Mercanis: procurement agents can prepare the RFQ, but buyers still approve the award

External-source research synthesis, not an OfferSignal client result. Based on the retained October 2026 research and source-specific dates.

A procurement team can lose days before a buyer ever compares a quote. The request may be incomplete, supplier records duplicated, certifications scattered and risk information held in separate systems. Mercanis addresses that operating problem as a cloud procurement suite rather than a generic chatbot: intake, sourcing and requests for quotation (RFx), supplier management, contract workflows, risk and compliance, and analytics are presented as connected areas. 3

Agents prepare the shortlist, not the commitment

Mercanis’s supplier-discovery description says a Discovery Agent searches market databases, trade registries, supplier networks, certification sources and related external data to build candidate lists. A Market Intelligence Agent monitors prices and market conditions. The claimed output is a ranked shortlist with scoring rationale, risk flags and market context. Mercanis also states that buyers retain control of consequential decisions, with qualification criteria, escalation thresholds and exception handling defining the guardrails. 6

That makes “agentic” useful but bounded: the system can research, evaluate, verify and recommend within a procurement workflow. The retained evidence does not establish unattended supplier award, contract signature or payment authority. For a manufacturer, the strongest fit is therefore before award—discovering capable suppliers, structuring an RFQ, comparing bids, checking compliance and preserving an audit trail.

What the named cases actually report

Amer Sports is the clearest quantified example. Mercanis reports a four-week implementation, 38% time savings in streamlined indirect procurement and 18% sourcing-spend savings over 12 months. It also claims transparency across global procurement teams, while the customer describes self-service for smaller tenders with category experts retaining strategic control. 5 These are provider-published customer-case results: the page does not disclose the baseline spend denominator, event count, control group, absolute savings or independent audit.

A Mercanis-authored 2026 article adds that Amer Sports grew from four to 31 sourcing events in one quarter and exceeded 2,800 active indirect suppliers. Those figures are also company-reported, with no published before-and-after method. 6

Oventrop is a relevant industrial reference: the manufacturer is described as having more than 1,500 suppliers, over 6,500 articles and more than €65 million in annual purchasing volume for production material and commodities. Mercanis says it digitally mapped supplier-management processes and automated manual work in the context of transparency, data foundations and German supply-chain due-diligence compliance. The case headline advertises four times more offers per sourcing event, but gives no denominator, period or event count. Treat that as a vendor claim requiring validation, not a general performance promise. 7

The industrial boundary is still important

The public material does not prove that the current agents understand drawings, revision-controlled computer-aided design, tolerances, material compatibility, tooling, machine capacity or engineering-change logic for non-stock-keeping-unit work. An illustrative manufacturer should test one governed RFQ with canonical supplier identities, historical spend, a released specification and attachments. The agent can discover and rank candidates, collect comparable bids and surface risk; engineering must confirm technical fit, procurement must approve the shortlist and an authorised commercial owner must approve binding terms.

Data quality is a dependency. Mercanis says fragmented or stale supplier identities, certifications, financial ratings, spend history, delivery performance and external risk data produce incomplete outputs. 6 Require source-linked recommendations, duplicate resolution, normalised units and currencies, effective dates, contract and revision lineage, permission-aware retrieval and exportable audit logs. Keep hard blocks on autonomous supplier creation, award, signature and payment.

A controlled proof of value

Public pricing, implementation rates, usage limits and total cost are undisclosed. Budget for licensing, configuration, data cleansing, ERP and application-programming-interface integration, security review, supplier onboarding, training and ongoing exception review. Mercanis says it works alongside SAP, NetSuite, Slack and Teams and names Unite, IntegrityNext, Prewave and Coface integrations, but actual availability depends on subscription and partner contracts. 4

A useful pilot should compare the existing process with Mercanis on cycle time, reviewer hours, bid completeness, qualified-supplier accuracy, realised savings and exception rate. It should deliberately include stale prices, duplicate suppliers, missing certifications and changed drawing revisions. The public evidence supports a credible procurement platform and reported customer outcomes; it does not yet validate every industrial workflow or independent return on investment.

Sources and next reading

• Mercanis product site
• Mercanis customer case studies
• Amer Sports detailed case
• Oventrop detailed case
• Mercanis integrations
• Mercanis API explainer

Explore OfferSignal · Focused Business & AI Workshop · All research

References