March 6, 2025
In today’s fast-paced business environment, companies are constantly striving to modernize their processes to stay ahead of the competition. The integration of Artificial Intelligence (AI) and Machine Learning (ML) has been a game-changer, especially in industries where efficiency and accuracy are paramount. The Project Elevate Formulation AI initiative is one such transformation that is revolutionizing the product formulation process within organizations. By replacing the traditional manual methods with an AI-powered system, the project has significantly improved both operational efficiency and cost-effectiveness.
In this blog, we will dive into how the Project Elevate Formulation AI initiative was developed, the challenges faced, the cutting-edge solutions implemented, and the long-term business impact. We’ll explore how AI and ML technologies helped automate the formulation process, optimized costs, and streamlined the entire workflow. We’ll also touch on how Royal Cyber, with its expertise in AI-driven transformations, played a key role in delivering this innovative solution..
The traditional formulation process within Organization’s Product Technology department was time-consuming, lacked efficiency, and required significant manual effort. There was a need for an AI-driven solution to automate the formulation process, reduce human error, improve accuracy, and optimize costs. The absence of a structured digital system hindered seamless formulation generation and cost evaluation, impacting productivity.
Formulation AI project involves finding the optimal combination and quantity of additives (from a list of over 43) to achieve a specific target strength in cement formulation. This optimization must be done within certain constraints, such as cost and additive availability. The challenge lies in predicting how different combinations and quantities of additives affect the final strength of the cement, considering:
- The current chemical composition of the cement.
- The current strength and the desired target strength.
- Cost constraints and additive availability.
Objectives
- Develop a user-friendly web interface.
- Implement AI and ML models for formulation predictions.
- Ensure robust user authentication and data security.
- Streamline communication with the ML model.
- Provide detailed audit logs and conduct security checks.
Planning Stage
To address these objectives, we formulated a structured project plan that involved:
- Identifying key pain points in the formulation process.
- Designing a roadmap for an AI and ML-based solution.
- Establishing clear project objectives and deliverables.
- Engaging cross-functional teams for requirement gathering and feasibility analysis.
- Implementing a phased development approach with regular milestones.
- Ensuring rigorous testing and security protocols.
- Security and Compliance Measures, Ensuring data integrity and role-based access control.
The Implementation Stage
The implementation process included the following key steps:
Creating a user-friendly web-based platform built using React.js and the Vite framework, with Tailwind CSS for styling. It is structured to support a modular and scalable architecture.
Key Features:
- Components & Shared UI: Includes reusable UI components like CountryBox, Language, Notification, and Selectbox, along with shared layout elements (Header, Footer, Sidebar).
- Pages & Routing: Implements authentication (Login, ResetPassword, VerifyCode) and core functionalities like Dashboard, CreateFormulationAi, and EditFormulationAi.
- Data Visualization: Uses charts (BarCharts, LineCharts, PieCharts, ScatterCharts) for analytics.
- State Management: Centralized using Redux slices (bookSlice, createExperiments, uniqueSlice).
- API Integration: Managed via a dedicated services/api.js file.
- Project Configuration: Uses .env files for different environments and ESLint for code quality.
The project utilized advanced machine learning algorithms to analyze and optimize formulation constraints, including chemical composition, target strength, cost limitations, and additive availability. The goal was to develop ML models capable of generating optimized formulations based on specific input parameters.
Model Selection Process
Initially, four models were selected after thorough evaluation and research:
- Random Forest
- XGBoost
- Decision Tree
- CatBoost
After further experimentation, Decision Tree and XGBoost emerged as the top-performing models due to their superior accuracy and predictive capabilities.
Key Considerations in Model Development
- Formulation Constraints: Ensuring the generated formulations adhered to chemical composition and cost limitations.
- Multiple Target Variables
- Strength Standards at Different Days: Predicting material strength across various timeframes.
- Raw Material (RM) Values: Evaluating the impact of different chemical components in the formulation.
To ensure data integrity and security, a robust access control system was implemented, aligning with client requirements and the industry’s best practices.
User Registration & Role Assignment
- Restricted User Creation
- Only Super Admins had the authority to create new users and assign roles.
- This was enforced due to security concerns and client-specific requirements.
Login & Authentication
- Company Domain-Specific Emails
- Only employees with organization-approved email domains could log in.
- This ensured secure access and prevented unauthorized users from entering the system.
Role-Based Access Control (RBAC)
Each user had restricted access based on their assigned role:
- Raw Material (RM) Database Admins → Access to the RMDB screen for managing raw material data.
- Sales Managers/Teams → Access to existing formulations and the main dashboard for monitoring sales insights.
- Formulators → Access to:
- Creating new formulations
- Adding lab results
- Main dashboard for tracking formulation progress
Data Flow Architecture:
- Flat Files: Raw data is stored as flat files.
- Microsoft Azure Blob Storage: These files are stored in Azure Blob Storage for further processing.
- Bronze Layer: Raw data is imported and compiled.
- Silver Layer: Data preprocessing occurs (cleaning, transformation, augmentation).
- Gold Layer: The data is aggregated at a business level for insights.
- Processed data flows into an AI Virtual Machine with higher computational power.
- Feature Engineering is performed to extract meaningful features.
- A Machine Learning Model is trained and applied to the data.
- Optimization is applied to improve results.
- The results are integrated into a UI (dashboard and predictions panel).
- The UI operates on a lighter virtual machine with lower specs.
- Users interact with predictions and dashboard insights.
Challenges and Resolutions
Challenges Faced
During the development and implementation phase, we encountered several challenges:
- Data Complexity: Handling diverse datasets and ensuring the ML model generated accurate results.
- User Adoption: Resistance from users unfamiliar with AI-driven formulation.
- Integration Issues: Seamless communication between the ML model and the web interface.
- Security Concerns: Ensuring encrypted data transmission and secure authentication.
- Performance Optimization: Balancing speed and accuracy in formulation predictions.
Solutions & Resolutions
To overcome these challenges, we implemented:
- Data Standardization: Cleaning and structuring the dataset to improve model accuracy.
- User Training & Workshops: Conducting training sessions to familiarize users with the AI-driven approach.
- API Optimization: Enhancing API performance for seamless integration between the web interface and ML models.
- Robust Security Measures: Implementing encryption, multi-factor authentication, and periodic security audits.
- Performance Tuning: Refining ML algorithms and optimizing system response times to improve efficiency.
Key Takeaways
The successful implementation of the Project Elevate Formulation AI led to several valuable insights:
- 40% Reduction in formulation development time.
- Cost Optimization by dynamically adjusting formulation costs.
- AI and ML can significantly enhance formulation efficiency and accuracy.
- User training and engagement are critical for seamless adoption.
- Enhanced Security with role-based authentication and encrypted data transmission.
- Continuous monitoring and improvements are essential to sustain long-term success.
- Collaboration between technical and business teams is key to a successful implementation.
Final Words
The successful implementation of the Project Elevate Formulation AI has proven that AI and ML are not just buzzwords; they are essential tools for transforming how businesses operate. By embracing automation, the organization has cut down formulation development time by 40%, optimized costs, and ensured a more accurate and efficient process. The project’s success highlights the power of AI in improving operations while enhancing security, data integrity, and user experience.
For businesses looking to achieve similar results, Royal Cyber is the partner of choice. With our deep expertise in AI and machine learning, we can help you modernize your processes, optimize your operations, and achieve your business goals with cutting-edge technology solutions. Contact us today to learn more about how we can help you drive your business forward with AI-driven solutions.
Author
Zeeshan Mukhtar
Learn how to plan an Optimizely CMS 13 upgrade with .NET 10, Optimizely Graph, Visual …Read More »- Learn how AI meeting notes automate summaries, action items, and insights from video meetings using …Read More »
- Boost AI discovery for ecommerce with AEO, GEO, and MetafyAI. Optimize product data, structured content, …Read More »
