Key Takeaways
- Coupa’s 2026 report claims average savings of $30 million to $40 million per billion dollars of spend, based on vendor-authored benchmarks rather than independent audits.
- Zycus Merlin AI projects $48 million to $90 million in annual value for a $1 billion enterprise, with its Generation 3 agentic AI completing 75-92% of transactions without human intervention, according to the company.
- Medius cites Ardent Partners research for its cost-per-invoice benchmark, giving it more independent grounding than most vendor figures in this category. Six procurement platforms have published their most granular ROI claims to date in 2026, covering vendor spend reductions, sourcing cycle compression and invoice processing costs. The catch: most figures originate from vendor benchmarks, not independent audits. What follows is an assessment of what each platform is actually claiming and where the numbers deserve scrutiny.
Coupa: Spend at Scale
Coupa reports that structuring intake alone produces a 19.5-percentage-point drop in manual invoice review, with upstream investments in sourcing and contracts compounding value further down the cycle. Users report significant time savings on routine spend reporting, according to the company. Actual results will vary by spend category, data quality and implementation depth.
Zycus Merlin AI: Agentic Source-to-Pay
Zycus projects $48 million to $90 million in annual value for a representative $1 billion enterprise. The specific claim to watch is the touchless rate: Zycus says Generation 3 agentic AI completes 75-92% of transactions without human intervention.
Merlin Agentic Sourcing targets a sourcing cycle time reduction of up to 60% and claims 8-15% in benchmark-driven negotiation savings. These are ceiling figures from the vendor’s own materials. How consistently they hold across diverse supplier bases and commodity categories is harder to verify from public material. Teams evaluating the platform should press for customer-specific data rather than benchmark averages. The risks of agentic AI in high-stakes procurement workflows are worth stress-testing before committing to a touchless rate target.
SAP Ariba: Native Integration
SAP rebuilt Ariba on its Business Technology Platform in 2025-2026, delivering native S/4HANA integration without middleware. The Joule AI copilot is embedded across procurement workflows. A Joule Bid Analysis Agent, expected in the first half of 2026, will automatically compare supplier bids factoring in total cost, shipping and payment terms, generating summaries that highlight trade-offs.
For enterprises already running on SAP’s stack, the native integration removes a genuine friction point. Middleware elimination reduces both latency and the data-consistency problems that undermine AI-assisted decision-making downstream. The bid analysis functionality, once live, addresses a specific bottleneck in direct materials sourcing where manual bid comparison is slow and error-prone. The case for SAP-native enterprises is straightforward: lower integration cost, and AI tools that inherit cleaner data from the outset.
JAGGAER: Proactive Sourcing
JAGGAER cites Gartner research (May 2026) finding that AI-driven cycle compression can reduce the full pre-negotiation phase from over 12 months to 20 weeks, positioning its JAI concierge as already delivering on that shift. That is a specific, testable claim, though it originates from Gartner’s general research rather than JAGGAER-specific customer data; procurement teams running multi-stage sourcing programmes should measure it against their own baseline cycle data.
Beyond sourcing, JAGGAER’s JAI automates requisition triage, purchase order creation and supplier onboarding end-to-end, grounded in company policies. The supply chain risk angle, surfacing single-source dependencies before they become crises, is where the platform’s positioning is most differentiated. Most procurement AI leads with cost reduction; flagging concentration risk upstream is a different value proposition, and one that resonates in post-pandemic supply chain planning.
Medius: AP as a Control Point
The numbers Medius cites are specific: invoice processing cost drops from an industry average of roughly $9.87 per invoice, per Ardent Partners research, to roughly $2.81 per invoice, a reduction of about 70%. Touchless processing rates reach 96.3% for PO invoices and 99.5% for non-PO invoices, according to the company.
The Ardent Partners sourcing gives the cost-per-invoice figure more credibility than a purely vendor-originated benchmark. The touchless rates, however, are Medius’s own. In 2026, Medius is extending into cross-product agents spanning contract-to-invoice resolution and invoice-to-payment orchestration, moving up the value chain from pure AP automation toward a broader source-to-pay position. For finance teams where AP volume is high and headcount is constrained, the cost-per-invoice reduction is the most immediately legible return. This is also consistent with broader patterns of AI delivering measurable workflow savings when applied to high-volume, rules-based processes.
Workday: Contract Intelligence
Workday Strategic Sourcing incorporates contract lifecycle management capabilities from its Evisort acquisition. One global procurement team cited in Workday’s materials reported $21 million in cost avoidance and value creation after adoption, a 75% reduction in time to extract key contract information, and the ability to process 40% more contracts 65% faster, amounting to a claimed 77x ROI.
The case study is compelling but carries the standard caveat: a single unnamed organisation, selected and published by the vendor. The structural argument is more durable than any single data point. CLM tools that sit natively inside the procurement stack, rather than as a separate legal-team system, surface contract obligations, renewal windows and commercial term deviations at the point where sourcing decisions are made. That integration is where leakage typically occurs in manual processes, and it is where the Evisort capability is most directly applicable.
Ivalua: Cross-Function Orchestration
Ivalua reports that AI can deliver substantial potential savings on total spend, contingent on process maturity and orchestration depth. The platform’s agentic layer identifies risks, delays and off-contract behaviour early, enabling dynamic adjustments before issues escalate. Case study data shows purchase order and sourcing cycle time reductions of 30% or more through automated approvals and task routing.
The orchestration pitch, connecting fragmented procurement, legal and finance workflows into a single decision layer, addresses a real enterprise problem. Large organisations running separate systems for sourcing, contract management and AP face data consistency problems that compound at every handoff. Ivalua’s argument is that the coordination layer, not any single AI feature, is where the durable savings accumulate. That is precisely the framing boards are now demanding: system-level ROI, not feature-level claims.
Originally published at https://autonainews.com/zycus-projects-up-to-90m-in-annual-savings-for-1b-enterprises/
Top comments (0)