Recently, I published this note on AI procurement in the Analyst Corner of the Weekly Spend Matters email (though I have more to say).
COVID-19 struck quickly and forcefully, crippling the supply chains of organizations that were unprepared for the majority of the world's supply chains. This not only exemplifies the need for improved supply chain visibility and control, but also enables organizations to more quickly procure and resource new products (PPE) and new sources of supply when needed. I have illustrated the need to do so. New modes of transportation were needed (as existing supplies were disrupted) and (as airlines slumped and small business drivers got sick).
There is no better time for vendors to advertise and promote their products. The vendors most vocally promote built-in best practices, project management, open APIs, supplier networks, virtual collaboration technologies, and specifically, these technologies include robotic process automation (RPA), machine learning (ML), ), powered by AI (“artificial intelligence”).
As organizations seek to understand their data, analytics take center stage, supplier discovery, once a nice-to-have, becomes critical, and supplier risk moves from a “nice-to-have” to a “must-have” in sourcing. did. Choosing a procurement platform (for organizations without frozen budgets). And of course, startups with AI/RPA and established suites can bring some capabilities here, particularly around resourcing, more proactive risk identification and mitigation, and/or more intelligent analytics. Having acquired/implemented it, I started shortlisting everything.
You can be more specific, but it's more important to focus on what you should see rather than what you see. Specifically, the focus is on SUM (spend management) and the organization's MDM (master data management), particularly suppliers and products, risk identification and mitigation plans, and external risk tracking and analysis, rather than RPA or AI. (which doesn't really exist), supplier networks, or even more advanced analytics (usually if the organization isn't making the most of what it already has).
Spend Under Management allows organizations to instantly understand what they're buying, from whom, where, and when. Master data management allows you to have a uniform and accurate list of the suppliers your organization currently uses, the products your organization procures, and the services your organization relies on. You can't identify risk unless you know what you're buying, who you're buying it from, and how important it is. More importantly, they also cannot identify the events that cause the risk to materialize.
As natural disasters continue to rise, with the risk of devastating pandemics, global political tensions rise (thanks to the election of populist leaders), and economic certainty becomes more precarious, risks will materialize. The ability to respond instantly is becoming increasingly important. . This is only possible when organizations are integrated with global risk monitoring software that identifies strategic product/supplier/service risks, identifies triggers, and constantly monitors for signs that risks are materializing. .
Conversely, Robotic Process Automation (RPA) can only automate tactical tasks based on predefined rules, while Machine Learning (ML) can analyze historical data and automate tactical tasks based on predefined rules. , all that AI (augmented intelligence, not artificial intelligence) does is identify decisions that would normally be made in normal situations (so that RPA can further automate them). It's just a matter of detecting that it's not proceeding, and how (or what to do about it).
We use RPA, ML, or AI (which can automate and eliminate much of the tactical work), advanced analytics (which can identify trends, opportunities, and outliers), or supplier networks (which can improve supplier discovery). ) is not discounting its usefulness. (valuable during a crisis), but none of these would have prevented the supply chain disasters that have resulted from the pandemic and the associated sourcing and sourcing challenges, nor have they been able to manage spending. He pointed out that organizations that do not have the appropriate infrastructure in place to do so are institutionalized. Data and risk are doomed to failure again and again.
And this isn't the first time I've published words along these lines. A while back, one of my articles on coronavirus response was “AI won’t save you.” Advanced technology doesn't work without good data, good processes, and great visibility into both.
Even today's level of AI is usually just augmented intelligence in specific situations. Nothing can be expected unless the system learns the behavior and that learning occurs only through supervised, iterative application. In other words, you can't expect anything. and long-term, semi-supervised learning on large, clean, well-classified, and well-curated datasets. You can't find a solution off the shelf, install it, and expect to get anything good from it if your data is a mess. In fact, this will cause supply chain failure faster than past debacles (Big Bang ERP projects, SCP projects, and other initiatives that plagued multi-billion dollar companies) — the author's classic on sourcing innovation Check out some of our posts This includes an article on how relying on ERP can go on the supply chain disaster register.
Similarly, without processes in place, any rational insights or recommendations that advanced analytics, ML, or AI systems can derive from data about next steps will be wasted. Because there is no reasonable and reliable way to do it. Act on those insights. Additionally, these processes must be powered by a robust software platform. That platform could be an S2P suite, process orchestration software (such as ignio or zapier), or specialized S2P project/program management (such as Per Angusta). As long as the process is well defined and the platform supports it, you'll be fine.
However, as stated in the Analyst Corner article cited above, there are two keys to getting the right data.
- MDM (Master Data Management) — specifically for supplier and product/service data
- SUM (spend under management) — regardless of whether the spend is “sourced” or not
Considering all the risk management solutions your organization craves, all the advanced analytics solutions you want to use, and all the automation you crave (because you're not going to have more headcount anytime soon), they're all important. is. It relies on data, and the more data the better. Let's look at them one by one.
- crisis management A supplier risk solution should be aware of all your suppliers, preferably where they are working, as well as the products you are purchasing and the bills of material (BoM) they contain.
- analysis Predictive (and prescriptive) requires large amounts of cost/metric data over long periods of time
- automation You need rules that you can follow based on data values/buckets/trends, etc., as well as integrate with all the software platforms you use.
Still, most organizations still have dozens of data silos, and the idea of MDM is that they are dumped quarterly into a data lake for spend analysis from the previous quarter. This alone will tell you which suppliers you recently started using, which suppliers you no longer use, which products are nearing the end of their lifespan (and where shortages may not be a major issue). ), which products are key growth products (and critical to success).
Additionally, Managed Spending (SUM) is limited to contracts negotiated by Sourcing. This is not enough. To truly take advantage of best-in-class Source to Pay (and related) technology, you need to have as close to 100% of your spend as possible under control. This means that everything except in-house payroll must go through the procurement system, whether you negotiate spend or not. And whether you can control your spending.
If your marketing department, legal department, or C-suite wants to control spending (on their own projects), or if your spending is less important and your involvement will save enough to make the effort worthwhile. If you can, that's fine, but you still need to keep going. Track suppliers, products, services, and other critical information through approved procurement systems, giving your organization complete visibility into your spending. Only then can risk monitoring insights provide appropriate alerts, automated ML-backed analysis systems provide you (and other spenders) with meaningful system-generated insights, and the platform All tactical procurement processes can be properly automated and require minimal human effort. .
So if you want a chance of surviving the next disaster, equip your procurement house to adopt the right technology and really realize the value that technology provides.
Have you signed up for Spend Matters' weekly email updates? In it, you'll find: Exclusive and valuable insights From one of our analysts. Must-read articles; of Events waiting for your RSVP; And there's also some self-promotion to help keep the lights on. Sometimes it can even be witty (we hope so). Sign up now and never miss a thing about spending, delivered to your inbox every week.
