Customer:

Globally known European structural clay products manufacturer. It has more than 55 years of successful operations all over Asia.

Business Challenges:

Although the company is digitally transformed, the management did not know what to with the data. The major problem was the error-prone and dysfunctional systems and its disintegrated channels to its customers; and the massive gap of talents in Artificial Learning (AI) and Machine Learning (ML)

Solutions:

A thorough workshop was conducted to first understand its current data readiness and maturity. Immediately, a semi-dedicated AI and ML expert were attached with the company for two months. The technology deployment mainly revolved around the Hyper automation (RPA + AI) solution. This includes:

1. Intelligent Document Processing (IDP): Configured Microsoft Azure AI to extract and match data on Point of Sales (POS)
2. Optical Character Recognition (OCR): An enterprise cloud-grade OCR engine to extract necessary entities from both printed and handwritten text in documents.
3. Robotic Process Automation: Designed end-to-end automation flows (cloud + desktop) with error-handling logic, automating their ERP and CRM applications to streamline business processes and improve data consistency.
4. API Integration: for real time master data validation.
5. Data Science: purposed-built custom vision model to match tiles by color and texture.
6. Natural language processing (NLP): algorithm model to analyse product descriptions and match them with the product database for accurate product code retrieval.
7. Support Services: allocated 40 man-hours per month for managing the AI/ML model via shared services.

Outcome:

1. Processing a PO now is less than 5 minutes; previously at 2 days
2. Accuracies on OCR and AI ML model improved by 86%, and 92.7%, respectively.
3. Returning customers improved by 13%; while NPS score +3.5 within a quarter.
4. Customer based improved by 17.4% in a quarter.

That smart recommendation engine greased the entire ecosystem by simplifying the process between customer and inventory. Product searchers (on website) once took 2 days for them to revert and if the matching doesn’t fit, it will require another 2 days. Now, its takes only 5 minutes as the platform is interactive and AI enabled. This resulted in more successful transactions and returning customers.

Quote by customer:

“Since implementing IDP platform, we've transformed how we handle customer inquiry and PO process. The automation and accuracy reduced processing times and errors, which had a direct impact on customer satisfaction.” - Regional (Asia) Vice President, Marketing and Customer Experience.