Case Study

Optimising quotation management with Agentic AI on Oracle Cloud

Technology Reply has supported Pusterla 1880 in adopting a multi-agent system to optimise the quotation creation process and refine pricing strategies, ensuring even more precise and customer-focused service.

Picture

THE CHALLENGE

To make the quotation process faster and more efficient, ensuring quicker and more accurate responses to customers.

SCENARIO

A new approach to managing luxury packaging quotations

Pusterla 1880, a company specialising in the production of refined packaging for over 140 years, serves global luxury brands. The growing demand for bespoke solutions and the complexity of production processes necessitated a more efficient way of managing cost estimates and operational workflows. The traditional quotation process relied on manual analyses by skilled personnel, consuming significant time and resources. To enhance the management of commercial proposals and ensure greater accuracy in cost estimations, Pusterla adopted an AI-driven system. This new approach intelligently integrates data management, optimises the quotation process, and supports strategic decision-making.

Picture

THE SOLUTION

Automation and multi-agent AI for cost estimation management

The solution implemented by Technology Reply, based on Oracle Cloud Infrastructure (OCI), integrates advanced technologies such as Autonomous Database (ADW), Oracle Database 23c AI, and Oracle AI Vector Search, ensuring seamless integration with Pusterla’s Oracle Cloud ERP to optimise cost estimation management and pricing strategies. The AI architecture is based on a multi-agent model, where each agent is specialised in a specific phase of the quotation management process, enabling complete automation.

The system can analyse customer briefs expressed in natural language, extract and structure technical information, generate detailed cost estimates, and formulate price proposals based on historical data and production specifics.

The solution consists of three main components, each optimised for a specific area of the process

AI Configurator
Assistant

A machine learning system that analyses historical data to provide more accurate cost estimates and optimise quotation management, ensuring real-time monitoring.

Notes Analysis
Assistant

A generative AI model that automates the extraction of technical specifications from customer briefs, improving data consistency and reliability through the use of LangChain Agent and Tools.

Estimate
Assistant

An AI module based on Oracle AI Vector Search and ADW that enables the retrieval of historical data and supports pricing decisions and commercial strategies.

HOW WE DID IT

Integration of the multi-agent solution with Oracle Cloud ERP

Technology Reply adopted a gradual approach to ensure effective integration of the multi-agent solution within Pusterla’s environment. The first phase involved analysing historical data, collecting and structuring information related to production, pricing, and quotation to train the AI agents with representative and reliable data. Subsequently, the AI ecosystem was developed and tested, ensuring compatibility with the existing IT infrastructure, particularly the Oracle Cloud ERP and corporate database. After a validation and optimisation phase, during which the system’s predictive capabilities were refined to guarantee accurate estimates and data consistency, the go-live phase was launched. The solution was implemented with a continuous monitoring system, allowing performance analysis and progressive improvements, ensuring efficient functioning and constant optimisation of pricing and quotation strategies.

THE BENEFITS

Better process management and increased operational efficiency

The adoption of the Agentic AI system integrated with Oracle Cloud allowed Pusterla to achieve tangible benefits, including:

Picture

Automation of
the quotation process

Reducing manual workload and improving accuracy.

Picture

A reduction of up to 90% in the time required to generate cost estimates

Thanks to integration with historical data and predictive analysis.

Picture

Optimisation of
pricing strategies

Increasing competitiveness and economic sustainability.

Picture

More efficient management of production workflows

With better resource and quality control.

Picture

Improved
data consistency

Ensuring uniform and reliable technical specifications.

Picture

Faster responses
to customers

With quotations based on real-time data analysis.