Less than a year ago, we had a blog titled "AI in inventory management - do you want to be first on the ball?" We told a little about PromoSoft’s development efforts to supplement SOLO's statistical models with Artificial Intelligence (AI) and Machine Learning (ML, machine learning). The ball has been rolling for a while and within a few months it will also roll into goal.
We are going to present blogs and articles with the results of the initiative and how the AI will lift digitized and automated purchasing processes to a new level. We already know that this will be the case. We also believe that we are facing a paradigm shift in terms of the role and importance of buyers in many companies, and a change in an often-traditional view of the purchasing departments' area of responsibility.
Just as economics departments had their lift by means of ERP systems are purchasing departments to get managerial approval when they see how AI and machine learning create new added value of great importance both economically and competitiveness.
Digitized and automated purchasing processes with SOLO have already become something of an alarm clock at management level in many companies. The insight into how important the purchasing function is will be even stronger when SOLO is supplemented with AI. That is our conviction!
A big thank you to you who contributed with test data
PromoSoft's AI project has been conducted together with Hans Salomonsson at the company Machine Intelligence and researchers at Chalmers University of Technology. What makes it possible to integrate AI with SOLO is that Machine Intelligence has developed solutions to what has been a major challenge for the application of machine learning: Big Data + Algorithms + Computer Power.
What has been an obstacle to a broader application of AI has been that the algorithms perform heavy calculations of a large amount of data. Computers with a lot of computing power have required expensive investments. What Machine Intelligence has done is to optimize the core of the AI systems so that they are faster and require less computing power. This while maintaining reliability. Thus, the way is open for industry-specific applications such as purchasing and inventory management systems.
An important driving force in the project has been the strong commitment from customers by providing real data as e.g., sales history (of course, this data is distorted during development work!) allowing good test conditions and comparisons of different AI models. A big thank you to you!
AI makes SOLOs statistical models even sharper
Already today, SOLO uses algorithms in the form of statistical models. For many demand planners, in combination with a digitized and automated purchasing and inventory management process, these have simplified many work steps. You can make smarter purchases with more accurate forecasts where the right items are in stock, at the right time and with the right volume. You save a lot of time, make sure to tie up less capital while you get full control over the entire process.
What now happens when AI becomes part of SOLO is that the statistical models are further sharpened by continuously training the AI algorithms and learning how to optimize the purchasing process and inventory management. The computers thus improve their algorithms themselves to make ever better purchasing decisions. The principle is that the larger the amount of data, the more accurate forecasts and greater precision in purchases. Therefore, over time, AI will also become an increasingly accurate tool when the system has access to sales statistics for several years. In addition, it will be possible with AI to predict the consequences of various external factors. SOLO will be able to determine which information is relevant and suggest which articles and volumes to purchase. This is not a utopia but a reality soon!
Is it dangerous to hand over control to algorithms?
As a demand and supply planner, are you in doubt about whether you should dare to trust "AI-based purchasing decisions"? Can I get rid of the job when the computer must make the decisions? How does AI work as a purchasing tool, really? The questions are many and we will in different blogs explain and talk more about how the technology for machine learning works and becomes a support in the purchasing planning work.
In psychology, there is a term for that hesitation one can feel before AI and algorithms: Aversion Algorithm. It is about the fact that many, regardless of industry, feel an irrational skepticism towards algorithmic decisions. A survey conducted by the consulting company KPMG (from 2018) shows that 2/3 of business leaders ignore decision material from data analyzes when it contradicts their intuition for strategically important business decisions.
We will therefore return with a blog about the CEO perspective on the purchasing function, and the development that is now taking place and where AI gives a new dimension to the concept of "digitized inventory management"!