Max Bruni in Supply Chain Brain: Beyond the Handshake: AI and End-to-End Management Can Protect the PO Process

Read the original article in Supply Chain Brain here.

Many times, several months before a retailer fills its shelves with a supplier’s products, the deal begins with a handshake. Two executives agree on an order amount, delivery date, and pricing, and then find themselves continually adjusting that deal for weeks until an order gets submitted and products hit stores.

For years, this has been the traditional way to kick off the purchase order (PO) management process — technically a handshake or verbal agreement leading to an email, EDI or fax to transmit orders — but it creates blind spots in the PO journey.

Typically, retailers use separate systems to manage the initial PO handshake deal, and then transition to an ERP for a final PO. Alternatively, the retailer manages the PO via manual spreadsheets, and later adds finalized details to an ERP. Either way, information can get lost, or changes to the PO can occur unseen by the rest of the supply chain.

Between the time of the “handshake” and when an actual PO gets processed by a vendor, a lot can happen. There can be a spike in material costs, say, or a sudden change in tariffs. These macroeconomic factors can have a big impact on the size of an order and pricing. In fact, even the slightest change to the PO can have a ripple effect throughout the supply chain, forcing business teams to adjust the allocation of product to certain stores, the financial plan, the assortment plan and the planogram.

A traditional PO process, where changes happen in a silo, can cost retailers significant revenue, with millions of dollars in financial commitments going unmanaged and untracked. However, with AI and analytics embedded into a unified platform, PO management can become more accurate, enabling retailers to get the right number of products at the right price and keep shelves replenished.

AI-Powered PO Management for Fast Fashion 

Few retail categories witness as many fluctuations in the assortment plan and product orders as fast fashion, making it ripe for an optimized PO management strategy. Department stores or fashion retailers with hundreds of stores around the world must keep core products in stock, while preparing for seasonal launches and assortment changeovers, as well as reacting to sudden changes in fashion trends.

Different brand managers handle men’s and women’s wear, children’s clothing, footwear, accessories, and so on. Each of them manages purchase orders, assortment plans and financial plans, and any change alters a store’s outlook. If each of these department managers is manually updating POs on siloed spreadsheets, they can miss pivotal information impacting the business overall. A drop in orders in men’s suits, for example, might open up the need for an order somewhere else. Merchants, planners and financial teams all need insights in real time to enable more agile decision making, proactively adjusting for margin, inventory flow and customer demand.

AI and streamlined analytics, coordinated in one place, can greatly improve the accuracy of PO management, right from the beginning of the process. If an executive at a sweater vendor shakes the hand of a buyer to produce a number and range of products, a streamlined, AI-powered PO system can immediately begin to track that order and manage expectations. When changes inevitably occur, AI can automatically calculate and report both the financial and inventory impacts happening at different selling locations.

Modern PO management requires an interconnected system where all business users can see changes to orders, impacting the supply chain down the line.

Retailers can improve total visibility across supply chain and can boost operations by conducting what-if scenarios, such as how an expected tariff increase might impact product orders and a product’s price. They can also deploy AI to automatically generate alerts about a vendor, such as if a sweater manufacturer routinely misses its delivery date, which can help managers expect and plan for upcoming shifts.

Further, deploying agentic AI can help automate decisions on replenishment orders and allocation; for example, an alert can note that there are multiple vendors supplying sweaters in one month and adjust orders for another month. And suppliers get onboarded to a retailer’s product information and supply chain system faster.

By moving PO management into a single, end-to-end system, retailers can eliminate issues with late entries, reduce expediting costs, and improve overall on-time delivery.

Preparing for AI-Powered PO Management

For retailers to develop a PO process backed by AI, there are some key steps to take. 

First, companies need to review the technical infrastructure of their supply chain and operations. If they’re using older, stand-alone technology systems, it will be tougher to deploy AI that monitors the entire journey of the PO. Also, it’s essential to have a centralized data management strategy. The heart of a supply chain is clean and enriched data, flowing from PO to delivery.

AI-driven automation can then power every step of the PO process, from order initiation and vendor communication to approvals and status updates. AI analyzes historical patterns of a vendor, like whether it delivers on time, optimizes the timing of reorders, and helps avoid overstocks or out-of-stocks.

The PO is an undervalued part of the supply chain. AI technology means retailers can now accurately monitor orders from handshake to point of sale.

Maximiliano Bruni is SVP, enterprise solutions architect, at Digital Wave Technology.