How sustainability data lays the foundation for your AI strategy

Inside many manufacturers, two areas of investment are moving in different directions.

AI investment is growing. Sustainability and product-data work is facing greater pressure to demonstrate its value. These may look like separate conversations. In practice, they depend on much of the same data.

The product, supplier, material and lifecycle information collected for LCA, product carbon footprints and compliance is also part of the foundation AI needs to support credible product decisions.

When that data is fragmented, AI can produce answers quickly without the full product context. When it is connected and traceable, the same foundation can support decisions across product development, compliance, cost, sustainability and supply-chain risk.

What has your sustainability team already built?

Producing an LCA, product carbon footprint, compliance assessment or carbon report requires more than running a calculation.

Sustainability and product stewardship teams may already:

• Map products to components and materials
• Collect information from suppliers
• Bring together data from PLM, ERP and procurement systems
• Reconcile records that do not agree
• Identify gaps in product and supplier information
• Document calculation methods and assumptions
• Check whether results can be traced and defended
• Repeat analyses across multiple products

These capabilities are not the whole of AI readiness. AI programs also need appropriate technology, governance, security and ownership. But they are part of the data work AI depends on.

This does not mean relabelling every sustainability project as an AI initiative. It means examining the data foundation already being built before treating AI readiness as a separate program.

Why does AI need connected product data?

AI works with the information it can access. In manufacturing, the full product picture rarely sits in one place.

Product structures may sit in PLM. Supplier information may be held in procurement tools and spreadsheets. Cost data may sit in ERP. Material and lifecycle information may be managed by sustainability or product stewardship teams.

Each source contains part of the answer. None necessarily contains the whole.When those sources remain disconnected, AI cannot see a complete and consistent product view.

Common problems include:

• Different systems describing the same product in different ways
• Incomplete supplier and purchased-part information
• Materials recorded at inconsistent levels of detail
• Teams working from different assumptions or baselines
• Data being reconciled manually for each analysis
• Outputs that cannot be traced back to their sources

AI can analyse an individual source quickly and still produce an incomplete answer.

Speed cuts both ways. Better data allows AI to analyse more information across more products. Weak data allows it to reproduce incomplete or inconsistent results at the same scale.

Much of the work that makes AI useful happens underneath the interface. It involves connecting systems, resolving conflicting records, filling gaps and creating product models that can be used more than once.

How do you know if your product data is ready to support AI?

These five questions offer a useful starting point:

1. Does sustainability data sit across several systems without a consistent product view?
2. Is it difficult to obtain consistent, verifiable information from suppliers?
3. Does the team spend more time preparing data than analysing it?
4. Can LCA and product carbon footprint results be traced and defended?
5. Does every reporting cycle require much of the same data work to be repeated?

If several sound familiar, the organisation may need to strengthen its product-data foundation before scaling AI-supported analysis.

This is not a formal maturity score. It is a way to identify where the immediate limitation may sit. The problem may not be whether the AI model can perform the task. It may be whether the information provided to the model is complete and reliable enough to support the result.

What can connected product data make possible?

Supplier data provides one practical example. When Microsoft developed a new LCA methodology with Makersite, the share of its total product carbon footprint calculated using suppliers’ primary data rose from an average of 20% to over 70%.

The approach also reduced modelling time and inconsistencies associated with practitioner decisions.

The value was not simply a faster final calculation. Microsoft created a more representative product model based on deeper supplier and material information.

Connecting PLM, ERP, supplier and sustainability data can also bring environmental and compliance information into the design stage, before key product decisions are fixed.

A shared product-data foundation can help teams explore questions such as:

• Which material options could reduce environmental impact
• Which substances create compliance exposure?
• Where are the main product-cost drivers?
• Which suppliers or materials introduce supply-chain risk?
• How could a material substitution affect cost, compliance and environmental impact?
• Which products in the portfolio should be reviewed first?

These questions span product development, procurement, compliance, sustainability and finance. The same underlying information can support several teams, even when each team is asking a different question.

Why does reuse matter?

Product and supplier data becomes more valuable when it can support more than one analysis. A product model containing component, material and supplier information may contribute to:

• Lifecycle assessment
• Product carbon footprints
• Compliance analysis
• Material substitution
• Product costing
• Supply-chain risk analysis
• Portfolio screening

The value is not limited to the first calculation the data was collected to complete. A reusable foundation allows different teams to examine different questions without rebuilding the underlying product picture each time.

It can also reduce the risk of teams reaching different conclusions because they are working from different product records, supplier information or assumptions.

How can sustainability teams broaden the business case?

Adding the word “AI” to an existing sustainability project is unlikely to make the case stronger on its own. A more credible approach is to show how the underlying data work supports decisions across the organisation:

Identify who needs the same data
Sustainability, product development, procurement, compliance and finance may depend on overlapping product and supplier records. Making that overlap visible can help position the work as shared product-data infrastructure rather than an isolated sustainability project.

Connect the data to specific decisions
The value becomes clearer when the business case explains what teams will be able to assess or decide.

For example:
-Compare material alternatives
-Identify compliance exposure earlier
-Understand product-cost drivers
-Review supplier and material risks
-Evaluate environmental impact during design
-Prioritise products for further analysis

This is more specific than saying that a project will simply “enable AI.”

Show the work underneath the output
The report, dashboard or AI response is the visible result. Producing a trustworthy result may require teams to:

-Connect data across systems
-Resolve inconsistent records
-Fill product, supplier and material gaps
-Document assumptions
-Build reusable product models
-Apply consistent calculation methods

Making this work visible helps explain why credible AI requires more than selecting a model or adding a new interface.

Show what can be reused
A product model developed for sustainability analysis may also support compliance, costing or supply-chain risk. Showing that reuse gives the investment a broader organisational case.

It does not guarantee access to AI funding. Budgets and ownership will differ between companies. But it can create a more relevant conversation with the teams responsible for data, product development and AI priorities.

Does regulatory uncertainty change the case?

Regulation has given many manufacturers a clear reason to invest in sustainability and product data. When requirements or timelines become uncertain, the immediate compliance argument can become harder to defend.

That does not resolve the underlying data problem. Manufacturers still need reliable product, material and supplier information to assess sustainability, compliance, cost and supply-chain risk.

The same foundation is also needed if AI is expected to support those decisions. The reason for investing may shift. The underlying data need does not.

Two questions AI pilots need to answer

AI pilots often combine two separate questions:

• Can the technology perform the task?
• Is the organisation’s data ready to support the task?

A polished demonstration may answer the first question, but it does not show whether the underlying data is complete, traceable and reusable enough to answer the second.

A useful pilot can also test whether the data is:

• Complete enough for the intended decision
• Traceable to a source
• Consistent across products
• Repeatable without extensive manual work
• Suitable for use beyond a small demonstration

This helps distinguish between a promising interface and a process the business can trust and use at scale.

The business case is broader than one function

Sustainability data and AI investment should not be treated as unrelated conversations.

Product structures, material intelligence, supplier data and lifecycle insights are already part of the information foundation AI needs to support manufacturing decisions.

The question is not only whether AI can perform a task. It is whether the underlying product and supply-chain data is connected, traceable and reusable enough to support decisions the business can trust. That foundation is what makes both sustainability and AI investment worthwhile.