Smyths Toys

Customer Service Automation · Enterprise Knowledge Search

Smyths Toys needed to handle growing customer-service volume across email, web, and social channels while giving AI agents access to live order, shipping, refund, and product data without slowing or destabilizing core SAP transaction systems.

Read the original story Explore all use cases

Resolved more than 300,000 customer emails

handled approximately 60% of the entire customer service workload

maintaining customer satisfaction above 90%

Reported impact

Business outcome

Smyths expanded Codie from order tracking into a broad omnichannel service platform, resolved hundreds of thousands of emails, and automated most peak shopping-period support while keeping satisfaction above 90%.

Solution

Smyths Toys built an agentic customer-service platform called Codie on top of Gemini Enterprise and a split Google Cloud architecture that separates transactional SAP workloads from heavier AI inference. Business-critical processes remain in an ABAP implementation using the Google Cloud ABAP SDK, while more compute-intensive orchestration and reasoning run in a Python environment on Kubernetes using the Google Agent Development Kit. Incoming requests are first classified through Gemini Enterprise Agent Platform, then handed to a supervisor agent that can delegate retrieval or action tasks to specialized sub-agents. Gemini's native function-calling pulls live information from SAP S/4HANA, carrier APIs such as Royal Mail, DPD, and DHL, and payment providers including Stripe and Worldpay. BigQuery augments the SAP estate with additional datasets, allowing Smyths to serve customer requests in real time while preserving the performance of mission-critical retail systems.

Data flow

Customer emails, web chats, and social messages are classified on Gemini Enterprise Agent Platform and routed to a supervisor agent. The workflow calls live SAP, carrier, and payment systems through Gemini function-calling, enriches the response with BigQuery-backed context, and then returns or completes the action before sending the customer reply.

Solution architecture

5 components · 4 layers

Compute
- Gemini Enterprise Provides the underlying Gemini models used for customer-service reasoning and function-calling in Codie.
Storage
- BigQuery Combines SAP and additional enterprise datasets so agents can reason over broader retail context.
Serving
- Google Cloud ABAP SDK Connects business-critical SAP processes to Google Cloud while preserving performance for transactional retail workloads.
Orchestration
- Gemini Enterprise Agent Platform Classifies incoming requests and orchestrates the supervisor-and-sub-agent workflow for customer-service tasks.
- Google Agent Development Kit (ADK) Supports the Python-based agent environment that handles complex AI tasks on Kubernetes.

Architecture clues

BigQuery is being used alongside enterprise SAP data to support larger analytics workloads
Codie uses two complementary environments: ABAP for SAP-critical processes and Python on Kubernetes for heavier AI workloads
Incoming requests are analyzed before being assigned to a supervisor agent
Specialized sub-agents can execute function calls against operational systems

Evidence from the source

Codie consists of two complementary environments.

During peak shopping periods, the platform handled approximately 60% of the entire customer service workload, while maintaining customer satisfaction above 90%.

Each customer interaction begins with an intelligent classification stage.

Through Gemini's native function-calling capabilities, Codie securely retrieves live information from SAP S/4HANA, carrier APIs such as Royal Mail, DPD, and DHL, and payment platforms including Stripe and Worldpay.

incoming requests are analyzed before being assigned to a supervisor agent responsible for planning how the enquiry should be resolved.

Transparency Evidence & methodology 100/100 reference value · 79% confidence · 1 source

Review how this reference was scored and trace each normalized claim back to its documented source.

Scoring methodology

How scores are calculated

Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.

Customer reference value100/100

Problem clarity20/20
Measures whether the customer's operational problem is specific enough to compare.
Implementation depth25/25
Rewards named components, data flow, integrations, and architecture detail.
Outcome specificity20/20
Measures whether the reported benefit is concrete rather than a generic claim.
Quantified impact25/25
2 metrics describe measurable impact.
Deployment scale10/10
Rewards evidence of a live or scaled deployment and meaningful integration scope.

Evidence confidence79%

Source authority20/25
The main evidence is an official vendor-published customer story.
Claim traceability21/25
10 of 12 key fields link to a named source.
Direct evidence20/20
5 direct source quotes retained.
Date provenance0/10
The publication date has no documented provenance.
Source transparency8/10
1 source reference document publisher, URL, authority, retrieval, and supported fields.
Implementation specificity10/10
Rewards source-backed components, data flow, quantified impact, and a specific outcome.

Traceability

Sources & provenance

Field coverage shows which facts each source supports. Dates are shown only when documented by the source.

Google Cloud
Smyths Toys customer story

First-party
https://cloud.google.com/customers/smyths-toys-ge

Source type
Vendor-published customer story
Supports
- Customer
- Business problem
- Solution
- Products
- Solution components
- Data flow
- Integrations
- technical area
- use case category
- Business outcome
- outcome category
- Metrics
- Architecture
- evidence quotes

Date evidence

First archived Oct 8, 2026 · Internet Archive · Independent First Wayback Machine capture; confirms the page existed by this date, not when it was published.
Open date source
Sitemap last modified Sep 16, 2026 · Google Cloud · Publisher metadata Publisher-reported modification timestamp; not a publication date.
Open date source

Not separately sourced Industry, Region, Company size. Treat these normalized attributes as lower-confidence until a dedicated source is added.

Similar use cases

See all like this

Google Cloud Est. · Apr 13, 2026 FairPrice Group Retail Asia Pacific FairPrice Group needed to keep pace with rapidly shifting retail consumer needs across more than 570 physical and digital touchpoints, personalize the in-store and online shopping journey, and equip more than its… Ad creation cost reduced by up to 100x Customer Service AutomationEnterprise Knowledge Search Chirp 2 · Gemini API · Gemini Enterprise

Google Cloud Archived · Oct 2, 2026 Hawaii DCCA Public Sector Honolulu, Hawaii The Hawaii Department of Commerce and Consumer Affairs (DCCA) serves a vital mission: protecting residents while ensuring a level playing field that promotes fair business competition across the state. Reclaiming Time: Streamlined operations and reduced manual work across 52 state boards and commissions by… Customer Service AutomationEnterprise Knowledge Search Cloud Storage · Contact Center AI Platform · Customer Experience Agent Studio

Google Cloud Archived · Sep 17, 2026 Home Credit Bank Financial Services Kazakhstan Home Credit Bank needed a safer enterprise AI environment that non-technical teams could use to build banking agents quickly without creating data-leak risk or relying on a long custom-development cycle for each new use… Achieved 100% migration to a secure AI environment Customer Service AutomationEnterprise Knowledge Search Agent Designer · Gemini Enterprise