Normalized claim
Call center volume: 50-60% decrease
call center volume decreased by 50–60 percent
PınarOnline and LimonCloud: Generative AI customer-support assistant on Amazon Bedrock
Use case typeCustomer service agentUpdated Jun 12, 2026
PınarOnline, the ecommerce platform for a leading Turkish food brand, worked with AWS Partner LimonCloud to build a generative AI assistant on AWS for customer support and sales capture. The assistant handles inquiries across call center, email, and WhatsApp, surfaces products of interest, and alerts the sales team when bulk-purchase intent is detected. The solution used Amazon Bedrock, Amazon Bedrock Guardrails, Amazon ECS, Amazon RDS, AWS Lambda, and Amazon SNS.
- Organization
- PınarOnline
- Industry
- Consumer & Food
- Location
- Turkey
- Published
- June 2026
Reported outcomes
7,000 users
monthly unique usersAdoption & scale
50-60%call center volume20-25%customer satisfaction18-21 interactionsmonthly interactions25 purchasesbulk purchases captured
Strategic outcomes
New business modelCaptured after-hours sales opportunitiesBetter decisions & insightImproved visibility into customer behavior and recurring issuesOther strategic outcomeAdded product discovery and sales-routing capabilities
Why do we believe this?Outcome claims, sources, and evidence checks
Normalized claim
Customer satisfaction: 20-25% increase
customer satisfaction rose by 20–25 percent
Normalized claim
Monthly unique users: 7,000 users increase
PoBo now supports roughly 7,000 unique users each month
Normalized claim
Monthly interactions: 18-21 interactions increase
generating 18,000–21,000 interactions
Normalized claim
Bulk purchases captured: 25 purchases increase
helped PınarOnline to obtain 25 bulk purchases
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- PınarOnline
- Provider
- AWS
- Maturity
- Unknown
- Linked source
- AWS Customer Story
No explicit deployment-stage evidence found.
Customer identity supportedSource describes one deploymentMaturity evidence evaluated
Primary read
Use case focus
Showing 2 of 2
- 1Customer service agent
- 2Workflow automation
- Customer inquiries were fragmented across call center, email, and WhatsApp, creating slow responses outside business hours.
- Missed bulk-order inquiries were causing lost revenue and the company lacked visibility into recurring customer questions.
- Built PınarOnline Bot (PoBo), a production generative AI assistant on Amazon Bedrock.
- Used Amazon Bedrock Guardrails for safeguards, Amazon ECS microservices for support functions, Amazon RDS for conversation logs and sentiment signals, AWS Lambda for asynchronous workflows, and Amazon SNS to alert sales when bulk interest appears.
- Customized the assistant to PınarOnline’s ecommerce workflows rather than deploying a generic assistant.
- Cut call center volume by 50 to 60 percent.
- Increased customer satisfaction by 20 to 25 percent.
- Supports roughly 7,000 unique users per month.
- Generates 18,000 to 21,000 interactions per month.
- Helped capture 25 bulk purchases in the first month by catching after-hours inquiries.
Architecture
PoBo is a generative AI assistant built on Amazon Bedrock. Amazon Bedrock Guardrails provide policy safeguards, Amazon ECS hosts microservices that support near real-time communication and logging, Amazon RDS stores conversation data, sentiment signals, and errors, AWS Lambda handles asynchronous workflows, and Amazon SNS sends sales alerts when customers express interest in bulk purchases. The system includes product discovery, lead capture, and reporting capabilities.
Implementation partners1
Sources & evidence1
Evidence: Medium50/100Evidence strength
- Customer explicitly identified
- Primary source available
- Quantified outcome available
- Technical implementation details available
Type: Customer StoryPublished: Jun 12, 2026Publisher: AWSEvidence: PrimaryConfidence: High
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Customer evidence
Provider evidence
AI patterns
Browse the catalog
Was this useful?
Community
Comments
No published comments yet.