If AI is the light at the end of the tunnel, why are there so few success stories to be found? How do we enable smart procurement?
Success with today’s broad set of complex objectives requires Procurement leaders to think strategically and process ever greater volumes of diverse information. Unfortunately, this is an area with significant room for improvement at most organisations. A survey of over 400 procurement leaders by Forrester found their top priority to be “improv[ing] business insight on purchasing activity through reporting and analytics.”
The obstacles to more informed, strategic decision-making are quite consistent. The study, entitled “Enabling Smarter Procurement” found three common issues
1. Firstly, despite efforts at automating processes, too much capacity is still consumed by operational or manual activities. Teams must free capacity to work on new projects, conduct analysis and plan, but are struggling to do so.
2. Secondly leaders struggle to access relevant insights when and where they are needed. The volume of information now available is of little help if not digestible, simply leading to information overload.
3. Compounding this, respondents also cited poor data quality as a key challenge. Duplicate supplier records, inaccurate data and poor integration between systems all were cited as sources of data quality issues.
Fortunately, new technologies are available that can empower procurement to address these and other challenges and rise to the occasion. AI in particular is finally coming of age and often viewed as the answer to many of Procurement’s challenges. The same survey found that 71 per cent of business leaders plan to adopt AI in procurement over the next 12 months. Yet if AI is the light at the end of the tunnel, why are so few success stories to be found?
A key reason is the approach taken to implementing AI solutions to date. As vendors struggle to burnish their innovation credibility, there has been significant marketing ahead of capabilities which has led to unmet expectations post implementation. As capabilities are now coming in line with past marketing, this problem will subside. Of greater concern, the innovation race has led to nearly an exclusive focus on the algorithms, leading to poorly designed implementations. Less innovative but equally important areas, especially data quality, are being ignored. AI relies on a solid foundation of data, in terms of volume and quality, so solutions that offer clever applications alone are sure to disappoint.
To remedy this problem, organisations must implement AI in conjunction with cleaning up their data, rather than using poor data quality as an excuse for inaction. Source-to-Pay suites that are built upon a unified data model partially address the challenge by generating clean data that can be mined by AI applications across all processes. For example, suites with a single supplier record can provide true 360 degree visibility of supplier performance and activity, and enable AI applications to predict potential risks.
That still leaves issues with existing data or data in other systems. Here, master data management solutions should be leveraged that can actually fix issues in back end systems, linking vendor and item master records across systems. This further improves visibility and the potential for new and better insights.
Empowering procurement to make more informed, strategic decisions is no longer an option. There is simply no other way to effectively meet the broad set of objectives now expected. Fortunately, new technologies are finally reaching the level of maturity where they can have a transformative impact. By implementing AI applications in parallel with initiatives to improve their data foundation, leading organizations are both enabling smarter procurement today and ensuring they are well positioned to leverage tomorrow’s innovations.