How we solved Pelangi's manual sales processing with AI

An educational publisher put its purchase orders through Axrail. Textract and Claude on Bedrock read and structure each order, so admins review only what needs a human eye.

40%

less time spent processing each purchase order

25%

lower manpower requirement over three years

AWS Premier Partner with GenAI expertise.

See how Pelangi transformed their order processing with Axrail

Watch how Pelangi's admin team replaced manual purchase-order entry with an AI-reviewed workflow, cutting time spent processing each order by 40%.

About Pelangi Publishing Group

Pelangi Publishing Group is a regional educational publisher founded in 1979 by a group of young teachers. Headquartered in Malaysia with publishing offices in Thailand and Indonesia, Pelangi produces educational materials for the Malaysian, Thai, Indonesian and Singaporean curricula, spanning every level from kindergarten to pre-university.

Every year Pelangi produces more than 400 new publications in five languages: English, Bahasa Malaysia, Chinese, Thai and Bahasa Indonesia. Behind that catalogue sits a large and growing inventory, wide SKU variety, and a customer base that runs from big urban schools to small rural ones.

Document Intelligence
Publishing
Malaysia
AWS
Amazon Textract
Amazon Bedrock

The challenges

Conventional sales channels
Pelangi's sales channels cover big, small, urban and rural schools. Many of its distributors still operate in conventional ways that involve large amounts of manual work.
Slow processing time
In the months leading up to a new school season, Pelangi is under pressure to deliver high volumes of book orders to schools over a short order-processing timeline.
Large, complex inventory
Across 40 years and more than 400 new publications a year, Pelangi's huge inventory and its SKU variety across subjects, levels, series and versions complicate how the team manages stock.

The solution

Customers’ purchase orders arrive in many different formats, so Pelangi’s team had to extract the details from each one by hand and key them into the business portal. Axrail replaced that reading with a generative AI solution built on Amazon Textract and Amazon Bedrock, taking the manual keying off both the operations team and the sales team.

How the solution works#

On the customer’s end, sales entry documents are scanned.
The scanner triggers an email notification that starts the workflow.
Amazon Textract extracts the key information on each document: Customer PO, Customer ID, Customer Name, ISBN or Bookcode, Quantity and Sample.
The extracted text is passed to Amazon Bedrock running Anthropic’s Claude, which structures it into key-value pairs.
The order database is populated automatically with the processed information.
Pelangi’s admins review and verify the extracted information on their web portal before the order moves on.

For most organisations, daily sales activity still looks like the diagram below: customers issue a purchase order, and the operations team manually inputs it into ERP software such as SAP or Navision. The process is time-consuming and prone to errors.

Workflow diagram of manual sales processing: customers send a purchase order to the operations team, which manually inputs it into the ERP or accounting system. — Before: every purchase order is keyed into the ERP by hand.

With Axrail Commerce, the customer journey stays unchanged. The purchase order goes into the Axrail Commerce platform, where Amazon Textract and Amazon Bedrock running Anthropic’s Claude extract the relevant information and automate the data input into the ERP.

Workflow diagram of AI sales operations using Axrail Commerce: purchase orders from any channel flow into the Axrail Commerce platform on an AWS data lake, which automates input into the ERP or accounting system. — After: Axrail Commerce reads each order and feeds the ERP automatically.

Conclusion

Axrail's solution solved Pelangi's manual sales processing challenges with AI, cutting order processing time and the manual workload while lifting customer satisfaction and team efficiency.

We believe education is an industry that needs digitalisation so future generations benefit from it. Pelangi trusted us with the problem, and the system now does the routine reading so its people can stay with customers.

40%less time spent processing each purchase order