At Coupa Inspire 2026, held May 11–14 at the ARIA Resort in Las Vegas, the company made its position unmistakably clear: Coupa is betting its future on agentic AI, and it wants to remove every barrier standing between its customers and adoption, including the one that makes CFOs most nervous: unpredictable compute costs.
Coupa Compose and the Agent Ecosystem
The centerpiece product announcement was Coupa Compose, a comprehensive environment for building, managing, and orchestrating a digital workforce of AI agents across procurement, finance, and supply chain. Compose includes Navi Agent Studio, a no-code interface now generally available, which allows business users to build and configure custom agents without engineering support.
Coupa currently has 20 agents in production under its Navi framework, with a roadmap to reach 65 persona-based agents by January 2027. The May product release expanded the catalog to include agents for sourcing event creation, bid comparison, risk sentinel, sanctions risk, autonomous opportunity analysis, and scenario ranking. The company also introduced Agent Studio, giving customers the ability to build their own agents for company-specific use cases. This is an acknowledgment that no single vendor can anticipate every workflow that matters to a given enterprise.
The framework follows a three-tier philosophy: Assist (make work easier), Advise (make work smarter), and Act (make work happen). The "Act" tier, where agents execute transactions based on predefined guardrails, is where the real transformation potential, and the real risk, lives. Coupa is positioning itself to move customers progressively along that continuum.
The Numbers: Making the Case for Agentic ROI
Coupa came prepared with data to justify the investment. The company claims that for every $1 spent on the platform, customers see $50 in return, with $300 billion saved through automation across its customer base to date. A few of the more striking figures shared during the event:
- Coupa's agents now handle 99% of invoices that pass through its system.
- The company's contract intelligence agent has reviewed over 21,000 contracts in six months, saving what it estimates to be 4.7 years of manual effort.
- On the customer support side, an agent handling supplier calls has driven queues down 73%, saving roughly 320 hours per month, while case loads are being resolved 41% faster overall.
- Across the broader customer base, Coupa claims 5 million hours saved.
- On the internal engineering front, the company noted that its teams are converting documentation to code in three days versus a prior baseline of 16 weeks, and delivered Digital Operational Resilience Act (DORA) compliance in just six weeks, timeframes that, if accurate, are significant.
The trajectory is clearly one of a company putting real resources behind proving that agentic AI delivers measurable returns, not just efficiency theater.
Strategic Partnerships and Acquisitions
Coupa made several important partnership and acquisition moves at Inspire that reinforce its agentic strategy:
Rossum Acquisition: Perhaps the most significant announcement was Coupa's acquisition of Rossum, an AI-first leader in intelligent document processing. Rossum brings a proprietary domain-specific language model trained on tens of millions of documents. The integration is aimed squarely at strengthening Coupa's invoice-to-pay pipeline, where unlocking data trapped in unstructured documents has long been a bottleneck. This is a smart play — doubling down on the transactional intelligence layer that agentic workflows depend on to operate autonomously.
Celonis Partnership: Celonis joined Coupa as an elite partner, bringing its process intelligence capabilities into the Coupa App Marketplace. The integration combines Coupa's spend management data with Celonis' operational analytics to create digital twins of business processes, helping identify where value leakage is occurring and where spend can be optimized. The combination of Coupa Compose and Celonis essentially allows organizations to take a process, understand its variations through a digital twin, and then deploy agents to act on the optimization opportunities.
Thomson Reuters Partnership: Thomson Reuters also elevated to elite partner status, expanding Coupa's capabilities in global tax and regulatory compliance. For organizations managing spend across multiple jurisdictions, having embedded tax intelligence within the procurement workflow is a meaningful friction reducer — particularly as regulatory complexity continues to increase globally.
Cirtuo (Acquired 2025): Coupa also showcased its earlier acquisition of Cirtuo, a leader in AI-powered category management, which is now fully integrated into the platform. Cirtuo enables a workflow from spend analysis through opportunity identification to category strategy and sourcing execution — a capability that feeds directly into the agentic sourcing agents Coupa is rolling out.
The Big Bet: AI-Native with Compute Costs Baked In
The headline move from Inspire was Coupa's pivot to what it is calling an "AI-Native" platform with outcome-based pricing. In practical terms, this means that for new clients and upgrades, Coupa is charging a flat rate for AI compute costs into its pricing model. New customers get unlimited usage for the duration of their term, providing a predictable payment stream rather than being exposed to the volatility of token-based pricing.
This is a notable strategic choice. As processing costs for large language models and agentic workloads continue to climb across the industry, many enterprise buyers remain hesitant to commit to AI-driven workflows precisely because they cannot forecast what those workflows will cost at scale. Coupa is essentially telling the market: we will take on that risk so you don't have to.
It is worth noting that some analysts have pushed back on this approach, pointing out the inherent difficulty in measuring value-based pricing in a way that satisfies both vendor and customer. The model is bold, but it remains to be seen how cleanly it translates once adoption scales across the customer base.
Catalyst: Tackling the Adoption Problem
One of the more pragmatic announcements was Coupa Catalyst, a new services offering that embeds forward-deployed engineers (FDEs) and business consultants directly into customer teams to accelerate AI adoption. This is a frank acknowledgment of a reality across the enterprise software industry: building the technology is not enough if customers cannot figure out how to implement it effectively.
Catalyst includes role-based learning resources, all offered free with no gate or fees. The company is also investing in training its partner ecosystem through the Partner XChange Program, equipping system integrators and consultants to deliver agentic support at scale. This is the kind of unsexy-but-essential infrastructure investment that often separates successful platform plays from ones that stall at the proof-of-concept stage.
Spend Pulse and the MIT Data Lab
One of the more intellectually interesting developments is Spend Pulse, the evolution of Coupa's analytics solution. Powered by data from Coupa's $10 trillion in cumulative spend and developed in collaboration with the MIT Data Science Lab, Spend Pulse functions as an intelligence layer that matches market data, internal data, and community data to surface actionable signals.
The MIT collaboration produced the Business Spend Index, a macroeconomic indicator built from Coupa's community data. The company claims this index out-predicted the PMI (Purchasing Managers' Index) by three months — essentially showing that real transactional data from Coupa's network captured economic shifts before the traditional survey-based indicators did. If this holds up under scrutiny, it represents a genuinely differentiated data asset, one built on the strength of Coupa's two-sided buyer-seller network operating in private markets where data moves faster than public reporting.
Investment Areas and Forward Look
Coupa laid out several key investment areas for the coming cycle.
- Sourcing and supplier marketplace capabilities are being positioned as the area of biggest impact for agentic AI, with enhanced discovery and matching.
- Contract management is being reframed as "strategy-to-contract" rather than the traditional "source-to-contract," reflecting a more upstream role for AI in shaping procurement strategy.
- Invoice-to-pay is being strengthened through the Rossum acquisition and Thomson Reuters partnership.
- The company also introduced LYRA, an agentic support bot, and highlighted the broader vision of autonomous spend management. This is a future state where AI agents handle the full cycle from intake through execution with human oversight at defined checkpoints.
The Analyst's View
The numbers Coupa shared, $30 to $40 million in savings per $1 billion in spend, with claims of doubling that with agentic capabilities are attention-getting. The 50:1 return ratio and $300 billion in cumulative customer savings provide a strong foundation for the value narrative.
What is harder to dismiss is the strategic coherence of the moves. The Rossum acquisition, the Celonis and Thomson Reuters partnerships, the Cirtuo integration, and the Compose/Catalyst launches all point in the same direction: Coupa is building toward a vision of autonomous spend management where AI agents do the work and humans set the guardrails. The investment in Catalyst and free learning resources signals an awareness that technology alone will not drive adoption. Customers need help with the journey and humans in the loop are critical to its success.
The agentic AI wave is still in its early chapters. For now, Coupa has positioned itself at the front of the pack in enterprise spend management's agentic future. The strategy is clear, the investments are substantial, and the early results are promising. The next 12 to 18 months will tell us whether the execution matches the ambition.
Disclosure: Incisiv attended Coupa Inspire 2026 at the company's invitation.




