Artificial intelligence (AI) is an increasingly tangible reality for private and public organizations. Many Purchasing Departments are preparing for the arrival of this technological groundswell which will profoundly transform them.
Purchasing processes (S2C & P2P) naturally handle a large volume of data. They are at the crossroads of other equally data-intensive processes (quality, compliance, accounting, sales, etc.).
This intensive use of data is what AI is based on. Indeed, through data, AI can extract relevant information and identify a logical path. As such, AI represents a "new frontier" in decision support.
We can imagine that Purchasing Controlling will quickly equip itself with this new technology and significantly affirm a very "cautious" positioning today.
The “Augmented” Purchasing Controlling
Until recently, Purchasing Controlling encountered several difficulties, which hinder its development:
1. To have access, in real time, to data from various sources (including external sources)
2. Clean up and reconcile data of very heterogeneous quality
3. Automate and maintain a high number of process
4. Create procurement analyses, including predictive and even "prescriptive"
The good news is that the convergence of "Big data", "Extract Transform Load", "Artificial Intelligence" and "Data visualization" technologies makes it possible to remove all these obstacles.
Our opinion is that the average cost of implementing an advanced decision support solution, oriented towards Purchasing, is divided by 5 or 10. If we compare it with mainstream solutions.
We recently "augmented" a purchasing business intelligence solution in three months. We have mobilized the following technologies:
1. API, Open Data and Big Data orchestrator: The implementation of an API orchestrator has allowed us to access a wide variety of data, including external data (Open Data). This data quickly reaching a significant volume (Big Data) was stored in the "cloud" to free itself from any constraint related to infrastructures.
2. ETL The implementation of ETL technology is the opportunity we have been given to clean, "straighten" and consolidate raw data from various sources and then industrialize these processes.
3. AI We used artificial intelligence algorithms to analyze data and search for recurring patterns (Machine Learning) ... Consequently, many activities related to purchasing performance measurement have been automated.
5. Data visualization (Descriptive, predictive, prescriptive "analytics") The purpose of a decision-making system is to produce indicators to understand what has happened (Descriptive), what can happen (Predictive) and what should be done (Prescriptive). New "data visualization" solutions already implemented enable the production of dashboards and descriptive or even predictive analyses.
The implementation of these new technologies will transform Purchasing Controlling in many ways. First, we can anticipate that the way Purchasing activities are reported will be impacted. Moreover, the dialogue between Purchasing Controlling and the other departments of the organization (especially, the "Finance" department) will evolve.
The role of Purchasing Controlling will undoubtedly be strengthened. From a simple interface with the Departments impacted by Purchasing, it will become the driving force behind the implementation of these decision-making solutions.
Training and recruitment needs will soon emerge. When they exist (which remains rare), purchasing managers are not familiar with the new possibilities available to them and even less with the techniques of "Data labeling", "Machine learning" or "Unsupervised learning"...
Which path to follow?
When new technologies emerge, organizations (whether public or private) have different adoption paths. Some take a cautious path, taking small steps, others take more ambitious but dangerous paths, others finally leave it to the former to lead the way and rather choose to "follow".
E-Sourcing and e-Procurement projects, launched at the very beginning of the 2000s with the emergence of the Web, marked a break with those of the last century. Pharaonic projects have given way to much more reasonable approaches, guided by "business" expectations. Today, almost all the solutions on the market are available in SaaS mode, they allow maintenance costs to be controlled but leave a small part to specific developments.
The new technologies that are now available will not only offer a new field of possibilities from a "functional" point of view, but they will also allow a further significant reduction in IT costs while giving much more agility to the Information System. It can be built extremely quickly, and, like a "Lego", its bricks can be modified to meet the organization's needs. IS becomes, once again, a real competitive asset.
While it is important to not waste this opportunity, we recommend a progressive approach as the impacts on the IS and on the business are potentially significant.
Getting to grips with this new technology and all its possibilities requires time and a little hindsight. As we have written further on, these solutions are structurally scalable, and the disadvantages of an "agile" approach are reduced.
We therefore recommend starting with the construction of simple dashboards but with a real business impact. It is a question of making an impact and starting the transformation process with all the support, starting with the Executive Management.
For example, we have chosen to design solutions that integrate data (internal and external) that until now have rarely been reconciled because they come from sources that are "distant" from each other. Once the bridges were created, we gradually increased the depth and use of the analyses.
The perspectives for immediate gains (as well as the interest in understanding technologies and working methods that will quickly become unavoidable) amply justify an investment. Like any "revolution", it is important to integrate these impacts into the strategic trajectory of the Purchasing function, particularly in terms of IS master plan and management.