4+ hours of campaign management cut to <2 minutes with AI Agent
Saves retailers tens of millions of euros annually
2-3% increase in net profit as a result of optimized pricing
Improves sustainability by reducing overstock and waste
Achieves industry-leading Net Promoter Score of 62
RELEX uses Google Cloud to help retailers optimize promotion efficiency, improving revenue and profit margin while reducing waste.
RELEX uses Google Cloud to help retailers optimize promotion efficiency, improving revenue and profit margin while reducing waste.
Overcoming the complexity of
modern retail planning
Google Cloud's early capabilities in data, ML, and Kubernetes gave us a massive head start and we're now fully leveraging their platform approach to AI. By building on their products, we're able to ship faster and scale our infrastructure without the usual growing pains.
Ludvig Storm
VP of Engineering, Pricing and Promotions, RELEX
As rising prices push shoppers to hunt for bargains, promotions now account for approximately 20% of all retail sales. Yet managing these campaigns is a complex and time-consuming task for retailers. Many juggle tens of thousands of weekly promotions, meaning they often simply repeat last year's plans, regardless of their effectiveness, rather than spending time planning new ones.
Poorly planned promotions do more than hurt the bottom line—they often lead to overstocking, which increases food waste, or out-of-stock scenarios that frustrate customers. RELEX Solutions helps retailers and manufacturers move beyond these manual limitations. Its pricing and promotions solution uses ML, large language models, and agentic AI to transform complex data into smarter strategies, allowing retailers to optimize every price point to meet customer demand while protecting their margins.
When building this solution, RELEX knew it needed a cloud provider that excelled in big data and ML to allow it to process such massive data volumes. The team chose to build on Google Cloud because of its advanced data capabilities.
"Google Cloud's early capabilities in data, ML, and Kubernetes gave us a massive head start and we're now fully leveraging their platform approach to AI," explains Ludvig Storm, VP of engineering for pricing and promotions at RELEX. "By building on their products, we're able to ship faster and scale our infrastructure without the usual growing pains."
Solving retailers' complex pricing challenges
with agentic AI
With Google Cloud, RELEX turns raw retail data into actionable insights. The company uses Cloud Storage to ingest vast amounts of customer data, including transaction histories, inventory levels, and promotional calendars. To process this data at scale, RELEX uses BigQuery as its central data platform to perform complex transformations on massive datasets with ease, alongside Dataflow to efficiently run specialized, large-scale processing jobs. Together, these tools provide the essential data foundation on which the rest of RELEX's solution is built.
Once the data is processed, it moves into the ML layer. The company recently migrated its forecasting models from a legacy, manually deployed environment to Vertex AI to take advantage of its MLOps capabilities. With Vertex AI, the RELEX team can manage the entire life cycle of its AI models, monitoring model performance, managing versioning, and ensuring its predictions remain accurate as market conditions change.
Promotional planning is a complex problem for retailers and a very challenging optimization problem for a service provider. With Gemini and Vertex AI underpinning our AI agents, we're solving that complexity and amplifying our recommendation and automation capabilities in a way we could never have imagined just 12 months ago.
Ludvig Storm
VP of Engineering, Pricing and Promotions, RELEX
This ML layer allows RELEX to provide advanced business intelligence that moves beyond simple sales tracking. By analyzing historical data, retailers can identify deep patterns in consumer behavior, such as how a discount on one item drives sales for complementary products, or how a promotion might inadvertently divert demand from other brands.
Now, RELEX is enabling retailers to turn these high-level insights into immediate action with an agentic layer built using Gemini models on Vertex AI, and orchestrated through Google Kubernetes Engine (GKE). This AI agent will allow retailers to interact with their data using natural language, effectively turning the platform into a pocket pricing consultant that can handle several critical tasks in a single conversation.
The AI agent will save retailers significant time during campaign planning by allowing users to specify objectives, rather than tactical execution details, to meet business goals, ensuring promotions meet targets without requiring hours of manual entry. A promotion campaign that previously took more than four hours to manage can now be executed in less than two minutes with RELEX's AI agent, built with Google Cloud.
The agent will also simplify onboarding. A new category manager will be able to quickly get up to speed with the company's vast amount of historical data, with the agent pointing to underperforming promotions or products that are over-promoted to help the new manager understand where they can have the biggest impact in their first 90 days.
By asking the agent to list promotions that are underfunded relative to the category average, retail buyers can also negotiate with suppliers more effectively to secure the funding needed to maximize incremental volume and profit.
"Promotional planning is a complex problem for retailers and a very challenging optimization problem for a service provider," Storm explains. "With Gemini and Vertex AI underpinning our AI agents, we're solving that complexity and amplifying our recommendation and automation capabilities in a way we could never have imagined just 12 months ago."
Unlocking future growth with
next-generation AI capabilities
Google Cloud's early strength in big data and ML allowed us to build the foundations to get where we are today. Now, with Vertex AI and Gemini, we're unlocking new potential. As Google Cloud continues to pioneer the AI space, we're excited to keep innovating to solve retailers' biggest challenges.
Ludvig Storm
VP of Engineering, Pricing and Promotions, RELEX
By providing a more user-friendly way to engage with complex data, RELEX empowers retailers to make better decisions, leading to savings of tens of millions of euros annually. Using RELEX to optimize pricing and target their promotions more effectively, retailers have seen a 20% increase in sales, along with a 2–3% increase in net profit.
RELEX's solution also drives significant improvements in sustainability, with more accurate volume forecasting helping retailers reduce overstocking and minimize waste. These clear benefits are reflected in the company's industry-leading Net Promoter Score of 62.
The company is now focused on making its AI agents even more robust and production-ready. While these agents currently act as assistants, RELEX sees potential in granting them increasing autonomy to tackle tasks independently, with the correct guardrails in place.
It also plans to list the new AI agent on the Google Cloud Marketplace to allow retailers to access and deploy RELEX solutions using their existing cloud commitments, helping the company expand its user base and grow the business.
"Google Cloud's early strength in big data and ML allowed us to build the foundations to get where we are today," adds Storm. "Now, with Vertex AI and Gemini, we're unlocking new potential. As Google Cloud continues to pioneer the AI space, we're excited to keep innovating to solve retailers' biggest challenges."
RELEX provides a unified, AI-native platform for retail and supply chain planning. It helps retailers, manufacturers, and wholesalers optimize demand, inventory, merchandising, pricing, and supply and production planning to improve availability and efficiency at scale.
Industry: Retail
Location: Sweden
Products: Google Cloud, BigQuery, Cloud Storage, Dataflow, Gemini, Google Kubernetes Engine (GKE), Vertex AI
