How Arena cut inbound supply chain disruptions by 78%

Arena is an Italian mid-market sportswear manufacturer running production on quarterly collection cycles. It concentrates inbound volume into short seasonal windows across a global supplier base, with most operating on a three-month factory lead time. Day-to-day supplier confirmations and packing lists are handled via email and Excel. The result is a high-volume, time-boxed inbound operation in which a single late or unconfirmed order can affect a launch.

Arena at a glance

• Industry: Premium competitive sportswear manufacturer
• Factories: 43, across 15 countries
• Markets: 128 countries
• Supply Chain Workforce: 33,139
• Revenue: €180M

How it ran before

Before working with Glacis, Arena ran inbound purchase order management on email, spreadsheets, and the attention of a small team. Every confirmation, change, and malformed file passed through a person before it became usable data.

Arena's team confirmed, tracked, and reconciled PO lines one by one, checking each manually against the volume and delivery date the supplier had agreed to.

How a purchase order moved through the team

The planner created orders and issued a confirmation request to each supplier. Suppliers were asked to confirm within two weeks. Sometimes the reply came back as the standard Excel template, but mainly as plain-text email, a PDF, or a modified spreadsheet.

The team read each reply, interpreted it, copied the confirmed dates and quantities into the ERP, and then updated internal trackers. When a supplier went quiet, operators sent reminders: a first reminder at roughly the one-week mark, and an escalation as the two-week mark approached.

Across the roughly three-month production cycle, the team ran production check-ins manually every two to three weeks to maintain visibility into whether goods would be ready as promised. As goods approached readiness, operators managed pre-shipping coordination as well. They issued booking reminders around 15 days before the shipping date.

Solution: Meeting suppliers where they already are

Arena needed a solution that could augment the way Arena and its suppliers already worked. Suppliers were never going to log into another portal, and Arena was not going to replace its ERP, which already held its orders. Consequently, we integrated our PO Lifecycle Agent into its shared inbox by adding the agent's email, and the agent took over coordination from there:

• Works with what suppliers already send - email, spreadsheets, and PDFs, turned into consistent records.
• Runs Arena's own SOPs and rules on every record, surfacing critical deviations automatically.
• Delivers a short, ranked list of deviations, not thousands of PO line updates to check by hand.
• Handles routine PO activities end to end, from confirmation to document validation, with a human in the loop.
• Manages by exception - the team sees only the deviations that threaten to become disruptions.

Business Impact: What changed for Arena

The clearest change was in reliability. With at-risk orders surfaced early and ranked by criticality, Arena's team resolved most routine issues before they became disruptions. Supplier OTIF rose from 80% to 86%, and inbound supply chain disruptions fell by 78%.

The second change was in capacity. Automating confirmation and validation work saved the equivalent of about five FTEs.

Document quality improved as well. Standardizing validation raised the quality of what entered the system and reduced manual errors at the point of entry. Fewer errors upstream meant fewer disputes downstream, and less administrative time spent reconciling them.

With better inbound visibility and proactive follow-ups, the team absorbed fewer surprises and spent less time firefighting. For a business built on seasonal collections, keeping inbound flow reliable protects speed-to-market and the profit margins that depend on landing a collection on schedule.

Where Arena goes next

PO confirmation was the first workflow Arena automated. With the business case proven, Arena is scaling AI agent implementation to other workflows with the same approach to document validation and delivery follow-up.

The takeaway is the one we hear from every team that starts this way. Pick one workflow that is painful, measurable, and bounded. Keep a person on approvals while trust builds. Then let the agent take on the next workflow as the evidence comes in.