International Institute for Analytics

Procurement Integrity (PI) represents a broader problem and bigger opportunity than most businesses recognize. Comprehensive PI programs continuously validate purchasing transactions, using data and analytics to trace patterns, spot anomalies, and reduce fraud, waste, and abuse. The problems uncovered range from occasional opportunistic fraud to ongoing organized fraud, from duplicate invoices and other improper payments to regular kickbacks, from conflicts of interest to ongoing collusion with suppliers. Continuous monitoring of anomalies in procurement and supplier due diligence processes reveal potential problems, including data issues and process breaches, and help focus the efforts of audit and other investigative staff.

Inquiry:

We’re new to HR analytics. What kind of projects should we look to tackle to show some initial wins?

Inquiry:

How can we measure the lift from coupons used and what’s the best way to deploy the models to do it? We know we have lift coming in from various combinations of promotions, but we need a more granular view. Unfortunately, we don’t have access to web data, although we can track the coupons used in our brick-and-mortar stores. What do you think?

Inquiry:

The analytics team and IT partners team just merged and we’re running into a number of challenges. We have different definitions of what it means to put thing into production, and we’re not communicating cleanly with the business stakeholders. How can we make this work?

Inquiry:

We’ve developed a successful natural language processing (NLP) algorithm and think it would be applicable to other areas of the business too. How should we deploy something like this so we can share it more broadly?

Inquiry:

We don’t currently have enough resources to tackle advanced analytics products. How can we help our leadership see and understand our needs? Also, how can we become more efficient with our large scrum team?

Inquiry:

We’re exploring ways to mature our fraud detection models. At the moment we rely on outlier-based detection, but we don’t have a ton of label data. How can we advance our outlier testing until we generate enough labels to build ML models? What else might you suggest?

Inquiry:

As a manufacturing company with a new data and analytics operating model. How does the role of analytics catalysts fit into our organization?

Inquiry:

As we work towards our goal of deploying more machine learning (ML) models into production, how do we position our MLOps needs to the organization? What needs to happen for MLOps to be successful?

Inquiry:

We’re considering buying a data catalog but struggle with the value proposition. What’s the best way to build a business case around purchasing a data catalog? Which vendors should we look at? For implementation, what are we looking at from a time-work effort standpoint?