How long before AI shows results in procurement?
Implementing AI in procurement takes time, as the initial fortnight is spent on groundwork, getting data in order, and establishing a common language for the AI to understand, after which the first question worth answering is usually something straightforward, like "what are our top suppliers by spend?".
You've just connected your procurement data to AsscherAi, and now you're waiting to see results from the AI. But what does that actually mean, and how long will it take to get there? Most people expect to see immediate changes, but the reality is that the first fortnight is spent on groundwork, not glamorous revelations.
Setting the foundation
The initial period is all about getting the data in order, which is not as straightforward as it sounds. This involves mapping the various data sources, cleaning up inconsistencies, and establishing a common language for the AI to understand. It's a tedious but necessary step, as the quality of the output is directly tied to the quality of the input. A category manager, for instance, needs to ensure that the supplier information is accurate and up-to-date, so that the AI can provide meaningful insights.
This process can be frustrating, as it feels like nothing is happening, but it's a critical investment of time. Without this foundation, the AI would be working with incomplete or incorrect data, leading to misleading conclusions. It's like trying to build a house on shaky ground - it may look fine at first, but it will eventually collapse.
The first question worth answering
Once the data is in order, the first question worth answering is usually something straightforward, like "what are our top suppliers by spend?" or "which categories have the most vendors?". This may seem trivial, but it's a vital test of the system, as it allows the procurement team to verify that the AI is working correctly and providing accurate information. A shift supervisor, for example, may use this information to identify areas where they can consolidate suppliers and negotiate better deals.
This initial question also helps to build trust in the system, as the team sees that the AI can provide reliable answers. It's a small win, but an important one, as it sets the stage for more complex queries and deeper insights. As the team becomes more comfortable with the system, they can start to ask more nuanced questions, like "which suppliers are at risk of disruption?" or "where can we optimize our logistics?".
When a leading measure should move
A leading measure, such as the number of queries answered or the volume of data analyzed, should start to move within a few weeks of the system being operational. This indicates that the AI is being used regularly and that the team is becoming more comfortable with the technology. However, it's essential to remember that these measures are not the ultimate goal - they're just indicators that the system is working as intended.
The real measure of success is when the procurement team starts to make decisions based on the insights provided by the AI. This could be something like renegotiating a contract with a supplier or identifying a new vendor to work with. When this happens, it's a sign that the AI is having a tangible impact on the business. For more information on how to implement AI in your operations, you can visit our operations page.
What makes it take twice as long
One of the most common mistakes that teams make is to underestimate the amount of time and effort required to get the AI up and running. This can lead to frustration and disappointment when the expected results don't materialize. What makes it take twice as long is often the failure to dedicate sufficient resources to the project, or to underestimate the complexity of the data and the queries being asked.
Most people get wrong the idea that AI is a plug-and-play solution that can be implemented overnight. The reality is that it requires careful planning, significant investment, and ongoing maintenance. When teams fail to recognize this, they end up taking twice as long to get the results they want, and often with subpar outcomes. If you're struggling to get your AI project off the ground, you can contact us for guidance and support.
When AI is not the answer
Finally, it's essential to recognize that AI is not always the solution to every problem. There are times when the data is too limited, or the queries too simplistic, to warrant the use of AI. In these cases, traditional methods may be more effective and efficient. It's crucial to understand the limitations of AI and to use it judiciously, rather than trying to force it into every situation.
Frequently asked questions
How long does it take to see results from AI in procurement?
The first fortnight is spent on groundwork, after which the AI can start providing insights, with leading measures moving within a few weeks.
What is the first question worth answering with AI in procurement?
The first question is usually something straightforward, like "what are our top suppliers by spend?" to verify the system and build trust.
What makes AI implementation take twice as long as expected?
Underestimating the time and effort required, and failing to dedicate sufficient resources, can lead to delays and subpar outcomes.
When is AI not the answer in procurement?
AI is not always the solution, especially when data is limited or queries are simplistic, and traditional methods may be more effective and efficient.