Top 5 Gen AI use cases for P2P teams

Paul Roiter

Jul 2024

5 minutes read

Gen AI adoption isn’t a question of if but when for P2P teams. Business is transforming and Procure-to-Pay teams need to keep up. With its ability to multi-task and perform out-of-the-box tasks, including summarisation, Q&A, classification and more, Gen AI is changing the game.

The onslaught of Gen AI in the mainstream media (see ChatGPT or artificial image generation) makes it overwhelming to decipher fact from fiction and know where to start.

Screenshot 2024-07-12 at 11.37.37

Is Gen AI an uncontrollable force ready to take over the world? The answer is probably not. Generative AI is a tool like any other and can be simple if you break it down into manageable, controllable pieces. More than manageable, it doesn’t have to be a massive time or monetary investment. When done correctly, you’ll see benefits quickly.

In this post we'll look at:

How Generative AI differs from other models
The Gen AI sweet spot: Vendor query management
Other Gen AI applications for P2P teams
Getting started with Gen AI

Gen AI vs. Machine Learning

Most people will be familiar with previous iterations of AI in the form of Netflix recommendations or self-driving cars. These applications are powered by Machine Learning, a technology that involves teaching a computer to do tasks that humans do.

Generative AI typically refers to the use of Large Language Models (LLMs). These models are trained using Machine Learning techniques to take existing written language and generate new, convincing text in response to a prompt. This model can then automate a range of language-based tasks without the need to create task-specific machine learning models, which require significantly more time and effort.

Learn how Generative AI builds on Machine Learning from our Head of Data Science below.

Watch the full webinar here.

Vendor query management: Your low risk, high value Gen AI application

Choosing the right application of Generative AI for your team is difficult when it comes to balancing effort, money, risk and results. Vendor query management is the low risk, high value sweet spot.

Without Gen AI: Emails from suppliers around payment queries, statement provision and endless questions bog down the best of AP departments.

Some AP teams have ticketing solutions in place, but they’re generic and don’t work well for AP/P2P teams. Others manage vendor queries manually via shared mailboxes. When vendor queries aren’t handled promptly, follow up queries are sent, and the size of the AP inbox and consequent workload continues to grow. This can damage supplier relations and impact the ease, speed and cost of doing business.

With Gen AI: AI-powered tools can automate much of this process. For example, Xelix’s Gen AI powered AP Helpdesk works alongside your existing systems to monitor your Accounts Payable inbox(es) and read and understand emails. It then categorises the emails, creates tickets and triggers relevant workflows based on your preferred ways of working, including responding back to your vendors.

Considerations: This application is quick to implement, and drive value because the models are already trained. You’ll see results quickly without effort from your team, making vendor query management the sweet spot for Gen AI applications in P2P.

Learn more about vendor query management using Gen AI below.

Watch the full webinar here.

Are other P2P Gen AI opportunities worth it?

Gen AI can help P2P teams in other ways too. Stay ahead of the curve by checking out the other use cases listed below.

Contract compliance

Without Gen AI: Teams currently have to manually check that pricing and payment terms align to contract. Not only does this take time away from other more important business activity, but it means results are prone to human error.

With Gen AI: Pattern recognition software will be able evaluate contract terms for you, saving you time and effort.

Value: For maximum effectiveness, you’ll need to provide good examples for the model and train it as you go. There’s a lot of potential financial benefit with this use case, but it requires time and investment.

Document capture

Without Gen AI: P2P teams sift through data to find what they need. As with many other current states, this takes time and is prone to human mistakes.

With Gen AI: Your LLM would quickly and more accurately extract the invoice and purchase order data you need.

Value: With newly available models, you should see a material increase quickly.

Vendor risk assessment

Without Gen AI: Vendor research is completed by hand with teams going to various individual sources to find the information they need.

With Gen AI: You can more easily summarise reports on vendors from public sources. For example, the models can evaluate multiple vendors and even query the web to ensure you have all the information you need on potential vendors.

Value: This is an easy and low risk application for Generative AI. Although, it might deliver less overall value than other use cases.

ERP querying

Without Gen AI: With many ERP systems, it can be hard to find the exact information you need quickly and comprehensively, especially if you’re doing it by hand.

With Gen AI: So long as you have an easy connection to the system, you can find answers quickly to these kinds of queries. For example, you could ask a Gen AI agent: “How many invoices were posted into my Oracle Fusion last Thursday?”

Value: You’ll need a strong connection into your ERP system to get the answers you need. The model’s ability to integrate with different systems will ultimately affect its value.

Getting started with Gen AI

With all the information and developments associated with Generative AI, it can be hard to know where to start. There are a few steps you can take to ensure you get the most out of it.

Implementing Gen AI

Pain points: Start with the pain points you're experiencing instead of going for the technology that looks the flashiest. Unless it solves a problem, don't onboard technology for the sake of it.
Low-risk processes: Your first Gen AI implementation should have a quick time to value, accessible data and the ability to test and learn.
Success criteria: Set clear KPIs and success criteria. As with any other project, you want to evaluate success from the start.

Generative AI is a buzzword that might seem daunting. However, when it's broken down into easy steps, it’s a no-brainer for the future of P2P teams.

You can learn more about the future of Gen AI for P2P teams by watching this webinar on-demand: Low-risk, high value: The Gen AI sweet spot for P2P teams.