMade in AustriaCountry of manufacture · View sourcesElectricity mainly from hydropowerProduction · View sourcesSteel body recyclableScrap metal · View disposal informationWhat lies behind this product ↗How do I dispose of the pot?Questions ✧Illustrative exampleFewer promises. More knowledge.02 / EXPANSION IN THE SHOPIntelligence at every
decision point.Evidence-based product knowledge is brought into comparisons, guidance at checkout, AI-assisted product questions and care advice after purchase. We are testing whether better explanations improve trust, interest and purchase completion.PLANNED APPLICATIONSPRODUCT PAGEExplain what matters to the purchase decisionCHECKOUTContextual guidance, with sourcesAFTER PURCHASEKnowledge about care, repair and disposal03 / RESEARCH & DEVELOPMENTUnderstanding what
products consume.We are researching how to derive material use, energy and water requirements across manufacturing, transport and use with probabilistic models. Known inputs, missing data, assumptions and uncertainty are disclosed before a figure is shown.RESEARCH PROGRAMMEMATERIAL FLOWComposition → mass → referenceRESOURCE REQUIREMENTSWater · energy · transport · useMODELLINGScenarios · probability · uncertainty04 / RESEARCH & DEVELOPMENTModelling CO₂.
Explaining uncertainty.We are investigating how to derive CO₂ from incomplete life-cycle data. This involves examining how resource use, process assumptions and alternative scenarios change possible footprints, using sensitivity analysis and scientific review.PRODUCT CONTEXTMaterial · process · transport · useDERIVATION MODELSystem boundaries and assumptions across the life cycleINTERPRETATIONEstimated CO₂e · uncertainty rangeWhat resources does a product require?ILLUSTRATIVE EXAMPLEView the life cycle: extraction → manufacturing → transport → useRaw materialsManufacturingTransportUseWaterUsed in the manufacturing processElectricityEnergy used in manufacturingWood & natural materialsUsed in the packagingMetals & mineralsMain material useCO₂ is linked to resource use.Material choice, electricity, transport and product lifespan feed into possible life-cycle estimates.PRODUCT-RELATED INFORMATIONEvidence, not assumptions.Material composition, manufacturing information and disposal guidance. Shown only where sources support the claim.Linked to a sourceOriginDepends on the evidenceUnknownsShown openlyRESOURCE KNOWLEDGE OF THE FUTUREWhat goes into manufacturing?View the material mass, water and energy requirements across a product's life cycle. Measured inputs remain clearly separate from research estimates.Material requirementsModelled scenarioWater & energyDepends on the methodUncertaintyVisible to customersCO₂ KNOWLEDGE OF THE FUTUREMore than a single CO₂ figure.An assumed lifecycle footprint, with the documented scope, assumptions and system boundaries. Figures would only appear after verification.Lifecycle estimateResearch conceptScenario rangeWith uncertaintyStatusUnder researchExample material mass · tonnesExample water demand · thousand m³Example energy demand · MWhExample CO₂ impact · tonnes CO₂eThousand m³Kitchen & cookwareHome & furnitureTextilesPersonal careOtherThousandLowerMediumHigherExample range:
