How AI Can Help Planners Avoid the ‘Guessing Game’ and Respond to Real-Time Performance

Fashion’s margins are under pressure. Companies are spending more to produce and ship merchandise while also navigating a promotion-heavy retail environment. The risks of getting inventory assortments wrong are rising, and margins can quickly erode if merchandise does not move.

Traditionally, planning for fashion seasons has been a purely predictive activity. Teams would forecast months or years in advance and create assortments in response to where it seemed demand was headed. During the recent Sourcing Journal “In-Season Agility: Using AI to Make Faster, Smarter Fashion Decisions” webinar, moderated by Sourcing Journal logistics editor Glenn Taylor, speakers explained why the typical ways of working no longer cut it.

“Planners are still being asked to deliver certainty in an environment that no longer behaves predictably,” said Liza Amlani, chief merchant at Retail Strategy Group. “So the old model built around making a few big bets far in advance—sometimes a year or longer—that is exactly where volatility creates that risk. Planners are managing things like late receipts, shifting demand, regional differences, channel differences and more pressure to protect margin, all at the same time. And the job is no longer just forecasting, it’s constant signal reading and decision-making.”

Compounding these challenges is the continued reliance on outdated technologies, with spreadsheets and analog tracking still prevalent. Operational siloes between design, merchandising, supply chain and finance teams translate into disconnected data if there are not technology systems in place to share information.

As Ken Weygand, solutions architect at Aptean, explained, fashion has several unique challenges for planning, including seasonal shifts and speedy trend cycles fueled by social media. Adding complexity, the industry also manages copious SKUs since each size and color must be tracked separately. Also making it harder to read the demand forecast are the emotional motivations behind purchases.

Alain Tessier, director of product management at Aptean, noted that pre-season planning still has value, but companies must anticipate the need for adjustments. Even the most data-driven predictions can be wrong when the actual season hits if there are consumer behavior shifts, new trend spikes or other changes. Therefore, the goal is not “perfect forecasts,” but being able to pivot quickly. “The real danger of locking into a pre-season plan is that by the time you realize it’s wrong, it’s often too late to fix it without taking a significant financial hit. You’re either sitting on inventory nobody wants, or scrambling to find product you can’t get fast enough. Both situations are pretty expensive.”

Typically, the tactic for dealing with excess or low-performing inventory has been markdowns or shifting goods to off-price channels. But as Tessier noted, relying on discounting has disadvantages including training consumers to wait for sales. However, brands that can anticipate or identify slow-moving merchandise sooner have more options at their disposal. For instance, Amlani noted merchants can shift where goods are placed to better align with demand, such as changing channels, locations, customer bases or timing to improve sales.

With traditional reporting cycles, companies may not catch issues for multiple weeks or longer. Artificial intelligence has the ability to constantly keep tabs on product performance, and it can automatically alert teams to margin risks in real time. By leveraging a combination of connected data management and AI, companies can get an earlier read on inventory issues, including any production delays that may impact stock. Reducing stockouts, AI can also identify where merchandise is performing well and help inform replenishment plans or indicate what alternatives might work as a substitution for unavailable styles.

Rather than replacing humans in the planning process or decision-making, speakers noted that AI is a tool to support and surface insights that would otherwise not be achievable via manual oversight. Human judgment then comes in to review and validate what the AI is communicating. Amlani pointed out that planners bring a perspective to the table, considering their brand’s DNA and customer profile, but layering in AI can help prevent planning from turning into a “guessing game.” Weygand noted, “Successful AI initiatives focus on transforming roles, not eliminating them.”

Aly Breeman, senior product manager at Aptean, cautioned that the quality of AI-powered analysis is only as good as the associated data management. If data is missing or wrong, or if systems are disconnected, AI will only exacerbate these issues. “Companies need to focus on data quality, governance and consistence before expecting AI to drive value,” she said. “In other words, the more aligned your systems are to the industry and the more disciplined your data management is, the more efficient AI will be.”

Watch the webinar on demand to learn more about:

Watch the webinar here.