Last week we asked what would you accomplish by 2020, which is less than 16 months away. Between 2008 and 2013, all the big analyst firms and thought leadership vendors painted a glorious picture of where Procurement would be by 2020 — a picture which isn’t even close to being a reality. Simply put, the vendors haven’t advanced technology to the point where it was supposed to be and, as a result, while you got more integrated, streamlined, easier to use platforms with friendlier, and sometimes even mobile, interfaces, you haven’t really obtained new functionality.

By 2020, our software was supposed to be smart. It was supposed to be doing most of our work for us. Tactical procurement was supposed to be a thing of the past. Paperwork was supposed to be over and done with. Data processing and verification automated. And Sourcing was supposed to be smart … not just more functional. But that’s what your average S2P platform is. More functional. All of the great advances we were supposed to have in the average platform aren’t there.

But, fortunately, if you’ll step outside the S2P marketplace and look at the best of breed players, across the entire space, you’ll see that most of the functionality you were promised is there, just in bits and pieces across a dozen or so best-of-breed players. So in addition to:

• Invoice Automation
• Supplier Identification
• Automated Supplier Discovery
• RFX Process Automation
• Should Cost Modelling
• Guided Workflows
• Automated Spot Buys

What else is there?

Cognitive Buying

This was the holy grail we were promised, but not the holy grail we were delivered. Instead of AI helping us buy better, we are running on glorified decades old tech that just helps out collect data and do tactical paperwork processing faster, but we still have to do all the data collection and review all the paperwork. Even dumb tactical work hasn’t been eliminated in the average platform. It’s no wonder that most people don’t even know what cognitive is.

But there are so many low-value, minimally strategic purchases that need to be properly sourced, and which take up way too much buyer time for the value that is delivered. But if a sourcing system could not only automate most of the work, but make rather obvious decisions based on simple process rules, that could eliminate most of the manual effort. A buyer would only have to review the suggested decisions at key points where there is a shadow of a doubt.

And even then, if there is a high probability a buyer would make a certain decision, why shouldn’t the system make them automatically if the dollar value is low enough, the strategic importance is low enough, or the risk is low enough that the decision doesn’t really need to be analyzed when the confidence factor is high enough.

So how much would you really need to make this happen? Not as much as you think. The core capabilities are:

• Rules-based workflow
  that allows the sourcing process to be well defined with gated decision-based branching points
• Should-Cost Models
  that allows the program to compute precisely what the average / expected market cost for the product / service should be
• Risk Models
  that capture both a product-based risk profile and a supplier-based risk profile
• Market-Data Feeds
  that contain current raw-material, energy, labour, and average mark-ups for the industry
• RPA
  that can automate execution of the rules-based workflow based upon automatically derived (and human initiated) decisions
• Behavioural Modelling
  that can monitor human decisions and actions and learn what a buyer would do under atypical or borderline circumstances
• Machine Learning
  that will take the outputs of behavioural modelling and human decisions and derive modified rule-based workflows, risk models, and automations to allow more fully automated processes in the future

And if you’ve been following along, you’ll know that there are a number of new best-of-breed systems out there with much of this capability, albeit most providers don’t have this in one system. However, as per yesterday’s post, Xeeva is pretty close. With the exception of behavioural modelling, Xeeva is pretty much there.

But if you happen to be a lucky buyer in electronics, you have a solution that meets all the criteria that goes by the name of LevaData. It’s one of the handful of providers trying to take Sourcing where it needs to be. Like every solution provider mentioned in these posts, they’re another company to keep an eye on.

And this is where Sourcing should be. So why isn’t it?