New Way Now: Smyths Toys delivers agentic customer support to meet peak Christmas demand

Learn more about other Google Cloud customers bringing their ideas into reality by exploring our Customer Success Library - https://goo.gle/3VJUjvS

Featured in this video: Adam Sexton, AI Development Manager, and Deepika Ramachandran, AI Application Developer at Smyths Toys

Executive summary: Global specialty toy retailer Smyths Toys generates 70% of its annual revenue during the Q4 Christmas season. This causes a massive influx of more than 1k customer emails per day and severe backlogs entering the new year. To scale customer support, the retailer worked with Google Cloud to build Codie, an agentic AI assistant powered by the Gemini Enterprise Agent Platform and Gemini Pro. Natively integrated with core SAP databases and scaled on Google Kubernetes Engine (GKE), Codie operates across web, email, and social channels. Today, Codie Web autonomously resolves 60% of customer queries, while Codie Email handles 90% of inbound email. By also integrating shipping data for real-time tracking, Smyths Toys has freed human agents to focus on complex inquiries, enabling the team to smoothly manage the end-of-year rush and enter the new year at inbox zero.

Challenge: Smyths Toys is one of the largest specialty toy retailers in the world, with an iconic multinational brand presence across Ireland and the UK. For a major brick-and-mortar and digital toy retailer, business operations are heavily seasonal; the company generates approximately 70% of its annual revenue in just the last three months of the year leading up to Christmas. During peak season, customer inquiry volumes surge exponentially. The company's customer support queue routinely faced 1k+ inbound emails every single day. Support teams are overwhelmed, leaving the business facing a massive backlog of unanswered messages entering the new year.

Solution: Smyths Toys chose Google Cloud for both its AI models and for having hands-on engineering collaborators, workshops, and advisory support to guide the team through their agent journey. The company built Codie — an agentic AI support solution powered by Gemini Enterprise Agent Platform and Gemini Pro. Using the Google ABAP SDK, they natively connected their core SAP database to Gemini to launch agents across web, email, and social media. Built on Google Kubernetes Engine (GKE), the platform automatically scales to handle intense holiday traffic spikes.

Result: With Codie Web, Codie Email, and agentic social media support, Smyths Toys' team can handle a wider range of customer service use cases faster than ever. On the website, Codie Web autonomously resolves 60% of all customer queries, intercepting 600+ of the more than 1k daily inbound questions. The remaining 400 emails are directed to Codie Email which replies to 90% of inbound email queries. Further down the funnel, human agents now have more focused time to resolve a daily average of 50 complex customer requests. Additionally, by integrating its agents with shipping partners’ data, Smyths Toys can now provide real-time tracking data to deliver customers more information about their important gift purchases. With these integrations, Smyths Toys can ensure all customers get the support they need and the team can enter the New Year at inbox zero.

Highlights and key takeaways from our interview with Adam Sexton, AI Development Manager and Deepika Ramachandran, AI Application Developer:

→ "At peak, we receive about a thousand emails a day. We’d typically face a massive backlog of those emails that would roll over into the new year. With Codie, at the start of 2026, we had zero." — Adam Sexton
→ “The underlying technology is important for us, but the real value-add from Google is the people we interact with and the support we get. It feels as if Google is responsible for Codie in many ways because they really helped us get it there.” — Adam Sexton
→ “We’ve federated some of the relevant data we require for customer service into BigQuery, and our agents use data from BigQuery to respond to customers in real time.” — Deepika Ramachandran

Google Cloud products used: BigQuery, Gemini Enterprise Agent Platform, Google ABAP SDK, Google Kubernetes Engine