Buyers at trading companies with large inventories have similar challenges. Having the right products in stock at the right time. Not buying too much and not too little. Anticipating seasonality and other factors affecting sales. In some companies, Excel is the only work tool.
Others have chosen systems that automate the routines. Now the development has taken another step when AI - Artificial Intelligence - and machine learning have been integrated into such a system.
A couple of years ago, a conference was organized on the theme "digital purchasing processes of the future". The focus was on the possibilities of Artificial Intelligence and Machine Learning in systems for supply chain management. One conclusion was that AI is a great but underutilized asset. They also called for system suppliers who dared to start AI investments.
One company that has done it is Promosoft in Gothenburg. In 2020, a long-term AI journey was started within the framework of an EU-funded project. The company's system SOLO has helped more than a hundred customers to automate and digitize warehouse management procedures.
- Our customers are trading companies with large inventories where the core business is storage and distribution. SOLO has helped these companies create significant business value. Time-consuming manual procedures are automated. Buyers can work more strategically. Reduced inventory values that result in reduced capital tie-up and improved demand forecasts are some examples. By harnessing the potential of AI, we wanted to create even greater business value, says Promosoft's CEO Nils Robertsson.
Better decision-making basis, safer forecasts and reduced capital tie-up
The investment in AI was started together with AI specialists from Chalmers Institute of Technology. In 2022, a first interim goal was achieved. Promosoft was then able to present an AI version where a machine learning model further increases the forecast precision of SOLO's statistical models. This becomes possible by feeding the AI model with large amounts of input data from business systems and also external information. AI "trains" SOLO to detect patterns and sales-influencing factors that sellers and buyers have difficulty seeing.
With machine learning, sales history can also be used smarter and more efficiently. How much will different products sell during the year? How much can a new product conceivably sell? When is the best time to have a campaign? How much should we keep in stock during the different phases of a product's life cycle up to end-of-life? AI clears up questions and delivers better decision support than is possible with experience and intuition.
In the long term, Promosoft will develop more machine learning models adapted to the requirements of different item categories such as high-frequency items, low-frequency items, seasonal items and new items. Magnus Törnqvist is system owner at Promosoft.
- The business benefit of using AI for inventory management is great and I am convinced that many companies want to take advantage of the opportunities. SOLO has been instrumental in automation and increased forecast accuracy, but with AI the system becomes even sharper. Safer forecasts mean less safety stock and further reduced capital tie-up. AI also means that the system has time to react to rapid changes in demand and thus minimizes the risk of product shortages and overstocking. A big advantage is also that order points for seasonally affected products are raised completely automatically. We see a very exciting AI journey ahead!