ChatGPT: Booster for Procurement

Zuiver Analytics uses AI software ChatGPT to help purchasing set up and improve the Spend Analysis, the Price Fluctuation Explorer and perform supplier, contract and sustainability analyses.

We conducted a study in May 2023 into categorizing suppliers using the language models hidden behind ChatGPT. The research was conducted using GPT-3.5 Turbo and GPT-4. For more information go to:
Categorization suppliers & ChatGPT

From April15, 2023 it is now possible to follow an online 2-hour hands-on workshop. Here you can experience how you, as a procurement professional, can use ChatGPT as a virtual assistant. The workshop is in English. For more information go tot Workshop ChatGPT for Procurement.

From April 7, 2023, Zuiver Analytics has been given specific access to the OpenAI experimental environment to test and experiment with the Plugin. With the help of the plugin, we are able to perform certain tasks that ChatGPT cannot perform itself but can request to perform this task. For example, working with current data.

More to read

Procurement Analytics

Procurement analytics is the process of collecting and analyzing procurement informtion (data) to form meaningful insights and effectively support decision making.

Examples range from historical procurement spend analysis reports to advanced analytics to forecast and budget for future decisions.

The need for purchasing analyzes arose from the desire of many organizations to obtain a consolidated picture of purchasing expenditure based on the data from the purchasing and invoicing systems.

Our mission is to create new and valuable insights using advanced data analysis techniques by integrating Business Intelligence with Data Science.

Spend Analysis

A Spend Analysis is considered the starting point. This is the process of identifying, collecting, cleaning, grouping, categorizing and analyzing your organization's procurement related expenses. This is done with the aim of gaining insight into purchasing costs and, where possible, reducing them and improving efficiency.

Our Spend Analysis consists of a Basic Spend Analysis and optional additional (advanced) analyses.

The Basic Spend Analysis contains an overview of the total spend, the spend per group, a world map, the ABC analysis, the Kraljic Matrix, Po-Non Po insights, Invoices stats, and the transaction details.

The options for additional analysis are:

  • The Purchasing Prices Index combined with the Price Fluctuation Explorer
  • Double Payments

Prices Fluctuation Explorer

In response to the rapid price increases in the market, Zuiver Analytics has developed a unique Price Fluctuation Explorer .

Get more insight into price increases and decreases with the help of a calculated index trendline and improve your budgets, supplier management, order times and order volume.

The indexed trendline is calculated over your purchases based on your own invoice and/or purchase order data.

The result is an indexed trendline, calculated based on a weighted average of your purchases. What you buy more of also weighs more heavily in the increase or decrease.

Explore the price increases per supplier, brand, product, project, etc. Determine which products are volatile and what this means for your own supplier price index (fluctation).

In order to generate the desired insights, your organization is learend to work with the calculated Procurement Price Index.

Product Classification (new)

Product classification is actually the first step towards a good Spend Analysis. A correct product classification is a prerequisite for a good understanding of purchasing-related expenditure and for shaping the purchasing strategy.

Classifying products and services has been a challenge within procurement for years. It is very labour-intensive and is therefore given little or no priority.

Due to the arrival of pretrained AI algorithms such as ChatGPT, we are able to classify products in a completely new way. The bottom line is that we let the AI algorithm do the most of the work. The dataset with the products is the input, the classified products the output.

We are still at the early stages of this development. The first results are very positive. If you want to test it on your own spend, for example in the context of a spend analysis, let us know. You can experience for yourself how much time you will save.

Double Payments

We assume that no organization wants to pay the same invoice twice. Unfortunately, it still happens regularly. The automation systems generally check whether the invoice number already exists. This is not always sufficient to prevent an invoice from being paid twice.

The Double Payments Algorithm analyzes your historical invoice data and 'smartly' determines which invoices have been paid twice. You are then able to reclaim the amount, whether or not double paid, from the supplier.

Reducing the double paid invoices from the historical invoice data is also an optional addition to the Basic Spend Analysis.

Get in touch to receive more information on how we can bring your data to actionable insights