A readiness checklist for procurement AI
Procurement teams can use AI to identify patterns in supplier performance by connecting relevant data sources, agreeing definitions in writing, and deciding permissions in advance, then asking questions like what suppliers are at risk of missing a deadline, and how does their performance compare to others.
Every week, procurement teams wait for answers to the same questions, from stock levels to supplier performance, while the rest of the organisation moves forward with its plans. This waiting game can be ended with AI, but only if the right groundwork is laid first.
Data connections are not optional
The most common mistake made when introducing AI to procurement is to underestimate the importance of connecting all relevant data sources. Without access to the maintenance log, for example, a shift supervisor cannot use AI to identify patterns in equipment failure, and so cannot plan preventative maintenance. This breaks the whole process, because the AI is only as good as the data it can see.
A procurement team that wants to use AI to compare supplier performance, for instance, must first connect the supplier database, the inventory management system, and the accounting records. If any of these are missing, the AI will not be able to give a complete picture, and so will not be able to answer the questions that procurement needs to ask. This is not a technical problem, but a practical one: if the data is not all in one place, the AI cannot use it to answer questions.
Definitions must be agreed in writing
Most people get wrong the idea that AI can somehow magically understand the nuances of human language, and so can work out what is meant by a particular term or phrase. This is wrong: AI is only as good as the definitions it has been given, and so these must be agreed in writing before the AI is switched on. A category manager, for example, must define what is meant by "supplier performance" before the AI can compare different suppliers.
This is not just a question of writing down a few key terms, but of agreeing a whole vocabulary that will be used by the AI. This vocabulary must be consistent, and must be shared by everyone who will be using the AI. If this is not done, the AI will not be able to understand the questions it is being asked, and so will not be able to give the right answers.
Permissions must be decided in advance
Another common mistake is to leave decisions about permissions until the AI is already up and running. This is the wrong approach, because it can lead to confusion and delay. Instead, permissions should be decided in advance, so that everyone knows who can see what, and who can ask what questions. This is particularly important in procurement, where sensitive information about suppliers and contracts may be involved.
For example, a procurement manager may need to decide who can see the details of a particular contract, and who can ask questions about supplier performance. These decisions should be made in advance, and should be based on a clear understanding of who needs to know what, and why. This will help to ensure that the AI is used responsibly, and that sensitive information is protected.
A first question worth answering
One of the best ways to get started with AI in procurement is to identify a question that is already being asked, and that the AI can help to answer. For instance, a procurement team may already be waiting to hear about stock levels, or to get updates on supplier performance. The AI can be used to answer these questions, and to provide real-time updates that can inform decision-making.
A good example of this is a procurement team that uses AI to track inventory levels, and to identify when stock needs to be reordered. This can help to prevent stockouts, and to reduce the risk of overstocking. The team can use the operations data to inform their decisions, and to make sure that they are getting the best possible value from their suppliers.
When AI is not the answer
While AI can be a powerful tool in procurement, there are some situations where it is not the right choice. For example, if the data is not available, or if the questions being asked are too complex, then AI may not be able to provide the answers that are needed. In these cases, it may be better to use other approaches, such as manual analysis or traditional reporting.
If you are considering using AI in your procurement team, but are not sure if it is the right choice, you can contact us to discuss your options. We can help you to understand what AI can do, and how it can be used to support your procurement goals. By working together, we can help you to get the most out of your data, and to make better decisions about your suppliers and contracts.
Frequently asked questions
What are the common mistakes made when introducing AI to procurement?
Underestimating data connections and not agreeing definitions in writing are common mistakes that can hinder AI implementation in procurement.
How can AI help with supplier performance comparison?
AI can compare supplier performance by analysing data from connected sources, such as inventory management and accounting records, to provide a complete picture.
Why is it important to decide permissions in advance?
Deciding permissions in advance ensures that sensitive information is protected and that everyone knows who can see what, preventing confusion and delay.
How can procurement teams get started with AI?
Procurement teams can get started with AI by identifying a question they already ask, such as stock levels or supplier performance, and using AI to provide real-time updates and inform decision-making.