As we have been discussing this week, 2020 is fast approaching and, to date, you haven’t accomplished much with respect to what the big consultancies and “thought leaders” told you that you would accomplish by 2020 because, simply put, most of the vendors haven’t advanced the technology to where it was supposed to be and, as a result, you haven’t really obtained much in the way of new functionality.

Remember that we were told that software would be smart by 2020. It would do most of our work for us. Tactical would be a thing of the past as process, and paperwork, would be completely automated and electronic. And, most importantly, the software would not only allow us to focus on the strategic, but guide us through the strategic process. But we are still in a state of affairs where we have to not only set up the RFI process, but manually work our way thought it each time — using dashboards to figure out where we left off and what we have to do. And, as indicated in our last post, work our way through RFIs where, really, the dumb computer should be able to manage it for us.

So, what can we get?

Continuing on yesterday’s theme, more than we expect if we are willing to go best-of-breed along-side our primary S2P, get good at integration, and better yet at value identification and management. We do have the ability to do the following, with specific platforms in specific circumstances.

Guided Workflows

From the minute we log in to the minute we log out, the system should be putting us where we need to be when we need to be there. If we are only doing one event, we should be taken right to the point we left off next time we log in. If there is an information request, a system notification, or alert that should take precedence, that should be front and center at the top of the screen — whichever screen that is. But we shouldn’t have to go to a starting dashboard that summarizes one event, click into the event, figure out where we left off, click-click-click to get there, and then pick up where we left off.

And if we are doing multiple events, we should log in and see the active events, a status of where they are, a prioritization of events that need to be addressed, a prioritization of alerts and notifications and requests that need to be addressed, and an appropriate interleaving of activities. And, if we want, an auto-pilot mode that will take us through each and every task in the appropriate order. All we should have to do is focus on the information, the strategy, and make the right decision. That’s what a good system should do.

Cognitive Buying

Not only should the system be capable of automated spot buys, but it should also be capable of executing significant portions of more strategic events, especially if the only reasons the buy is not an auto-buy is because the dollar value is above a threshold or the component is a critical component to a primary product line. In the first case, there’s no reason the system cannot still auto-pilot he event all the way to an award selection. If it can identify the right suppliers, validate the cost models, validate the bids are in-line with the models, validate the supplier information, and validate that a potential award meets the business requirements, then it should be able to auto-pilot the scenario to the point where an award is recommended and all the user should have to do is evaluate the award, verify everything is okay, and approve — at which point the system auto-pilots the remainder of the event. In the second case, there’s also no reason that the system can’t auto-pilot the majority of the event. All a user should have to do is approve the supplier selection (because quality, OTD, and regulatory compliance is key, and they need to have human acceptance of risk factors) and the final award recommendation. Remember, the point is to allow buyers to focus on high-dollar strategic categories, not low-dollar ones where only availability needs to be guaranteed or high-dollar ones of non-strategic importance.

And while even this is not the end of the road … it’s where we were supposed to be now … but it’s but it’s definitely the case that not all of this is in your average S2P platform, or even your average best-of-breed system.

So just where could you be?

Last week we noted that 2020 is fast approaching and asked what you had accomplished, especially given all the 2020 predictions that were made five to ten years ago.

But then we didn’t wait for the answer, we gave it to you because, simply put, most of the vendors haven’t advanced the technology to where it was supposed to be and, as a result, you haven’t really obtained much in the way of new functionality.

We were told that, by 2020, software would be smart. It would do most of our work for us. Tactical would be a thing of the past as process, and, most importantly paperwork, would be completely automated and electronic. But, alas, we’re still in the situation where you not only have to manually build cost models over and over, but even manually populate them. And even then you can’t automate the RFX/Auction process when the best option is, well, completely obvious. Why? Twenty years after the introduction of this technology, why?

So, what can we get?

Continuing on yesterday’s theme, more than we expect if we are willing to go best-of-breed along-side our primary S2P, get good at integration, and better yet at value identification and management. We do have the ability to do the following, with specific platforms in specific circumstances.

Smart Cost Modelling
Here’s the deal, we’re at the point where we have well defined bills-of-materials in an ERP for just about every product ever manufactured. So we know what goes into the product, how much, and key elements of the production process (as well as estimates of energy requirements). We know how much we paid, how much average labour rates are in the locales we produce in, how much energy costs per kWH, and can estimate the remaining overhead and margin.

So, if we know all this, why can’t the system automatically create a cost model, automatically populate it with relevant (and current market) costs, and give you an accurate true should cost model that can be used to accurately identify which suppliers are on the mark and which aren’t. There’s no reason it can’t. The reality is that most systems just don’t.

Automated Spot Buys
Most of our RFIs, Auctions, and even spot-buys are small and/or for non-strategic products and services, many of which are one-time or MRO buys. The vast majority of the time all we need is the lowest cost product or service that meets our need that can be obtained in the required time frame. But still we have to spend hours and hours and hours setting up the RFI, inviting suppliers, verifying the products or services are acceptable, getting bids, comparing them, identifying an award, making the offer, getting a signed contract, and sending the PO.

But if we have approved suppliers, detailed requirements that can be qualified and quantified, gated workflows to make sure only those that meet all the requirements can bid, cost models to verify that costs are in-line with expectations, the ability to capture and verify delivery times, carrier costs, canned terms and conditions, the ability to force a supplier to agree to the conditions up-front, integrated e-signatures, and auto e-Document distribution and management, why can’t the system just run the process, make the award, get the contract, and send the first PO fully automated? There’s no reason. None at all.

And this is not the end of the road … but it’s definitely the case that not all of this is in your average S2P platform.

Last week we noted that 2020 is fast approaching and asked what you had accomplished, especially given all the 2020 predictions that were made five to ten years ago.

But then we didn’t wait for the answer, we gave it to you because, simply put, most of the vendors haven’t advanced the technology to where it was supposed to be and, as a result, you haven’t really obtained much in the way of new functionality.

We were told that, by 2020, software would be smart. It would do most of our work for us. Tactical would be a thing of the past as process, and, most importantly paperwork, would be completely automated and electronic. But, alas, we’re still in the situation where you have to manually define RFX processes, manually identifier suppliers, and manually build cost models. What’s up with that?

So what can we get?

Continuing on yesterday’s theme, more than we expect if we are willing to go best-of-breed along-side our primary S2P, get good at integration, and better yet at value identification and management. We do have the ability to do the following, with specific platforms in specific circumstances.

Automated Supplier Discovery

Why do you have to manually find, identify, qualify, and select suppliers for each and every RFI you do? When the platform should know the suppliers we’re using? The supplies we invited to bid last time? The suppliers that were pre-qualified as viable options? The suppliers that have registered with the products and services you need? The suppliers in the network platform you use? The suppliers your co-opetition have vetted for you? And so on.

Based upon the RFI, your particular product and service requirements, your supplier history, your community history, and so on, your platform should be able to use machine learning and advanced AI techniques to identify the best suppliers for your AI given all of your needs, the supplier’s performance history, the supplier’s current financial and CSR status, and available community data. But most platforms today can’t even suck in a default list of suppliers? Why … it’s NOT that hard.

RFX Process Automation

Most platforms allow you to define templates that define survey questions, basic cost models, and basic gated flows on registration and participation, but that’s about it. When it comes time to starting a new event, it’s up to you to find the closest template, alter the questions, cost models, and gated flows, double check everything, and drive it. But why?

Why can’t the platform identify the right questions based upon the specific products and services? The right gated flows based upon the industries, categories, geographies, and raw materials? The right models based upon the constituent raw materials, development processes, and level of cost detail available? The right starting data from market indexes? Most of this just requires the right rules, and last generation machine learning. Very little tactical effort should be required in an average RFX process, especially if you’ve sourced a similar product before.

And this is not the end of the road … but not all of this is in your average S2P platform.

Last week we noted that 2020 is fast approaching and asked what you had accomplished, especially given all the 2020 predictions that were made between 2008-ish and 2013-ish when pretty much every big consultancy or vendor had finished their own 2020 bandwagon and created the 2020 parade.

But then we didn’t wait for the answer, we gave it to you because, simply put, most of the vendors haven’t advanced the technology to 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.

But we were told that, by 2020, software would be smart. It would do most of our work for us. Tactical would be a thing of the past as process, and, most importantly paperwork, would be completely automated and electronic. But, alas, we’re still in the situation where you need OCR because many smaller suppliers max out at PDF when it comes to tech, and many in non-tech industries think PDF is scanning a paper invoice and embedding the image in a PDF. And if the platform doesn’t have PO flip, made easy, over a low bandwidth connection (there’s still no fibre [or even copper, they dig it up for scrap] in the middle of the backwoods, especially in developing countries), this is what you get.

So what can you get?

The answer is, believe it or not, more than you expect, if you’re willing to go Best-of-Breed along-side your S2P, get good at integration, and better yet at value identification and management. Because we do have the ability to do the following, with specific platforms in specific circumstances.

Invoice Automation
There are a number of platforms out there with intelligent OCR that can use large, global, supplier databases to match names; address databases to identify algorithms (and match to locations); global banking databases (with known swift codes, branches, etc.) to identify banking information; catalog and service databases to identify what is being billed; and smart code match algorithms to identify corresponding POs, contracts, etc. and identify the majority of relevant information. And, on top of that, some of the better applications have document structure identification capability where the application can be told what information is generally where, will learn the supplier’s invoice structure over time, and extract the information with higher and higher precision over time.

Moreover, the intelligent algorithms will only ask for confirmation of data that doesn’t pass a confidence interval, which will increase over time. So, even if you have hundreds of thousands of invoices, within a few months after roll-out, you won’t even have to look at 100 of them. Long gone are the days of EDI only. When you go from having to evaluate over 15% of invoices manually (which is the typical error rate) to less than 0.2% (as the platform can automatically identify, and correct, errors or bounce back invoices that can’t be matched), the time savings is phenomenal

Supplier Identification
Once upon a time, if you needed a supplier, you look at who bid last time, who sent you catalogs, and who your buddy recommended. Not very good. Then you got e-RFX and supplier portals where they could self register, but then you had an avalanche of supplier (catalog) data to go through that made looking for a needle in a haystack easy. Then you had networks which classified by high level category or product classes, but when you had specific needs, custom manufacturing requirements, raw material limitations, etc., it didn’t help much.

But now you have platforms that can use AI technologies and community feedback to identify suppliers that provide the products you are looking for, using the production techniques you are looking for, in the locations you are looking for, with the quality and reliability you are looking for, and the status/certification you are looking for, ranked according to most likely to be selected to least likely. This type of supplier discovery and automatic qualification using 3rd party platforms and community intelligence can often save a person-week of time on an RFI.

And this is just the start … but not all of this is in your average S2P platform.