IT Orgs Have the Tools To Make AI Trustworthy with Coupa’s September 2026 Product Release

IT Spotlight, September 2026 Product Release

Procurement wants agents that can launch sourcing events. Finance wants instant answers about payments and cash positions. Supply chain wants connected data to guide planning and execution. For CTOs, that means connecting AI to the systems where work happens, giving it access to reliable data, and defining what it can do.

Coupa’s September 2026 release, R46, advances that foundation across all of enterprise spend. Open-standard connectivity brings external AI environments into Coupa workflows. Unified item and supplier data help connect existing systems. New agents execute tasks within defined policies and access controls.

Together, these capabilities give IT leaders a practical way to support growing demand for an agentic enterprise while managing the architecture, security, and administration behind it.

Connect your AI ecosystem to enterprise spend workflows

Your enterprise AI strategy may span Microsoft Copilot, Claude, and custom-built agents. To deliver value across procurement and finance, those environments need a way to access spend data and execute actions within the controls your business already relies on.

R46 introduces Model Context Protocol (MCP) integration through Navi™ Connect, an open-standard framework that connects external AI environments directly to Coupa. With more than 30 out-of-the-box, intent-driven tools across procurement, invoicing, contracts, and expenses, IT can give agents access to defined capabilities for reading data and writing back transactions.

That access follows the user’s permissions. External agents operate within Coupa’s governance framework, and transactions they write back remain auditable through unique revision history identifiers. IT gains a controlled connection between the AI environments employees use and the system where spend decisions become business actions.

For CTOs, this creates a common integration approach that can support a growing ecosystem of agents. It reduces the need to build and maintain a separate custom connection for each AI environment, making it easier to extend enterprise AI into spend workflows as business needs evolve.

Give AI and enterprise workflows a consistent data foundation

Connecting an agent to enterprise data is only part of the work. That data also needs to carry the context required to complete a transaction. An item selected during sourcing must retain the information needed to create a purchase order in its originating ERP. A supplier participating in a negotiation needs a consistent record across the applications involved.

R46 strengthens those connections by unifying item data, linking direct material sourcing to ERP workflows, and bringing supplier records and participation into a shared experience.

Unify item data across multiple ERPs

For enterprises running SAP, Oracle, and Microsoft Dynamics across business units or acquired companies, item information can be fragmented across systems. Bringing it into a procurement catalog requires preserving the attributes each ERP needs for downstream processing.

The Unified Item Repository addresses this through a three-tier model: Item, PlantItem, and SupplierItem. This structure brings item master data into a single, plant-aware catalog while retaining plant-specific and supplier-specific context. Mandatory ERP attributes are stored, indexed, and searchable within Coupa.

For IT, that context matters at the integration boundary. Purchase orders and invoices can draw on the required origin data, helping reduce failures caused by missing attributes. The repository also helps reduce catalog bloat and the reconciliation work involved in maintaining item information across multiple systems.

Connect direct material sourcing to downstream execution

R46 also integrates the Coupa Sourcing Optimization (CSO) sourcing module with the enterprise Item Model. Configurable header data, such as Enterprise ID, helps translate differing backend structures and maintain a continuous data flow from the ERP into sourcing and back to the ERP for purchase order creation.

This connection helps preserve item context as work moves from negotiation to execution. Teams can source against preconfigured item data and carry that information into downstream purchasing, reducing manual reentry and the integration effort required to bridge disconnected stages of the process.

Bring supplier data and access together

R46 also connects Core and CSO supplier databases through the Coupa Supplier Portal. Buyers can search for and create suppliers from Core, while suppliers can participate in both Core and CSO sourcing events through a single login.

A shared supplier record reduces duplicate entry and the administrative effort of maintaining separate databases. Consolidated access also simplifies participation for suppliers, helping them engage across sourcing events without managing separate logins.

Together, these improvements address a foundational IT concern: keeping data consistent as it moves across systems and workflows. They support more reliable transaction processing today and strengthen the data foundation for expanding agentic AI across enterprise spend.

Extend automation with explicit controls and human oversight

As AI agents take on more operational work, IT needs to understand the boundaries of each capability: what data an agent can access, which actions it can execute, and when a person must review the result. R46 applies those controls to specific workflows, giving enterprises practical ways to expand AI adoption.

Policy-governed sourcing execution

Autonomous Sourcing Event Creation converts eligible intake requests into structured sourcing events. The agent resolves the applicable sourcing policy, selects a template, replaces stale template data with actual request details, and recommends qualified suppliers.

It then either publishes the event or routes it to the sourcing team for review, with human oversight for anomalies. Automated decisions remain policy-governed, auditable, and subject to human oversight. This gives procurement greater execution capacity while preserving a review path for work that requires attention.

Human confirmation for contractual obligations

AI-Powered Obligation Tracking & Management extracts commitments from executed contracts and attachments, including payment terms, deliverables, SLAs, and compliance clauses. It converts those commitments into structured records with owners and timelines, supported by tasks, status workflows, and notifications.

Human confirmation is required before AI-extracted obligation records become active in Coupa. That checkpoint lets teams validate the AI’s interpretation before using it to manage contractual commitments, while keeping ownership and fulfillment traceable within the contract lifecycle.

Access-controlled cash intelligence

The Cash Management Data Agent gives treasury users conversational access to account balances, statements, and cash flows. Its scope is read-only, and access follows existing Legal Entity Security Group entitlements.

Users can retrieve financial information through Navi while maintaining the access boundaries already established for that data.

For CTOs, these examples make agent governance concrete. Each capability has a defined scope, with controls suited to the work it performs. That creates a clearer basis for assessing where AI can act autonomously and where human judgment remains essential.

Expense creation agent on mobile

Expenses is one of the most-used areas of Coupa Mobile, and R46 brings Navi into that workflow. The Expenses agent creates expense lines and reports, and mobile users can capture them with the camera or speak to Navi by voice (the agent operates within the user's existing Coupa permissions, and users review the generated expense lines before they submit). The same agent is also available to web users.

Reduce the technical effort behind everyday operations

Enterprise platforms require ongoing configuration as business needs change. For IT and system administrators, even routine updates can create work when incoming data must match rigid import specifications. Mapping fields, correcting formatting errors, and troubleshooting failed uploads can slow implementation and consume specialist time.

R46 simplifies that work with AI-driven Expense Configuration & Mapping. The capability automatically maps fields in imported CSV files to Coupa’s specifications, helping administrators align external data structures with the format the platform requires.

An interactive interface lets administrators review and adjust those mappings. AI handles the initial matching, while administrators retain control over how the data is configured. This makes it easier to maintain expense categories without repeatedly restructuring source files or relying on external integration experts.

For CTOs, the benefit extends beyond a faster import. Easier configuration can reduce demands on technical teams, lower the cost of maintaining expense data, and give administrators greater self-service capability. As organizations expand their use of Coupa, these improvements help keep routine maintenance manageable and technical resources available for broader business priorities.

Manage Coupa Mobile with Microsoft Intune

Many organizations only allow employees to use business apps on their phones when those apps fall under existing mobile application management policies. Coupa Mobile now supports Microsoft Intune app protection on both iOS and Android, so IT can apply the same MAM policies to Coupa Mobile that it applies to its other managed apps.

Admins request enablement through Coupa Support and configure an app protection policy for Coupa in Intune. At sign-in, the app checks the user's Intune credentials and device compliance against the policies IT has set, and then continues to the instance's configured identity provider. Instances that aren't enrolled keep the existing sign-in experience, which lets organizations move business units or regions over on their own schedule.

A platform foundation for the next wave of AI

For CTOs, scaling agentic AI requires a foundation that can support new use cases without adding unnecessary complexity. R46 advances that foundation across enterprise spend: open-standard connectivity for external AI environments, consistent item and supplier data, defined controls for agent execution, and simpler administration.

Those improvements serve immediate business needs while giving IT a clearer path to support future adoption. Procurement, finance, and supply chain can expand their use of AI through shared platform capabilities, with permissions, data context, and oversight built into the workflows where work happens. And with Microsoft Intune support and Navi capabilities such as voice and receipt capture, the same governed AI foundation now extends to the mobile devices employees use every day.

Coupa’s acquisitions of Rossum and Tonkean reinforce the broader vision of a unified agentic workflows across the platform. Looking ahead, planned expansion of AI-native document processing will address complex invoices and other documents, including purchase orders, contracts, and sourcing materials. The broader roadmap aims to bring these acquired capabilities into the unified Coupa experience.

See the September 2026 Release in Action