We’re partnering with some of the most innovative retail and e-commerce brands to elevate their use of data and AI—enabling seamless omnichannel experiences that are efficient, sustainable, and drive profitable growth.

We help our clients unlock the full potential of AI across the entire value chain—from procurement to shelf. In today’s retail landscape, leveraging data and AI must go beyond automating individual tasks. Modern data platforms must enable accurate demand forecasting and supply chain optimization to meet the growing number of customer touchpoints, power immersive experiences, and significantly reduce waste.

Turning your data into a powerful, strategic advantage

The retail and e-commerce landscape is evolving at unprecedented speed. Omnichannel strategies have multiplied customer touchpoints and expanded fulfilment options, while the influence of generative AI continues to grow. Retailers now face mounting pressure to anticipate customer behavior, manage demand volatility, and respond swiftly to churn risks.

By embedding robust data and AI platforms at the heart of your business, you can eliminate data silos and unlock end-to-end AI value streams. This includes algorithms that uncover hidden customer behavior patterns, automations that support diverse fulfilment models, and predictive tools that enable dynamic pricing while minimizing waste. From digital assistants that enhance product discovery to AI-powered demand forecasting that optimizes logistics and reduces warehouse costs, a holistic approach to AI adoption empowers retailers to become more agile, efficient, and sustainable — gaining a significant competitive edge in the process.

Optimizing Call Center Forecasting and Performance

Booking.com is one of the world’s largest online travel agencies, processing more than a million reservations daily across 2.7 million properties in over 220 countries and available in more than 40 languages.

Challenge

Booking.com’s Customer Service (CS) centers employ thousands of agents who manage more than a million phone and email interactions every day, across 43 languages. Scheduling a workforce of this scale and complexity has long been a challenge. With rapid expansion into new markets, forecast accuracy for CS workload—a critical metric based on a 13-week daily horizon—began to decline, in some markets by more than 25%. The objective was to cut forecast errors in half with a system that agents and managers could easily understand and adopt.

Solution

We developed a prototype that could seamlessly scale from proof of concept and tested it in two language markets—one European and one Asian. Our approach was guided by four key principles:

• Understand: Capture the business dynamics that shape demand patterns.
• Simplify: Combine data intelligently and focus on the most impactful drivers to reduce complexity.
• Validate: Test multiple techniques rigorously and manage outliers through identification and analysis.
• Improve: Build a system with a learning loop to continuously correct forecast deviations.

Outcome

The total workload forecast error dropped by 60% in the pilot European language and by 50% in the pilot Asian language. This improvement was largely driven by a 75% increase in Contact Volume Forecast accuracy, supported by a 60% gain in Handling Time Forecast accuracy. When rolled out across the remaining 41 Customer Service languages, all but two showed measurable improvement. These two outliers will serve as the initial focus of our test-and-learn approach.