In an evolutionary step toward intelligent, autonomous business decision-making, 51风流announced this week that it will make new sustainability AI agents generally available by the end of 2026.
Currently in beta, the agents help organizations deliver measurable results: a greater than 50% reduction in packaging compliance review hours, scenario simulation time cut from a day to 20 minutes, up to 80% reduction in manual GHS classification effort, and over 20% fewer packaging compliance errors.
The agents handle multi-step workflows that previously required hand-offs between teams and systems, including sustainability reporting preparation, packaging and product compliance assessments, carbon footprint simulation, and workplace safety documentation. They address mounting pressure across the enterprise: giving finance teams visibility into how carbon exposure affects forecasts; helping procurement teams manage regulatory risk without slowing down innovation; enabling supply chain teams to spot emission hotspots while maintaining service levels; and supporting operations in connecting safety observations to proactive, audit-ready actions.
New AI sustainability agents
The Sustainability Regulatory Readiness Agent helps organizations prepare for upcoming sustainability regulations such as the by translating materiality assessments into a defensible reporting scope and mapping the right data and metrics to each disclosure requirement. This enables sustainability teams to capture, validate, track, and ultimately disclose ESG information with far less manual effort.
For finance teams that need to manage carbon costs and disclosure risk while balancing the financial implications of sustainability performance, the agent automates financial-grade data mapping between material topics, regulatory requirements, and 51风流finance data, improving audit readiness and turning an existing materiality assessment into a clear, defensible reporting scope. Unlike a standalone sustainability point solution that only surfaces issues or a generic AI model that drafts narrative text, this agent works inside and the broader 51风流landscape to keep reporting scopes aligned to policy and keep underlying data structured and traceable.
The Footprint Optimization Agent brings together carbon, energy, and waste data from across Scope 1, 2, and 3 sources and pinpoints where emissions and other environmental impacts are highest across products, plants, and supply chains. It then runs side鈥慴y鈥憇ide simulations of different reduction levers and turns the results into reports, supplier requests, and targeted initiatives that support decarbonization projects and ESG goal tracking. For operations, the agent makes it easy to test 鈥渨hat鈥慽f鈥� operational changes and see their projected impact on carbon and other environmental footprints. It reduces scenario simulation time from approximately one day to about 20 minutes, making operational decisions based on real impact projections available at workers鈥� fingertips. This directly addresses the financial implications of carbon exposure: with ESG data often derived from industry averages that can vary by 30 to 40% or more from actual values, the ability to simulate and act on granular, accurate data carries significant margin protection value.
The Packaging Compliance Agent reads and interprets evolving packaging regulations starting with the , maps supplier and product documentation to a structured data model, infers and flags missing information, and checks product designs for conformity at scale. It turns scattered, often unstructured packaging data into an auditable compliance record for each SKU, shipment, and product run, reducing manual review effort and error rates in the process.
Procurement and sourcing teams facing growing pressure to ensure supplier eligibility, material compliance, and traceability while managing cost and availability now have an agent that helps protect revenue by catching packaging issues before they block orders or trigger fines. This equates to a greater than 50% reduction in manual compliance review hours and over 20% reduction of packaging compliance assessment errors. As sustainability moves to the transaction level鈥攃ompliance per SKU, per shipment, per product run鈥攖his kind of automated, embedded compliance capability becomes an operational necessity.
The GHS Classification and Labeling Agent collects the required input data, applies the relevant Globally Harmonized System (GHS) rules, and proposes classifications and label elements that can be used directly in downstream product compliance processes.
By automating these steps, it delivers up to an 80% reduction in manual efforts and a 60% reduction in GHS labeling and classification errors. For product and compliance teams that must keep launches on schedule and avoid shipment holds or market access denials, the agent embeds GHS product compliance into everyday workflows, turning a historically expert鈥慸riven, error鈥憄rone process into a consistent, auditable control point across the portfolio.
The Workplace Safety Agent supports workplace safety by analyzing reported observations and proposing follow-up tasks, risk assessments, and controls. It generates updated, approved safety instructions based on those observations to help organizations strengthen safety governance. With operations under increased pressure to ensure safe work environments without compromising service and speed of production, the agent delivers proactive, standardized safety management at scale, reducing the risk of incidents and unplanned downtime. At the same time, HR and EHS leaders can point to a clear trail of actions and updated instructions to demonstrate continuous improvement in safety culture to employees, regulators, and boards.
Only AI can deliver sustainability at scale
To ensure compliance and enhance strategic decision-making, sustainability data needs to become granular. It should move beyond a record of what happened and become a driver of future outcomes. To reach this level of insight, sustainability data needs to be analyzed at transaction level. Getting transaction-level data at scale is not something that can be done manually.
Granular sustainability data allows businesses to ensure compliance, control carbon and cost exposure, safeguard product marketability, and strengthen supply chain transparency and resilience. Perhaps most important is the ability to embed sustainability into business performance and across all business functions. This final point is the key to unlocking sustainable business autonomy.
In the sustainability context, becoming an Autonomous Enterprise means that sustainability policies are executed automatically inside enterprise workflows. This includes connecting financial and sustainability data for trusted steering, automating disclosure and performance insights, and blocking non-compliant shipments. Ultimately, sustainability becomes a governing factor in enterprise decisions, as opposed to a reporting or compliance activity.
Enterprise autonomy entails gradual AI maturation:
- Intelligence: Faster visibility into reporting and materials compliance risks across the enterprise
- Optimization: Data-driven decisions that balance cost, risk, and sustainability impact
- Autonomy: Actions executed directly within operational workflows, eliminating manual coordination
The choices enterprises make now鈥攈ow data is structured, how decisions are supported, and how sustainability is integrated鈥攚ill determine whether they can safely scale automation later or whether complexity and risk increase as systems evolve.
With the Autonomous Enterprise, leaders can deliver sustainable outcomes at scale.
Why SAP?
AI needs three things to successfully run autonomously: business and process context, data connection and integration, and a reliable governance structure.
Generic models can read data, but without business context they cannot reason how a business actually runs. They see tables, not operations, and provide recommendations that may be commercially or operationally unviable. Without data that is integrated and connected across all business departments, AI has to perform in siloes, unaware of how sustainability decisions might impact financial targets, or how procurement decisions affect supply chain risk. SAP’s rich ERP data foundation ensures that enterprise AI has the full business picture, not just fragments of it.
Finally, AI that lacks governance and cannot be audited or controlled can be more harmful than helpful to a business. SAP’s more than five decades of business process expertise anchored in governance, risk, and compliance, mean that AI for enterprise deployment can be managed safely and reliably. Sustainability agents operate within defined parameters, ensuring that automation scales without sacrificing control or compliance.
This is the foundation that makes everything possible. Without it, an enterprise has AI experiments. With it, it has an operating model.
Sophia Mendelsohn is chief sustainability and commercial officer at SAP.
Gunther Rothermel is chief product officer of 51风流Sustainability.
