Rory O’Connor, founder and CEO of Scurri, explains how brands and retailers can cut the cost to serve at every stage of the chain
In fulfilment, the debate about whether AI will deliver value is already over. The real question is how quickly retailers can scale it across the chain to convert fragmented processes into a connected, self-improving system.
The marketing conversation around AI has focused heavily on discovery and personalisation. But every front end promise still has to be delivered by fulfilment teams under pressure to be faster, more accurate and more transparent. This pressure is intensified by the fact that consumers are already using AI extensively and have no intention of paying more for AI-enabled delivery. Scurri’s research shows that 44% of shoppers reject any surcharge for smarter fulfilment, even though most expect AI to improve delivery updates, order allocation and performance during peak.
Intelligent fulfilment systems
This environment demands that AI create efficiency rather than added cost. Intelligent Fulfilment Systems deliver exactly that. Instead of being confined to a single workflow, AI becomes the connective layer orchestrating decisions across inventory, warehouse processes, carrier selection, issue detection and returns.
Forecasting for end to end efficiency
Accurate forecasting is the foundation upon which the rest of fulfilment stands. When AI reduces forecast errors by 20-50%, it not only positions stock in the right location but also cuts waste and operational friction. Better forecasts lead to fewer lost sales, lower warehousing costs and significantly reduced administrative effort.
The impact is even more pronounced in social commerce, where platforms like TikTok Shop can create rapid, unpredictable surges in demand. With 72% of shoppers believing AI is essential to fixing fulfilment in these channels, real time forecasting allows retailers to mobilise inventory precisely when and where it is needed. In this sense, forecasting becomes the first in a chain of intelligent decisions that lower cost-to-serve long before the parcel enters a warehouse.
Warehouse automation to convert accuracy into productivity
Once the demand picture is clear, AI-enhanced warehouse automation converts that accuracy into throughput. Benchmarks show that AI improves picking and packing efficiency by 10-20%, while automation systems like DHL’s DHLBot deliver 40% greater sorting capacity with near-perfect accuracy. Unilever’s AI-enabled supply chain achieved similarly significant productivity gains alongside major emissions reductions.
These improvements translate into fewer mis picks, shorter processing times and more intelligent slotting. When forecasting feeds into warehouse logic and labour scheduling, the facility stays continuously calibrated to demand. Capacity therefore increases without expanding infrastructure or adding headcount, a crucial advantage in labour constrained environments.
Dynamic carrier allocation to cut cost to serve
The last mile remains the most complex and expensive stage of fulfilment, but AI gives operations leaders new tools to manage it. Dynamic carrier allocation evaluates every parcel in real time and selects the most cost-effective and reliable carrier for the conditions of that moment.
At scale, this reduces delivery delays by up to 50% and improves fleet utilisation through far more accurate volume prediction. Even simple routing optimisations have proven substantial; UPS famously saved millions of gallons of fuel annually by reconfiguring routes. With consumers expecting real time tracking and proactive updates, intelligent carrier allocation is now essential to both cost control and customer satisfaction.
Anomaly prediction to fix problems before they escalate
Traditional fulfilment detects issues only after customers report them. Intelligent anomaly detection reverses this model. AI identifies delays, stalled parcels, mismatched delivery promises, system failures or weather risks before they create downstream problems.
Although many consumers still prefer human support for complex issues, a large proportion welcome AI resolving simple problems instantly. Proactive issue detection therefore reduces inbound contact volumes, prevents unnecessary refunds and improves overall customer experience. It also frees operations teams from constant firefighting, allowing them to focus on system-wide improvement.
Returns optimisation to protect margin
Returns remain one of the highest-cost components of fulfilment, but AI allows retailers to treat them as an optimisable workflow rather than a fixed cost. Predictive models can anticipate return volumes, identify common causes and determine the most efficient route for each item, whether that is back to a warehouse, to a store or directly to a resale or refurbishment partner.
AI also enables instant low-risk refunds, consolidates multiple returns to reduce cost and emissions, and prioritises circular pathways that maximise recovery value. With many shoppers saying AI-managed returns make them more likely to repurchase, optimisation at this stage directly protects loyalty and long-term margin.
The power of intelligent fulfilment systems
When forecasting accuracy informs warehouse slotting, which shapes carrier allocation, which feeds into anomaly detection and returns orchestration, each gain amplifies the next. The result is an ecosystem that reduces cost per parcel, increases throughput without expanding facilities, lowers emissions and strengthens resilience during demand surges, particularly those driven by social commerce.
