Back to Hub

AI in Supplier Management: Tomorrow (Part 2)

05/13/2019 By

In Part 1 of AI in Supplier Management: Tomorrow, we began our discussion of some of the AI-enabled capabilities that you can expect to find in tomorrow’s supplier management platforms, where we define AI as assisted intelligence (because, as we have discussed, there is no true artificial intelligence in enterprise platforms today and there won’t be tomorrow either). AI is a buzzword, not a reality. But we don’t need true AI to achieve software that can radically increase our productivity. Reaching assisted intelligence will add multiples to our efficiency and effectiveness.

In our last article, we discussed how tomorrow’s supplier management platforms will offer smart, automatic, supplier profile update (suggestions) — taking the headaches out of profile maintenance that results in most profiles being out of date in a supplier management system shortly after they are created; market-based supplier intelligence that is more in line and reflective with reality — and not just the experience of an anomalous customer subset; and real-time relationship monitoring that paints a relatively full picture of the relationship, not just a point-based performance picture.

So what else will tomorrow’s platforms do to help you focus more on the strategic side of supplier management? Let’s look at the next three areas:

  • Automated resolution plan creation, monitoring and adjustment
  • Automated risk mitigation strategy identification
  • Optimized real-time resource re-alignment

Automated Resolution Plan Creation, Monitoring and Adjustment

Any decent supplier management platform contains corrective action management capability that can be used to identify, track and monitor the progress on issues that arise during the buyer-supplier relationship — whether they be related to a delivery, a product, a category, a geography or just a single account manager-buyer relationship.

Some issues are small — OTD has slipped from 97% to 95%, minimal acceptable levels, and it’s just a matter of both parties tracking the next two dozen or so deliveries and getting them in on time to get the OTD percentage back to a comfortable level.

Some issues are not so small. Maybe the defect rate has risen to 3%, when it’s supposed to never exceed 1%, and is rising still. Unlike the OTD statistic, where you are just tracking ship dates and receive dates, this could require an extensive analysis into the production process, raw materials and even the product design. As a result, this would also require a fairly sophisticated project plan that would need to be tracked, monitored and updated regularly.

And some issues are just normal. A single batch of products has a higher-than-normal defect rate and replacements need to be expedited. The implementation services didn’t port the data properly — three years of historical data were lost. And so on.

For these issues in particular, it can sometimes take longer to create the resolution plans than it can take to actually implement them. That’s why tomorrow’s systems will come with a set of template resolution plans for a wide variety of issues along with embedded assisted intelligence (based on automated reasoning and machine learning technologies) that will be able to automatically select, and customize, a starting corrective action plan template that a buyer or account manager can tweak to deal with an issue.

Automated Risk Mitigation Strategy Identification

As we discussed in our first set of articles on AI in Supplier Management Today (Part 1 and Part 2), the best platforms today can, by monitoring internal (objective and subjective) metrics and external supply risk vendor feeds, identify potential risks. They can also take advantage of community intelligence to identify potential risk indicators.

Whether the risk is late deliveries due to a port disruption, no deliveries at all due to a shortage of materials or a supplier of bankruptcy, brand damage due to a recent detection of slave labor in a supplier’s supply chain, or one of a hundred (or thousand) other risks, today’s best platforms can detect them.

But the best platforms tomorrow won’t stop there, they will analyze your entire supply chains to identify potential risks before they happen. Sole-sourcing of critical products and materials will be identified, use of products or materials where supply barely satisfies demand, heavy supplier concentration in small geographic areas at risk for natural disasters, and so on will be identified as well, so that the platform can continually monitor for early warning signs.

But, as you probably guessed from the title of this piece, tomorrow’s platform will do more still. Just like they will come equipped with dozens of templates for corrective action management, which will be automatically tweaked to the situation at hand using automated reasoning or machine learning technologies, they will also be equipped with dozens of standard risk mitigation action plan templates for dealing with sole source disruption, logistics disruption, natural disaster disruption — and all by category, etc. which will also be tweaked as a suggested starting template as soon as a risk is identified.

For example, as soon as a mine collapse threatens supply to a primary supplier, an alert will be sent off to the account manager to deal with the issue. The default mitigation plan will be to work with the supplier to secure enough raw materials from an alternate source until the mine is expected to come back online and produce again. This will involve triggering an instant sourcing event with other market suppliers for quick material lock-up, probably an auction because they can be set up and executed quickly.

If enough supply couldn’t be secured fast enough, the mitigation plan would also have a fallback plan of shifting demand to another product that didn’t require, or at least required significantly less of, the raw material. This would trigger a corrective action plan that would involve sales switching strategies, and possibly renegotiation with existing clients to switch to a different product.

At each step, the right parties would be alerted, and if the plan was not executed fast enough, their superiors would be alerted as well to ensure rapid progress, with the frequency and severity of alerts and measures depending on the projected impact of the risk materializing, and the other risks in the queue that need to be dealt with. A supply shortage that could cost the organization millions would be escalated quickly while a supply shortage that might only cost $100,000 would be de-prioritized until the big risk was dealt with.

However, a complete overview of all risks being tracked, being detected, being dealt with and that need to be mitigated would all be summarized on multiple risk tracking dashboards, with underlying analytics around cost impact projections, longevity and effort to deal with them. And as soon as a risk manager selected one, it would be one click to bring up the recommended mitigation strategies (ranked), another click to generate a recommended mitigation plan, a few clicks to customize it, another click to identify the recommended resources and assets, a few more clicks to customize those, and a final click to select it and alert all of the recommended parties. What could take weeks of research and planning could be reduced to minutes when the general strategy for dealing with a situation is known, even if not by the risk manager.

Optimized Real-Time Resource Re-alignment

Corrective action plans and risk mitigation plans have something very important in common — people. In both cases, people are required to create them, approve them, execute them and monitor them. Resources will need to be constantly assigned, monitored, replaced and re-assigned as new corrective action plans or risk mitigation plans are created, and take priority over existing plans and/or can only be addressed by organizational resources with the right education and/or experience.

It will often be difficult for a project manager, or even a resource manager, to determine when to remove an organization’s best problem solver from a critical corrective action project to address a less critical risk mitigation project, or vice versa, even when the manager can’t think of someone else who could address the less critical risk mitigation project effectively, even when there is another moderately experienced problem solver that could step into the critical project. Especially if the manager is directly responsible for the critical corrective action project but only affected by the risk mitigation project.

But a platform that has access to the entire HR database, complete with skill profiles, curriculum vitaes, project history and project success/failure data (and resource effort correlations), can objectively determine the best matches, the chances of success when those matches are assigned, the relative impact they will have to other resources, and objectively recommend when resource re-assignments should happen for the good of the organization.

The platform will make these recommendations “on-line” in real time as new projects need to be staffed, and off-line when new data comes into the platform that affects the criticality, chance of success or relative value of the project. If, all of a sudden, the extent of a disruption is amplified (because new data shows that the projected supply shortage of a key material is double what was originally predicted), or the amount of alternate supply that materializes at the warehouse is half of what was committed to by an alternate source, or sales for the new product are ramping up faster than expected, the platform will instantly update the critical rating, the relative value and even the chance of success, alert the appropriate managers and make recommendations for resource re-allocations to maximize the chance of success (and minimize the forthcoming loss).

In other words, the supplier management platforms of tomorrow are going to — gasp! — actually help with supplier (project) management, and go beyond the supplier information management (SIM) focus that permeates the market today. Not that SIM is bad — it’s very good, and as Aravo, the grandaddy of SIM, demonstrated over 15 years ago, critical to success — it’s just not enough now that we are about to enter the third decade of the 21st century.

Topics
AI - Artificial Intelligence