Author Archives: thedoctor

A Shiny New SaaS or AI Wrapper Doesn’t Make Tech Any Better

Just like painting a hammer bright shiny pink doesn’t change it’s fundamental function, putting a new shiny SaaS wrapper on a traditional desktop application or adding a Gen-AI interface to allow for a “conversational” interaction doesn’t fundamentally change what the application can do.

What an application can do depends upon the data model it can support, the core algorithms that process that data, and the workflows that connect them together to take raw inputs and produce necessary outputs. If the data model is not sufficient, the algorithms not appropriate, and the workflow lacking, a shiny new wrapper won’t change anything … the software will be no more effective than the software that is being replaced.

Pick any significant application, and the best results usually depend on intense or complex calculations, using a proper algorithm that works on a proper model populated by the right inputs, and if any piece is missing, the solution doesn’t work. In our area, it’s Source to Pay, and that starts with sourcing. In sourcing, the right decision is that which results not in the lowest bid, but the lowest lifecycle cost of the purchase, which takes into account not just unit costs, and not just shipping and tariffs and interim warehousing costs for landed costs, but also utilization/waste costs, local warehousing and inventory costs, (amortized) service costs, disposal costs, and even carbon costs if they vary by option. It considers all of the available product/SKU options, plants, shipping routes, and localized plant/warehouse/store needs and uses optimization and analytics to identify the optimal award that minimizes the overall cost while maintaining service levels and minimizing risk. If the solution doesn’t allow you to build the right models, collect all the options, identify the plants and routes, and determine optimal mixes that meet your criteria, then it’s not a modern sourcing solution no matter how SaaSy it is, how new it is, or how much BS Gen-AI gets shoved into it. A good application solves your core problem. If it doesn’t do that, it’s not good. And at the end of the day, it doesn’t matter how slick and SaaSy it is, because if the only application that gets it right is a green screen desktop application, then that is the best solution to your problem. (We hope it’s not — but given how little there is behind many of these SaaS apps, which are built to look good by developers with little to no knowledge of the domain they think they can satisfy with simple algorithms, and sometimes just fancy interfaces to a classic desktop application wrapped in a web container which slaps on a web-friendly API interface to the classic app and classic algorithm — we can’t say it’s not going to be the case that you have to keep using that decades old green screen application.)

At the end of the day, it’s algorithms that work, and the reality is that these are often the algorithms that were developed decades ago by leading minds, stress tested and sharpened by brilliant minds, proven to work, and just waiting for the computing power to catch up to where they need it in order to shine. (The best data structures and algorithms text book ever written is over 35 years old. Most of the revolutionary developments were between the 70s and 90s.) MILP is decades old, but we really didn’t have the computing power to solve large, complex, real world models until about two decades ago (and then only if you didn’t mind waiting a few hours to a few days for a scenario to solve). But now we can solve them in minutes, if not seconds, and that allows for next-generation strategic analysis and planning, as long as you have a modern platform that uses a modern algorithm that can take advantage of multi-core cloud processing capabilities, the right data model, and the data inputs you need.

And therein lies the hitch — it all comes down to the data model, algorithm, and application design — not the UX, the intake and orchestration, or the “conversational” Gen-AI interface.

Remember this the next time someone tries to sell you a shiny new interface or an upgrade to what you have. Remember that most upgrades are because software stacks change, functionality that should have been in the last release is finally added (since many SaaS companies now release untested alphas), or major security or performance issues are resolved. Now, you need the fixes for sure, but you shouldn’t be paying any more than the maintenance fee for those. If the buyer rolls them in “functionality updates”, you should insist you get those for free. If you got buy without the missing functionality (either because you had complementary systems or added it yourself), then do you really need more untested functionality now?

And at the end of the day, the primary reason software stacks change is that if they didn’t, you’d have to buy a lot less tech, and then the investors wouldn’t make money. Not all tech stacks offer significant improvements in functionality or even security. They just allow developers to work on the new hotness and enterprises to force you into spending more money, without any guarantee of more value in what you’re delivered.

So don’t get fooled by new tech. Do your homework. Sometimes the best tech is the old busted hotness.

P.S. Yes, Joel the number 666 is ruining Procurement*, but not necessarily, or just, in the way you appear to believe it is.

* see the Mega Map

Chief Sustainability Officer: USA Edition

A version of the graphic below has been making the rounds on LinkedIn for a few months (and the doctor wishes he could point to the original source of this [on LinkedIn], but either Google mis-indexed it [as the link goes to a user’s profile page] or it’s gone), and a more recent version can be found in this post.

These are great … if you are based in the EU. However, they are not so great if you are based in the USA, as outlined in our first quarter post on how in the corporate world, sustainability/ESG is NOT a priority. So, the doctor decided to correct it for you if you are based in the USA. Enjoy!

Despite Attempts to Simplify It, There Are MANY Categories of ProcureTech Solutions

When selecting a ProcureTech Solution, you have all the following buckets:

Function X Classic Type X SaaS Category X Integration
Sourcing
SXM
CLM
Analytics
e-Procurement Best-of-Breed Standalone App
(full function)		 Suite EcoSystem
Invoice-to-Pay Mini-Suite Lightweight App
(task specific)		 I2O Ecosystem(s)
ESG/Sustainability Suite Bolt-On
(extends a module)		 Open API
GRC
Category/Cost Intel
Niche (Legal, Marketing,
Hospitality, SaaS/Tech, etc.)
I2O

And if you do the multiplication, that’s 297 combinations … and that’s just the tip of the iceberg when there are 10 core areas of SXM, multiple niche areas being addressed (some classic solutions were just for print/telco), multiple buckets of risk management solution, generic and scope-3 specific sustainability solutions, different approaches to intake-to-orchestrate, and that’s just addressing the functional areas of Source-to-Pay+.

Then you have the situation where some vendors only offer a single best of breed (BoB) module, others offer a mini-suite, and others still offer a mega-suite with all of the core modules and often a half dozen more on top of that.

While most are SaaS apps these days, they vary from heavy standalone apps that implement full functions to lightweight apps designed for specific tasks (that are usually missing from larger standalone apps that purport to completely cover a function but don’t) to bolt-ons that offer advanced functionality, but require a core module to work on top of.

One also has to consider how you integrate them into a comprehensive workflow that supports Source-to-Pay+. Sometime modules integrate into one-or-more suite ecosystems out of the box (like the SAP Store or The Coupa Store), other times they just come with a (semi) open API, and now some, not built for integration, are integrating into one or more of the new orchestration ecosystems.

And while functionality should come first, you have to consider all of these other factors as well because if you select a suite for a module, you’re probably locking yourself into the other modules you need as those the suite offers due to cost and integration cost considerations, if you select light-weight or bolt-on apps, then you better have something to integrate them into, and you better be sure the ecosystem has all of the modules you will need to implement over the next five years or so before locking yourself into an ecosystem.

So even though THE REVELATOR believes that everything is going to be a bolt-on or an app and that’s all your going to have to worry about, unfortunately the ProcureTech world is NOT going to make it that simple. Overlooking traditional category and integration can completely destroy the value you require if you can’t easily integrate with complementary modules/apps (and especially if you are in a [primarily] direct industry and need to integrate with supply chain applications for the data you need to make good supply chain aware decisions).

However, it will be interesting to see the primary solution category, breadth, and integration of ProcureTech Solutions (by, and independent of, function) in the future.

Gen-AI is Bad for Consulting Firms … But Even Worse For You When the Consulting Firms Blindly Use It!

A recent post on LinkedIn noted how there’s a wave of AI products flooding the consultancy and advisory space and how they are, frankly mediocre, overpriced wrappers on public models with minimum innovation, if any.

This is sad, but true, and it’s not the worst of it. The worst of it is that some of the Big X firms are training tens of thousands of consultants and f6ckw@ds on these tools to generate hundred page pitch decks and three hundred page strategy and implementation guides of standard generic, meaningless, drivel to deliver to you as “highly tailored guidance and expertise from their leading partners with 20 years experience delivering high-value projects” and charge you tens of thousands of dollars for the privilege.

This is especially egregious when you can use free/cheap (and I’m talking put it on your personal credit card cheap because you won’t notice the fee that is less than your monthly coffee charge from the coffee shop) to build the exact same pitches, strategy, and implementation guides from the thousands of freely available documents on the web in a few hours with a few generic prompts over a Sunday morning coffee. (And then, when the coffee kicks in, realize it’s all a load of cr@p and put in the bit bucket, but at least you will know what a load of cr@p looks like in pitch deck, strategy guide, and implementation plan form and will recognize it the next time an overpriced Big X tries to sell it to you for a ridiculous price tag and will have learned something from the exercise.)

Now that there are companies selling overpriced “custom” products to these consultancies, the situation is only getting worse, especially when the “customization” is just a wrapper with some pre-engineered prompts that aren’t well tested, only work at a point in time, don’t really give the consultancies what they need, and sometimes translate mediocre inputs to inputs that are even worse. Moreover, when you consider the price is sometimes a 100X multiple on the products they build on top of, it’s disgusting. Consultancies are paying more for less, and, in return, you are paying even more for even less!

Which makes no sense when the current publicly available LLM tech is being offered cheap (to try and hook you on it, even though, as we’ve repeatedly explained, the tech is not ready for prime time and will never deliver more than a fraction of what they are promising), and new implementations will get a lot cheaper. Just look at how DeepSeek undercuts the cost by a factor of 100 and gets 90% of ChatGPT (as long as you don’t mind exposing all of your secrets to the CCP). LLMs are nothing more than a fancy next-gen “deep learning” Neural Networks that construct responses vs. serving up canned responses (which is why hallucinations and lies are a core function, not an error that can be trained out) which gets us closer (but no cigar) to decent natural language processing (NLP) for the express purpose of the generation of desired outputs from inputs, but not there (and now, in addition to all the false positives and false negatives, we had to deal with, we now get to deal with hallucinations and lies as well). It’s not secret magic, it’s layers and layers of interconnected statistics and probabilities that no human can understand, in rather standard models that any Theoretical CS and Applied Math PhDs can build, and implementations that are better and cheaper are going to keep appearing as time goes on.

This means three things to any consultancy thinking about using these custom “AI” solutions

  • you still have to be even more tech savvy to use them to any degree of effectiveness
  • it’s not “the art of the prompt“, it’s the art of the training (even though they don’t really learn because they are NOT intelligent) because that determines the maximum level of effectiveness you will ever reach with them (and you need to provide them with sufficient correct data, which needs to be in the high gigabytes at a minimum, and, preferably, in the petabytes)
  • you don’t have to worry about when they are right (enough), which will happen between 90% and 95% of the time with proper training and proper prompting, or when they are obviously wrong, which will happen a very low percentage of the time (say 5% to 9%), but when they are oh so wrong but the response is constructed in a way that is oh so convincing that an above average person in intellect and experience wouldn’t know otherwise (that danger zone between obviously wrong and good enough that is likely only 1% to 2% of the time).

Now remember that your consultants aren’t that tech savvy, and you should know right off the bat incorporating and using these is going to be difficult and time consuming. (There’s a reason we are constantly advising you to be very careful about using Big X for tech selection and tech projects, and that’s because, even though they say it is, it’s NOT their forte. They weren’t built on tech, and they don’t have the best talent in tech — that talent goes to the big tech companies who can offer the 500K salaries to leading devs or the wild-west startups that leading devs think are cool.)

You only have so much clean and complete data you can use for training. You can’t just throw in the 1000s of decks you’ve built as you can’t share work you’ve explicitly created and sold to past clients, and the AI won’t anonymize the decks and suggestions (even though you think it will). It won’t know that “Ford” is the name of your client and might think that “Ford Data” is another term for shallow data and copy sections from that custom strategy straight into your pitch deck for General Motors (and chances are your overworked junior consultant won’t catch it when skimming that 200 page deck with only 2 hours to go before the meeting). And we know what happens then … (and it ends with the consultancy not keeping either client).

It will take a lot of analysis to identify those 1% to 2% of cases where it is very, very wrong but so convincingly right that you will miss some. What happens when you do and give your client advice that explodes in their faces? (We’ll let you answer that one.)

And for you as a consumer, if your consultancy is using this Bogus AI tech, it means that:

  • the situation that results from solution delivered might be even worse than the situation you started with (as should be evidenced not just by the tech project failure rate that is approaching 92% but the fact that 42% of projects are being abandoned during implementation!)

A solution designed by Gen-AI is not a solution. A real solution is a solution designed by human intelligence that uses real, augmented intelligence, to research and validate that solution. Remember that if you are going to hire a consultant!