Advisor

How AI and ML Are Optimizing Procurement

Posted May 4, 2021 in Data Analytics & Digital Technologies
Robot shopping

I recently began examining how artificial intelligence (AI) and machine learning (ML) are changing procurement. In this Advisor, I explain my initial findings.

The application of AI and ML to purchasing and other procurement activities is relatively new; however, we are seeing increasing interest among organizations seeking to apply these technologies to optimize their procurement and other supply chain processes by:

  • Automating and optimizing time-consuming tasks

  • Generating additional insights via the application of predictive analytics and other ML algorithms as well as natural language processing (NLP) and image recognition to analyze large, complex data sets

Automating Time-Consuming Tasks

ML can help automate tedious procurement tasks such as purchasing approval workflows, reordering, and scheduling deliveries. Automation reduces costs and frees up employees so they can focus on more high-value activities.

Analyzing Large Data Sets

Organizations want more insight into their spending, supplier relationships, product/materials selection, and other procurement activities so they can identify opportunities for cost reduction. ML can automate the analysis of large amounts data to detect anomalies and performance indicators pertaining to the key aspects of procurement (e.g., price fluctuations in key components and raw materials).

ML-based analytics can generate insights into purchasing behavior by allowing purchasing, procurement, financial, and business professionals to see who is buying what, the costs incurred, and from whom they are buying.

Monitoring and analysis can take place in real time (or near real time) to uncover buying behavior and patterns that can help such professionals make better decisions and assist them in formulating policies and guidelines around employee spending.

Market Trends and Developments

The most important industry trend today is the availability of cloud-based platforms powered by AI analytics, which are designed specifically to support various procurement operations (e.g., spend analysis, supplier risk management, supplier information management). Many of these providers are relatively new companies. Some are well-established enterprise software providers (e.g., SAP). Key players include:

  • Amazon Business — uses ML to optimize B2B operations

  • Eyvo — hosted procurement applications

  • Medius — cloud procurement platform, including spend analytics

  • RapidRatings — supplier risk management

  • riskmethods — supplier risk intelligence/risk management

  • SAP/SAP HANA — uses ML, predictive analytics, and image recognition to streamline the purchasing process with visual search and automated purchase order approval

  • Sievo —  AI-powered procurement analytics

  • Simfoni — AI-powered spend intelligence

  • Stampli — payments

  • Suplari — spend intelligence/supplier risk analytics

  • Tealbook — supplier information management (uses AI to constantly update and provide a clean source of supplier information)

  • Tradeshift — B2B marketplace for procurement

These platforms are continually evolving with the addition of new AI capabilities. Basically, these providers are trying to “Amazonize” the corporate procurement experience by embedding automated analytics, product recommendation engines, visual and personalized search, predictive merchandising, supplier risk ratings, and other AI-powered capabilities (which have been used to transform the consumer e-commerce experience) within the various analytic and data management components of their respective cloud procurement platforms.

Cloud procurement platforms are appealing to end-user organizations for several reasons. First, their applications feature prebuilt ML, predictive, and other AI analytics that are embedded within the workflows of the various procurement processes and tasks they are targeted at.

These platforms also typically provide prebuilt end-to-end data integration and automated data transformation capabilities specifically built for procurement, including for cleansing, normalizing, and classifying data sourced from various enterprise sources. This offers a single source of clean, integrated data that lets organization assemble a more encompassing view of procurement activities (e.g., for spend, supplier analysis, cost analysis, and product/materials selection).

Key Applications of AI/ML in Procurement

There are a number of examples of AI/ML used within procurement functions, including:

  • Assisted search and purchasing recommendations

  • Predictive analytics for stock in transit

  • Spend analysis

  • Supplier risk management

  • Automated purchasing review/approval

Assisted Search and Purchasing Recommendations

ML can assist buyers in finding what they need quickly. This includes the use of advanced image recognition (i.e., based on deep learning neural networks) to power visual search engines. Employees can just snap a picture of a laptop or some other piece of equipment they need to order, and they are presented with a list of possible choices from the catalog.

Predictive Analytics for Stock in Transit

This is a frequently requested feature among organizations. ML algorithms are used to monitor the status of products/materials as they transit the supply chain. Predictive analytics serve to identify unforeseen delays and other issues before they happen. This allows procurement professionals to respond proactively to possible supply chain interruptions. Instead of having to scramble to find replacement products or materials, they can take steps to mitigate the problem in a timely manner, thereby avoiding product and production lines being caught off guard and eliminating unnecessary costs.

Spend Analysis

Spend analysis is one of the most sought after capabilities of procurement solutions. It involves analyzing current and historical spending data to identify cost reduction opportunities, improve strategic sourcing, and reduce procurement costs. NLP and ML serve to automate the various processes associated with spend analysis, with the former performing search and vendor/supplier matching and the latter classifying findings (from NLP analyses) into specific categories.

Spend analysis tools can feature sophisticated analytics combined with advanced visualization capabilities and filters so that users can explore key aspects of spending, including by category, supplier, region, and so on. Users can also separate long-term patterns from new anomalies by exploring trends over time.

A number of providers now offer AI-powered spend analysis platforms, including Sievo, Simfoni, and Suplari.

Supplier Risk Management

This is another much sought after capability by organizations. In supplier risk management, ML and predictive analytics are used to monitor and identify the financial health of suppliers, allowing organizations to protect mission-critical supply chains and mitigate against future risk.

Supplier risk management applications have become quite comprehensive when it comes to functionality offered. For example, take riskmethods. In addition to monitoring and identifying new and emerging supply chain risks by analyzing data extracted from different sources in real time, riskmethods can assist with assessing supplier health by detecting vulnerabilities at the category level. Another component, riskmethods' Action Planner software, lets purchasing professionals collaborate across the organization and with suppliers in order to proactively mitigate risk and avoid costs.

In an example of how cloud procurement platform providers are moving to expand their platforms' capabilities, Suplari has partnered with RapidRatings to integrate the latter's data about suppliers' financial health into Suplari's Spend Agility platform. The purpose is to enable procurement, sourcing, supply chain risk, and finance professionals who manage spend from thousands of direct and indirect suppliers using the Suplari platform to receive an alert notifying them in advance about a supplier’s degrading (or improving) financial position (see Figure 1). This allows them to take timely actions to mitigate risks to the supply chain, such as finding a new supplier and negotiating a new contract.
 

Suplari Spend Intelligence Cloud with RapidRatings supplier financial health ratings.
Suplari Spend Intelligence Cloud with RapidRatings supplier financial health ratings. (Source: Suplari.)


Purchasing Review/Approval

Here, AI is used to automatically review and approve purchase orders, making the process more efficient and a better experience for the employee. Automation can include the use of the search-by-image capability discussed above.

SAP offers a good overview of how an automated purchasing review/approval process could take place using the visual search and ML capabilities in the SAP HANA database.

An employee needs a new laptop computer. So he snaps a photo of a colleague's laptop, and an ML algorithm matches the image to potential product photos in the catalog. The employee then selects the correct model from the list and adds it to the shopping cart to check out. The responsible manager is notified that she has new items to approve, and the request is directly guided to her inbox for her approval. She then uses an ML-based feature that predicts the confidence level of an approval based on historical data. With a high confidence level, she approves the laptop purchase request, resulting in a purchase order (PO) being auto-generated.

This PO could then be tracked by an operational purchaser. Realizing that a follow-on document for the original purchase requisition has been created, he looks at a visualized document flow that shows the auto-generated PO and its number. To get order execution details, he enters the number in the “Monitor-Purchase-Order” operations app whereby he receives a status order visualization and an ML-based prediction that the laptop will arrive a day early.

Conclusion

Various AI technologies, including ML, predictive analytics, NLP, and image recognition, are helping to reshape procurement operations. These developments are most apparent in the emergence of AI-powered cloud procurement platforms.

In this Advisor, I provided some example use cases of how AI is now utilized to automate and optimize procurement activities. The reality is, as with all things involving AI, both market and technological developments are evolving rapidly, and we can expect to see AI continue to have a significant impact on procurement for the foreseeable future. If you are interested in learning more about the use of AI in procurement, I recommend you check out the in-depth guide from Sievo Oy, “AI In Procurement.”

Finally, I'd like to get your opinion on how you see AI impacting procurement, including any industry trends, solutions, or issues. As always, your comments will be held in strict confidence. You can also e-mail me at experts@cutter.com or call +1 510 356-7299 with your comments.

About The Author
Curt Hall
Curt Hall is a Senior Consultant with Cutter Consortium’s Data Analytics & Digital Technologies and Business & Enterprise Architecture practices and a member of Arthur D. Little’s AMP open consulting network. He has extensive experience as an IT analyst covering technology trends, application development trends, markets, software, and services. Mr. Hall's expertise includes artificial intelligence (AI), cognitive systems, machine… Read More