Sourcing Automation vs. AI in Strategic Sourcing
There are many buzzwords regularly tossed around in this age of tech-based startups and digital transformation. Two such phrases are “Automation” and “Artificial Intelligence” (AI). It’s common to mix up the two, lump them together, or otherwise muddy the water of what is already a complex relationship between the terms. You’ve heard it, you think (or know) you need it, but you may not fully understand it. The great news? We’ve all been there.
Being at the forefront of new technology means we’re intertwined with many phrases and ideas people have been exposed to, but don’t yet understand. In this article, I dive into the difference between sourcing automation and artificial intelligence in predictive procurement and sourcing and how Arkestro uses both to drive impressive results for our customers.
Procurement Automation is the transformation of otherwise manual tasks into an action conditioned to trigger an unaided response to an event or series of events. Wow, what? I feel you. In short, automation automates manual tasks.
Great examples of automation include the “Welcome” email you get with a 15% off coupon after signing up for the newsletter of your favorite online retailer, or the recurring charge on your credit card for your Netflix subscription. A response (sending an email or charging a card) was conditioned to occur upon an event (registering for a newsletter or signing up for Netflix).
No one has to sit there and refresh their newsletter subscription list and manually fire off a welcome email. This is also why you’ve never heard of Netflix’s accounts payable department. The efficient nature of automation creates a force multiplier for the organization that utilizes it and removes friction associated with scaling usage or growth. In short, these tasks generally require some up-front legwork but streamline the process once completed.
Artificial Intelligence & Machine Learning in Strategic Sourcing
Artificial Intelligence (AI) in Strategic Sourcing is exactly as it would sound: intelligence demonstrated by machines. Machine Learning (ML), a subset of AI, is the ability for a computer or device to ingest information from its environment and respond to this data in a way that optimizes its ability to reach its defined goals. A great example of machine learning in action would be Tesla’s self-driving autonomous vehicles. A Model S sedan, when self-driving, will intake information from its environment through sensors (such as high-contrast road striping paint or distances from surrounding vehicles) and make the operational decisions that maximize its chances of successfully delivering the passenger to the intended destination safely and within the bounds of the law.
For a more day-to-day example, let’s take Netflix again. When you sign in to Netflix and select your account, you don’t generally use the search function to find a show, right? Most likely, you browse through all of the recommended categories of TV shows and movies to locate your next binge-worthy obsession (seriously, have you seen Season 3 of Ozark?).
This is powered by a powerful recommendation engine that takes into account which shows you’ve watched in the past and what genre blend you consume. For the throwback Netflix users out there, do you remember how cool it was to rate shows or movies after watching them? You still can today, but in the early years of the streamlining platform, this was the crux of their recommendation engine. As technology has advanced, it has removed the need for us to manually rate content to receive updated recommendations by defining triggers that are more seamless and require less user intervention.
This is why what is shown to you differs from what is shown in your spouse’s or childrens’ accounts. It’s continuously receiving feedback from every decision you make to skip, start, or watch a show, and refines its recommendations over time. What makes AI and ML truly intelligent is how malleable it is. Whereas automation is set up to execute preset tasks with a very specific outcome to each triggering event (although some automations can become very so complex they are indistinguishable from AI to the unindoctrinated), ML learns and modifies the outcomes as it receives new data and feedback. Mind-blowing stuff, right?
Apples vs. Apples
Here is a fantastic example of how automation and artificial intelligence differ in the same context.
Suzy Marketer sets up a great email workflow that triggers a welcome email once a new subscriber registers for her company’s newsletter. It will also send an automated follow-up email if the first email is not opened within three days. That’s automation.
Jane Marketer (not related) sets up an email flow in a platform that responds to open rates and manipulates the subject line copy to increase the percentage of opens the email receives. It will also intelligently manipulate the email content to drive further engagement and encourage recipients who open the email to take a desired action, be it scheduling a demo or clicking through to take part in an active sale.
The email platform is continuously receiving signals from recipients of the emails to refine the look, feel, and content of the email to maximize its goals of driving conversions. This is artificial intelligence, specifically, machine learning. Both are similar in nature but very different in execution.
The Power of Arkestro Predictive Procurement Orchestration
Now the questions remain: Does Arkestro use AI? Does Arkestro use automation? Does Arkestro use both? How does one Arkestro?
Quite simply, the answer is “yes” (to all but the last question, more on that later). Let’s take a short step back and think about how these concepts fit into our solution.
Arkestro uses automation to reduce the supplier management time required to execute a strategic sourcing project properly. Using a set of automated notifications and processes, Arkestro keeps suppliers informed of where they are in the bidding process and what the next steps are.
Whether it’s in collecting supplier surveys for a new RFP or notifying suppliers the next round of bidding has begun, Arkestro uses automation to bring clarity to a process that has historically been confusing for suppliers and has taken way too long to execute. For the buying organization, this reduces the toil and cost of doing business with your strategic suppliers, enabling increased collaboration and transparency that fosters better partnerships. This makes suppliers pretty darned pumped as well.
For the AI side of the equation, Arkestro uses behavioral analysis of participating suppliers, combined with a few other key data points, to provide individualized suggested pricing to the suppliers at the outset of the bid. As suppliers navigate through the automated negotiation process, they are provided with feedback on how competitive their most-recently submitted offer is relative to the pool of suppliers. Think about that for a moment compared to your current negotiation process. You may be thinking any of the following right now:
“What? Why would we do this?”
Studies have shown that going first in a negotiation provides substantial benefits to the party who makes the first move. By providing an initial price, it sets an anchor for the negotiations, thereby narrowing the range of acceptable offers at a price point set by the buyer. You could just as well collect quotes one-by-one from your suppliers to manually compare and contrast, but if they’re all over the place, is this the best process?
“I could just tell all our suppliers price X, and it would do the same thing, right?”
Ah hah, our first devil’s advocate! Fortunately for me, I can say, “no.” This is where the intelligence itself comes into play. Each supplier’s recommended price is entirely different based on several factors, such as our evaluation of how eager they are to win the business, incumbency, relationship with the buyer, and a host of other insights we can glean. We also take into account an intelligent zone of acceptable outcomes, whereby we identify the range of commercial terms that would be acceptable to both the buyer and supplier, as guardrails for the recommended pricing. For instance, as a buyer, you may be willing to pay more for a supplier you have a longstanding relationship with. In that instance, all other things being equal, the recommended price for that supplier may be a bit higher than it would be for a supplier invited to participate for the first time. Would you be willing to spend 6x more? Probably not, and we use that in our calculation through the zone of acceptable outcomes.
“What if a supplier doesn’t like this price?”
Suppliers have just as much power as they’ve always had to provide quotes that align with their own business goals. They are welcome to reject the suggested pricing and bid higher (or lower) than the pricing that was provided or bid only on select line items within the request. If they reject the price in favor of providing a higher price, the provided anchor still has influence in pressuring the supplier to provide more favorable terms than they would have otherwise.
“This is required for all bids?”
No, no, no. The power is ultimately in your hands. We believe in providing our customers with the toolset they need to do their best work, and enable buyers to turn the recommended pricing on or off when setting up their bidding event in the platform. Every bidding event is different, and we expect users of Arkestro to utilize the platform in a way that suits their needs best for each bid.
“Recap how the AI and automation work here again.”
Through a combination of historical pricing data and behavior mapping of supplier interactions within the platform, Arkestro provides recommended target pricing as an “anchor” to the supplier that encourages savings over the baseline throughout the negotiation. The automated execution of the negotiation itself reduces cycle time for buying organizations (enabling organizations to run more strategic sourcing projects), and the tailored AI-generated target pricing puts downward pressure on offer prices that wouldn’t exist otherwise. This is what we mean when we say “Better Quotes Faster.”
The Birth of Predictive Sourcing Enablement
Sourcing Enablement is the blending of these two concepts, automation and artificial intelligence, that elevates a sourcing professional’s ability to execute activities that positively impact their buying organization. Whether this is reducing the average time spent per RFP or providing intelligent insights to make a more informed award decision, Sourcing Enablement is how smarter organizations make the best decisions and achieve better quotes faster. This is why we say Arkestro is Sourcing Enablement for the Modern Business.
Now, back the final unanswered question, “How does one Arkestro?” If you’re interested in finding out how sourcing enablement with Arkestro can make an impact at your organization, request a demo today.