Simfoni–AI-powered reports for Spend Analytics

Simfoni received a scalable, large language model (LLM) solution that automates the creation of key insights for their customers. What previously required hours of manual work is now an efficient automated process bringing immediate value.

The company and product

Simfoni is a global spend automation platform that transforms business procurement with insights and strategies derived from data analysis. Their AI-powered solutions help Chief Procurement Officers and CFOs control spending, streamline processes, and support sustainability goals.

Simfoni provides leading global enterprises with Spend Analytics, eSourcing, and Tail Spend Management solutions. Their platform utilizes machine learning and artificial intelligence to accelerate and automate key aspects of the procurement process, saving customers time and money. With offices in the USA, Europe, Australia, and the Middle East, Simfoni works with its customers and the wider vendor community to improve vendor diversity and achieve strategic corporate social responsibility objectives.

Simfoni was recognized as a "Provider to Know” by Spendmatters for 3 years running and ranked as a Top Value Enabler in Spend Analytics.

Client’s feedback

Scope and Highlights

AI Development

Large Language Model (LLM)

Product Design

technologies: Python, Django

Team members

Proof of Concept (POC)

The Challenge

Simfoni faced a critical business challenge with its Spend Analytics platform. Their customers spend data can run into the billions of dollars. While dashboards help navigate the data, it was hard and time-consuming for customers to make the most of Simfoni's analytics insights to achieve their strategic objectives.

Simfoni first tested an AI-driven Factpack presentation highlighting significant changes in customer spending patterns and linking them to wider trends in the supply market. While this approach dramatically improved customer satisfaction and platform adoption, and demonstrated the strengths of an AI-driven approach, Simfoni wanted to go further that just providing facts, they wanted to provide actionable insights.

This led to a brand new product idea called Opportunities–an AI-powered tool to identify cost savings opportunities across procurement data automatically. With hundreds of clients, thousands of suppliers, and varying prices for similar products, manually spotting potential savings became increasingly impractical for their analytics team.

Key challenges

Ensuring the accuracy of data processed by the AI tools required additional data preprocessing and cross-checking.

Limiting the AI prompts to 2-3 iterations to maintain context and produce reliable results.

Prompting the AI to generate the specific chart types and visualizations required for the report format.

Ensuring the accuracy of data processed by the AI tools required additional data preprocessing and cross-checking.

Limiting the AI prompts to 2-3 iterations to maintain context and produce reliable results.

Prompting the AI to generate the specific chart types and visualizations required for the report format.

The Process

We leveraged multiple technologies to transform raw data into meaningful insights, starting with the Python Pandas library, which efficiently handles large CSV files. This tool allowed us to quickly load and preprocess extensive datasets, ensuring smooth initial data management.

By integrating Claude, a large language model (LLM) through AWS Bedrock, we could perform a sophisticated analysis of spending data. Based on client data, our team developed specialized scripts and prompts that enable Claude to generate insightful slide reports. To maintain data integrity, we implemented a validation step in which the numeric data generated by the AI was carefully reviewed and corrected if necessary.

For final reporting, we generated PowerPoint slides using the python-pptx library. Recognizing the library's limitations, our team developed custom workarounds and implemented low-level slide manipulation techniques to achieve the desired presentation format.

Simfoni platform was also integrated with a news provider, allowing the Factpack to include summaries of relevant industry articles.

Core efforts included:

Preprocessing extensive datasets and ensuring data is clean, structured, and ready for in-depth analysis.
Utilizing Claude's natural language processing to analyze spending data.
Applying an integrated approach to transform complex data into clear, actionable insights with minimal manual intervention.

Simfoni application screenshoots

Team’s voice

The Outcome

The final version, while called a Proof of Concept, was largely functional, and was quickly integrated into the live product by Simfoni’s team, under the name of “Insights AI”. What previously required hours of manual work now happens automatically.

The AI system delivers comprehensive outputs, including detailed spend change summaries, identification of key industries of interest, and processing news related to specified industries.

Any Simfoni customer can now generate an Insights AI report at the click of a button and receive ready-to-send PPT presentations for their clients, who can take further actions like accessing specific dashboards, viewing reports, or performing in-app tasks like classification.

For the "Opportunities" initiative, our team developed a sophisticated AI-driven script that analyzes procurement reports using large language model (LLM) technology. The prompt mechanism can uncover nuanced patterns and identify strategic savings opportunities, including targeted supplier consolidation approaches. Complementing the technical solution, we crafted a comprehensive PPT template with visual potential cost optimization pathways.

Key results

Increasing the value of the key Spend Analytics platform with insightful presentations for customers.

Simfoni saves time and effort while offering comprehensive data refresh updates to all clients at scale.

Adding AI-powered tools to Simfoni’s offering significantly enriched the company’s portfolio.