How an ML algorithmic suite provides 60% relevant insights in significantly less time

Eugenie.ai provides an algorithmic suite that analyzes data at scale to generate actionable insights.

The Context

With the constant influx of massive and complex data, it’s becoming increasingly difficult to derive any useful insight from it. Not only are human-led methods such as visual explorations time-consuming and error-prone, it also introduces cognitive biases in the decision-making process. This consequently leads to missed signals for opportunities and threats to business.

Eugenie.ai addresses this problem through its framework of spotting anomalies, exploring the possibilities, and exploiting the results that have bottom line impact.

The Software

The value proposition rests on three pillars:

1. Spot - identifying the most critical areas in the business. The application parses the entire historical data in order to analyze expected business performance and its prediction range within a given time period.
2. Explore - helping businesses understand the rationale behind outliers. The application figures out the rest of the KPIs in the data that drove the anomalous behavior of the target KPI and reports how they contribute to the change in the target KPI.
3. Exploit - analyzing the impact on the business by changing several parameters. The application simulates different business scenarios while users understand its effect on the anomalous behavior that was previously spotted.

Some of the functional algorithm blocks are:

• Outlier - detects anomalies (items, events, or observations which do not conform to an expected pattern or other items) in the dataset. This helps businesses make tactical decisions and take corrective actions to mitigate the risks.
• Early Warning - detects areas of the business that will not meet the target outcomes.
• Hotspot - finds the most critical areas of the business by highlighting the products or business units which have the most bottom line impact. It compares the product or business unit against its peers as well as its past performance and gives a comprehensive picture of business areas that are growing exceptionally well or declining abruptly.

The Workflow

The Impact

Some of the key results from clients are as follows (name not provided due to confidentiality)

• More than 60% of insights curated were found relevant by human experts.
• Insights-as-a service along with the narratives for root cause helps various types of end-users, such as a business analysts and data scientists, in their data analysis tasks.