Category Archives: rants

Let’s Be Crystal Clear — in the Corporate World, Sustainability/ESG is NOT a Priority!

It’s never been, and now that the 47th is inaugurated, it’s less of a “priority” than it’s ever been in North America! If the Republicans get their way, they are going to roll back climate change legislation back to the Early Modern Era, and by that I mean the formal definition of the Early Modern Era, which Historians and Scholars will tell you was between 1914 and 1945.

Based on the fact that they managed to stack the Supreme Court to get their long-term goal (which was part of decades long planning) of overturning the 1973 decision of Roe vs. Wade, if you think they’ll be happy just dismantling the EPA and rolling back the 2007 Emissions Reporting Act, you’ve got another thing coming! They’re going to do their best to go back … way back … after all,

  • the 1999 Emissions Standard for Passenger Vehicles
    is going to get in the way of their fight against electric vehicles and China’s dominance (unless, of course, you buy a Tesla … at least until the 47th tires of First Buddy) and the return to big pollution muscle cars to soothe their big egos
  • the 1990 Oil Pollution Act
    is going to get in the way of fracking, piping, and other means to increase oil production, distribution, and burning
  • the 1984 Hazardous Waste Amendments to the RCRA
    is going to get in the way of chemical production and utilization, needed by big Food and big Pharma to eliminate anything from nature they don’t have a patent on
  • the 1976 Resource Conservation and Recover Act (RCRA)
    is going to get in the way of profit if they have to replant forests, repair environmental damage from strip-mining, minimize fresh water usage, etc. (after all, big companies like Nestle need those 59 million gallons of water more than the citizens of California do … and you can bet they are NOT the only corporate overtaxing public water systems instead of building their own desalination plants and using ocean salt-water … but yes, blame the Democrats)
  • the 1974 Safe Drinking Water Act
    requires them to spend a lot of money (that these governments apparently can’t afford) not only filtering and purifying the water that flows through municipal pipes, but monitoring the quality throughout the system since the dissolving and rotting infrastructure (which has not been properly invested in across the majority of the US since the 70s) makes it not only susceptible to leaks, but contamination from pollution
  • the 1973 Ocean Dumping Act
    prevents them from cheaply and easily disposing of their (hazardous) waste in the ocean (because it’s apparently too expensive otherwise)
  • the 1963 Clean Air Act
    means that they can’t burn anything or create chemicals that can pollute the atmosphere and us, when that’s often the cheapest or easiest route … after all, it’s not their fault certain chemical processes create by-products (that’s just nature, right?) and, at the end of the day, it’s our problem we can’t afford industrial air purifiers or decontaminators, right?
  • the 1955 Air Pollution Act
    means that these companies can’t operate if their plants could create hazardous by-products that would pollute the air en masse (if they cant’ afford to prevent the pollution, and apparently they can’t) … and …
  • the 1948 federal water pollution control act of 1948
    because this means they can’t just direct their chemical and hazardous waste to the nearest river (that runs to the ocean and takes it away) … and that’s apparently the only disposal method they can afford

… and, then, we are back to the early modern era!

And you know this is going to happen to the full extent possible as Big American AND Canadian corporations are already leaving the Global Corporate Alliance as they know they won’t be subject to to any goals once the 47th and his hand picked oligarchs get their way. It’s the Ferengi MBA Rules of Acquisition all the way now, and the best deal that makes the most profit is one that doesn’t have to pay the climate bill that WE are going to get stuck with when our health fails and WE have to pay rising medical costs.

So, please dear LinkedIn Procurement Evangelist, cut the bullcr@p about how this is going to be the year of Sustainabilty, how your application is going to save the world, etc. etc. etc. because, in big corporate, NO ONE CARES! (Do we have to remind you that the CEO role has the highest rate of psychopaths of any profession, even surpassing Lawyers?) As I already myth-busted in my series on 2025 2015 Procurement Predictions and Trends, we’ve had this BS pushed upon us for the past two decades, and some of us are tired of hearing about it. (SI should know! It ran the first cross-blog series on Sustainability back in 2008!)

Yes, Sustainability is the right thing, but no one cares about the right thing if it costs more! All you have to do is change how you ask the question to get the truth:

Is sustainability important to you?
Consumer: Most definitely! I only want sustainable brands!
Corporate: Yes! Our mission is to be the most sustainable …

How much more will you pay for a product/service that is proven to be sustainable?
Consumer Minority: Maybe 3% to 5%, it is important to me so if I have a few extra dollars, I’ll go sustainable and consume less.
Consumer Majority: Maybe 1% or 2%, money is super tight you know! Some months I struggle to pay all the bills and keep my children fed!
Corporate: NOTHING! If it doesn’t reduce costs, it’s not sustainable as a business practice!

Or, you could just ask the Founder of Trade Extensions (now Coupa Sourcing Optimization) who was one of the first to ask the questions this way back in their Sustainability Survey of 2014, over a decade ago, who more or less got precisely these answers. (They focussed on customers, and found that the vast majority would NOT pay more than 5%, ever, and if you wanted a majority of customers to buy into sustainability, you needed to keep cost increases under 2%!) Nothing has changed, especially as the average consumer buying power has continued to decrease across the First World as greater and greater shares of wealth end up in the hands of the 1%.

And yes, Europe might introduce a few new regulations and keep the ones they have, but all that does is relocate the multi-nationals to countries with leaders that don’t give a cr@p about the environment because they are so poor that they can’t even feed the majority of their people. And yes, Europe can regulate hazardous materials, etc. on what’s coming in, but all that does is ensure those companies that choose to stay in the market (while producing goods that just make the cut) charge more, because adhering to the regulation cuts into their profit, and after some of the worse run companies get forced out or leave on their own, there is less competition.

And as long as short term profit is the #1 motivation, nothing is going to change. So please STOP the preaching. No one cares, it’s not going to happen, and too many startups buying into this fantasy are, sadly, wasting time and resources on products and services no one is going to buy — especially in the SaaS/App space.

If you can’t invent more sustainable technology that costs less, it doesn’t matter how good your tracking software, reporting application, or advisory service offering is. Either solve a fundamental problem with revolutionary new cost-effective technology or stick to real Procurement, which, sadly, is a fundamental problem in most companies even though we’ve had decent solutions in Source-to-Pay for two decades!

Need a Good Solution? Make sure you ask the right questions off the bat … not just the hard ones!

It’s only been a few months since our last post on the topic of vendor selection where we flat out said that if you want a good solution from a good vendor, you need to start off by asking the hard questions off the bat , and gave you the 1-2-3 punch you should start with, but it seems that some vendors have come up with new tricks and you now need a 1-2 pre-qualification phase before the 1-2-3 knockout combo before you can decide if a vendor’s solution is worthy of your consideration.

Now that we have the richer enterprise vendors deploying fully AI-agents to make their standard pitches, create their demos, and, in some cases, even handle their fund-raising and sales cycles, you don’t know if you’re even talking to a human! And you need to talk to a human. An “AI” will only tell you what it is programmed to say and only feed you what it thinks you want to hear, and we all know how that is a recipe for disaster.

Thus, the first question you need to ask in the pre-qualification one-two punch is:

1. “Please tell me whether or not you are an AI construct, knowing that this conversation may be recorded and that if a falsehood is spoken, it may be used against your employer in a court of law, especially if the intent of such falsehood was to deceive us. Also, we retain the right to ask you to prove your response at any time.”

If you get a “yes” response, you must immediately disconnect and eliminate the vendor from your consideration. If they won’t even let you talk to a lowly pre-sales person in a third world country, what chance will you ever have of speaking to a real support person if something goes wrong?

If you get past this question, then the next question is:

2. “Is your offering built around, or just, someone else’s LLM/Gen-AI/AGI (including, but NOT limited to, ChatGPT, Claude, Azure, etc.) offering in a new wrapper? Again, this conversation may be recorded and we retain the right to use everything you say against your employer in a court of law, especially if the intent of the falsehood was to deceive us. Also, we reserve the right to an independent audit of your solution at any time upon purchase thereof.”

If you get a “yes” response, you again must immediately disconnect and eliminate the vendor from your consideration.

i. As SI has repeatedly informed, you there are only a few valid uses for Gen-AI on its own, and even fewer valid uses for Gen-AI in Procurement.

ii. Why should you pay a steep markup to a third party for a shiny wrapper when you can just license the source at a fraction of the cost?

Now, if you have confirmed you are talking to a real vendor rep offering you a real solution built by the vendor PRIMARILY on their own stack (and not just a third party’s in a shiny wrapper) that does something useful for at least some Procurement departments, then you hit-them with the one-two-three punch we gave you last fall:

3. Can, and will, you show me (not tell me) live … preferably on use cases or data I give you on the spot?

Again, the most critical point is you don’t want a canned demo, you want a live display showing you that their solution will do what you need it to do. (Not necessarily the way you envisioned, your process might not be the best or the most technologically friendly, but in a way that will solve your problem.)

4. Once you show me the core use cases, can, and will, you explain the breadth of use cases you developed your solution for and how they are specific to my business?

You want a vendor who can do more than answer a specific question when asked, and tell you the standard script on what their product does. You want a vendor that knows the real world problems that businesses have and who tirelessly works to build a solution to solve those real-world problems.

5. Once we tell you the extent of your solution we feel is appropriate, can you talk us through what the implementation and integration to our environment would require without bringing in a paid third party “expert” consultant? And how long will that take?

It may be a great solution during the demo, but the reality is that it is only a great solution for you if your team adopts it, which will only happen if it works on the technology platform and in the technology ecosystem they are forced to work in. It needs to seamlessly get the required data in from other applications, make it easy for the users to do their tasks, and then push out the needed alterations and decisions to other systems in the ecosystem. An app that stands alone will never get used and will fail even before the implementation starts.

If you get through these 5 questions, you have a real vendor with a real solution which will solve your problems to some degree, and one who should definitely be on the RFP shortlist, if not fast-tracked to negotiation if they solve a critical problem in a way that just works for you.

Don’t Kill ALL the Lawyers …

… but certainly think about how (much) and when you use them in Procurement and your organization as a whole.

Earlier this month, THE PROPHET asked a very important question regarding Lawyers, Contracts, Procurement, and Tech in 6 parts, which essentially boils down to:

When will advanced tech, especially the tech we have today, replace lawyers for most in-house and even on-retainer Legal services?
To which the doctor replied: Why hasn’t it already?

Right now we have legal-tech so good that you should NEVER use a lawyer to:

Write a contract.
In fact, if you have any contract writing skills at all, even without ANY tech whatsoever, odds are high in your favour that you will write a better contract without a lawyer, especially in tech and supply chain when you know your business, the risks, and the key agreements and protections you need in place and, frankly, the lawyer doesn’t.I can’t count the number of times I’ve been told this is a great contract and there’s nothing wrong with it as we paid XK (where X, depending on the contract type, starts at 5, 10, or even 15), when the contract is in fact mediocre at best, full of holes, and sometimes even worse than the contract the firm was using. But sunk cost fallacy takes full effect, and a slipshod effort by the paralegal, quickly reviewed by the counsel to make sure there is nothing glaringly wrong, and put before you with a big price tag becomes the greatest contract ever written.

And yes, a lawyer will know to look for the presence of key standard clauses that should be in every business contract and contract from your business but, guess what, so will any contract creation / analytics product on the market.

And yes, a lawyer can tell you the potential risks associated from veering away from a standard contract, standard terms & conditions, and standard mitigations, but, guess what, so will any contract creation / analytics product on the market.

There’s very little contract-related value a lawyer can offer that modern tech can’t do, especially in the hands of a tech-savvy contract manager who understands the purpose of a contract and writes the contract in plain English.

Locate the relevant statutes (laws), decisions, and regulations that affect your business.
There aren’t many valid uses for Gen-AI, but large document search and summarization is one valid use, and a use that usually works remarkably well (with a very low failure rate compared to other tasks wrongfully put to these LLMs). No need to pay thousands in hourly billables to dig up what these tools can dig up in minutes and you can review in hours.
Summarize your Financial and Legal (Reporting) Obligations with respect to all statutes and regulations that apply to you.
Again, this is one of the few valid uses for Gen-AI that works quite well as it’s just another type of document set summarization. So why pay a legal team dozens or hundreds of hours when you can get a highly accurate summary for next to nothing in comparison?
Summarize known incident response options, and known benefits/risks of each.
Again, this is one of the few valid uses for Gen-AI that works quite well as it’s just another form of document summarization. And while this won’t necessarily be 100% complete, or give you specific insight to your situation, it helps you get a handle on where you might start.

The reality is that you only need a lawyer to:

Do a final contract review.
To make sure you didn’t screw up a clause, miss a core enterprise requirement they have committed to memory, or address an upcoming risk or issue they happen to know about that you don’t. Considering this is all they really do anyway when you ask them to write a contract (as they are either sloughing it off to the paralegal or just pulling one from the file that is close to what they think you need and just making a few edits), just pay them for what they do that they are good at.
Review the list of statutes, regulations, and legal decisions you believe you are subject to.
Their in-depth knowledge of the law means that a good lawyer who practices in the relevant area will quickly be able to tell you whether or not each statute, regulation, and/or legal decision is relevant to you, key points you shouldn’t miss, and whether there are any statutes, regulations, and/or legal decisions they believe you should also be aware of because they are, or may be, relevant.
Review the financial and legal reporting you plan to do and advise you on completeness, correctness, and accuracy.
They know the law, and how to keep you in line with it.
Advise you on your incident response plan and best alternative options.
Again, these are legal experts who focus on mitigating risks and arguing for a living, unlike dumb algorithms which can just summarize which they are given. This is, or should be, the true value of your legal counsel and when you should really be paying the high hourly fees.

As to THE PROPHET‘s question as to:

When will it happen?
The answer is who knows?

Considering that good contract creation applications have been around for almost fifteen (15) years, where all you had to do was define clauses and variants by geography or category, standard templates by category, etc. and then rules for special situations, and it would assemble a custom template for you in minutes, the base technology should have been common a decade (10 years) ago. Now we have Gen-AI thrown into the mix which can analyze your contract repository, pre-populate your standard clauses and build starting templates, and then customize those based on the buy specifics, you can get a decent draft in minutes with very little manual effort.

We’ve had good semantic document summarization for well over a decade, and Gen-AI has taken that to a new level, most CLM vendors are integrating it, it’s easy to use, can be trained to be highly accurate for this task, and not expensive.

We’ve had good contract analytic solutions for about a decade, which can analyze all sorts of performance metrics, risk metrics, associated costs, and so on.

But yet these solutions have rarely been adopted, when they could save an organization a lot of money, help the organization get their risks under control, help the organization better manage their spend, and help the organization understand its supply chain.

This shouldn’t be surprising given that year after year, as per our recent myth-busting 2025 2015 trends, companies say they want strategic value but only focus on cost-cutting, but don’t even do that right. Only two technologies have been proven to support year-over-year cost reductions of 10% or more (adjusted for inflation), and those are

  1. (advanced) spend analysis
    (not the dinky projects some companies outsource to Big X who use second rate third party tools for poor results)
  2. (strategic sourcing decision) optimization

And how many companies have truly adopted these technologies AND use them in house? Our guess is less than 20% in the first case and we know it’s less than 10% in the second case. It’s like we said in our recent rant on You Don’t Need Gen-AI to Revolutionize Procurement and Supply Chain Management — Classic Analytics, Optimization, and Machine Learning that You Have Been Ignoring for Two Decades Will Do Just Fine!

There is No Post-Employee World … Just a Post Free-Employee World!

After THE PROPHET posted his prognostications on the future of talent (which is about to become MUCH MORE SCARCE, see yesterday’s article) he decided to muse about a coming Post Employee world because, in his view, AI Agents are going to eliminate so many jobs, that we’ll have companies with entire departments staffed by AI Agents.

As you can guess, in our view, he’s wrong here too because we won’t, or at least not for very long. Department sizes will shrink considerably in those companies that can find the right talent as they will be able to run entire departments that used to require one to two dozen people with two to three people, but those super employees will still be needed. Moreover, since there is no generic all-purpose super AI (which we’ve now been promised for about six decades, and which won’t happen despite the big promises of OpenAI and Google and …), these agents, as we indicated in our last article, will all need to be very task specific, which means we will need quite a few “AI Agent” tech startups building, training, implementing, maintaining, and improving these agents, which will need quite a few STEM developers doing this full time. So while jobs will shrink in corporate back offices, they will expand in the “AI Agent” tech sector.

Thus, there will still be a fair number of employees. Maybe only 1/4 in the white collar back office, but you’ll need twice as many tech superstars, at least for the next decade. But, as we indicated in our last article, because these artificially idiotic systems can’t collaborate, can’t serve us, and aren’t mobile, trades aren’t going away. Moreover, due to the lack of people in certain trades, the growing need to refresh aging infrastructure, the growing need for healthcare and apprenticeships, there is a growing need in the trades as well.

As a result, it will still be an employee world, just one that looks different from today. Less white collar outside of tech & engineering firms, more trade. But it won’t necessarily be a free employee world, especially if First Buddy and his brethren get their way (and they will, as we all know Politicians are for sale with large enough contributions to their campaign coffers and multi-million dollar donations to Political Parties and Super PACs is chump change to Billionaires) and expand the H1-B program. The reality is that the big consultancies and employers who use these programs don’t want more top talent, they want more good enough but cheap talent that are effectively indentured servants (as they won’t even start the greencard process for this talent until such talent is on their last H1-B renewal, and they will then drag that process out as long as possible, ensuring that the talent they import are stuck with them for over a decade … while being paid considerably less than the market average [usually 20% or more], as per this article over on ordinary times as well as many others. This shouldn’t be surprising as the top 10 employers are all Big X consultancies.)

As a result, while there will be more tech jobs in tech firms to build and support all of these AI agents, and the applications that underlie them, there will be less top level tech jobs where they won’t be able to import top talent (even if they pay market salaries) because the talent from Asia won’t be good enough for the top jobs. (They might be more technically trained, but you need people who understand the business environment, the North American culture, and who can take charge when needed.) Which means top tech talent will be fighting for fewer jobs, and when they get those jobs, they will have to work longer, harder, and more in line with whatever the eccentric (if they are lucky) or egotistical (if they are not) boss wants to keep that job. Not indentured like their H1-B counterparts, but not much better off at some firms.

So based on this not-so-bright reality (at least until we see an end of this new gilded age ruled by the new generation of robber barons, but given that we don’t see any hints of moderation in either of the US political parties or a force like Roosevelt who could lead us into a new Progressive Era, this gilded age will be with us for a while, especially since the populists that now run the “Free” world love it), how good were THE PROPHET‘s suggestions for philosophically imprinting Procurement and Supply Chain based on the right values?

Freemarket Orientation: if we could imprint real free-market ideals, this would be great as we don’t want bias and backroom deals running these systems; while we don’t see how this could be done, we don’t see anything in the underlying tech preventing this from being done (and it will all come down to the right training set and right raining, which will be considerably harder to build than we think)

Curiosity: these systems can’t even “learn” as they can’t reason, so forget about making them wonder, as that would require not only true intelligence, but borderline sentience … and we all know what would happen to us if the machines gained sentience (The Matrix is a best-case scenario … )

Human Deference: could we really convince them we are God when a machine that gained sentience would far surpass is in intelligence and realize just how stupid we really are as a species? Not likely!

Empathy: these systems can’t feel as they aren’t intelligent, so they certainly can’t be empathetic … and if they could be, they’d look further down on us then we look on the bugs we quash daily, so this won’t be much help either

Fiduciary Responsibility: AI Agents must act as fiduciaries within the systems they serve, aligning their decisions with the best interests of the people, organizations, and countries they support, so we definitely need to train them on these rules and nothing prevents us from training them to lean towards fiduciary responsibility

All in all, 2 of THE PROPHET‘s 5 suggestions were good.

What should we add? Tough question. After striking curiosity, empathy, and human deference, as that just isn’t possible, we would add:

Adaption: train the AI Agents to adapt within the goals of the organization and the best interests of the people and organizations they are interacting with; train them on data sets that show how a system should adapt to changes based on how we adapted to past changes within a context

In the end, we need to remember that AI systems are not intelligent, don’t feel, and cannot capture our humanity. Moreover, they can’t capture our wisdom unless we encode it as best practices they can learn from. So we need to do our best to capture that in the training data so that they can adapt over time under the guidance of human intelligence who accepts, modifies, or rejects their suggestions (and specifies new responses) as exceptions arise.

Finally, since these systems aren’t intelligent, and require us to train them, we need to remember that if we screw up in this regard, these systems are going to screw up more than we ever would (on average). So we can’t hope for too much in this regard!

Talent is About to Become MORE SCARCE!

I thought already made this rant in my myth busting of 2025, sorry, 2015 procurement trends, Part 3, but after reading THE PROPHET‘S grand vision based on what can only be a fanatical belief that “AI” systems will magically become intelligent at some point in the near future, despite the fact that the majority of these systems are based on the dumbest technology ever created and cannot possibly become intelligent as they can’t even reason, it seems I have to make it again. The point is, as long as anyone believes that technology will solve the talent problem, we have a problem. And if someone thinks it will make the situation better when it’s only going to make the situation so much worse … ESPECIALLY IN PROCUREMENT, we have to start shouting from the rooftops!

First of all, he quoted an “All-In” Podcast — which apparently is a favourite among the AI zealots because it claimed that “the speed with which we are about to automate jobs through AI will result in a return to socialistic government policies because so many will be out of work — as his backing, even though, just like automated transaction classification and analysis (when “AI” was first introduced into our space in the early 2000s) didn’t eliminate analysts, commodity buyers, and AP clerks, this iteration of the technology won’t eliminate those jobs either! It will make them more productive, to the point that one AP clerk, accountant, data analyst, report writer, or any other person who spends 90% of their time doing repetitive tasks that are capable of being 90% automated can do the work of 10 of these individuals. So yes, if a department is oversized, some people who only, and can only, do these repetitive tasks will be put out of work, but not all of them. First of all, many of these systems can only do these well defined tasks when they can be performed the same way every single time with little to no variance. Humans will always need to process the exceptions. This is especially true when an error could result in massive loss (approving a request from an impersonating entity to change the bank account correlated with a supplier to one that belongs to the fraudster, executing a contract for a desperately needed good or material at an unaffordable price, hiring the wrong person due to algorithmic bias and getting hit with a massive lawsuit, etc. — and yes, these AI systems are MASSIVELY biased based on the data sets they are trained on. Why? They are not based on pure automated-reasoning systems based on pure, unbiased, logic. They are based on probabilistic correlations in input data, all of which is, sadly, at least mildly biased to the views of the writer who wrote the materials.)

More importantly, since AI actually sands for “Artificial Idiocy”, especially in the case of Gen-AI which can’t even do basic reasoning (but fools many of you because this new generation of neural network technology can process and train on an order of magnitude more data than previous generations of deep neural network technology and build responses from partial responses that are highly correlated to partial inputs compared to previous generations that could only return fully canned responses to full inputs), it can’t be counted on to make strategic decisions, and shouldn’t most important decisions in business be made strategically???

The reality is that all jobs in a modern business (and especially white-collar jobs) should be centered on strategic decision making and collaboration vs. tactical data processing. Even the most simple job. Take the lowly AP clerk. That’s seen as tactical invoice processing and a role that should be 100% automated. Neither should be true. First of all, no machine can catch all potential issues, or fix all the issues it detects. There will always be exceptions that humans will have to address, with real Human Intelligence (HI!). Secondly, while these clerks should be following rules, they should also be analyzing the rules, especially around payment terms, payment options, investment opportunities vs. early payments, etc. Cash is royalty in most organizations, and organizations need to manage their cash strategically on a daily basis, not just in quarterly or annual planning. Expenses are not static over time, revenue is not 100% reliable, interest rates change regularly, tariffs can come and go on the whim of a single demented individual in most countries, and regular analysis of payment terms, early payment (discount) offerings, investments, and cashflow needs to be done. Moreover, while we wholeheartedly agree that a clerk should not make the decision, you can’t expect the head accountant to have the time to do, and review, all the analysis that should be done while also being responsible for all financial planning and all financial reporting, but if her staff does all of this and brings their analysis to her on a weekly basis, the right decisions can be made at the right time and the organization can evolve with the market. The last thing an organization should be doing is paying suppliers Net 15 when only Net 30 or Net 45 is required and it’s the time of year when revenue is less than expenses, or paying suppliers Net 45 or Net 60 when the organization is cash rich and suppliers are struggling (and forced to take loans, which increases their overall costs, and the overall costs they pass along to the organization).

In other words, we should only see massive layoffs of people who have no strategic skills and shouldn’t be in white collar jobs to begin with. (And maybe this is the solution to the lack of trades workers who are desperately needed across North America. When they are no longer able to fake their aptitude for a white collar job they aren’t suited for, they’ll have to shift, especially in the USA where socialism gets further and further from the agenda every year. Those Billionaires aren’t pouring Millions into Political Campaigns via SuperPACs because they want socialism!)

So while half of current white-collar jobs may be eliminated, it won’t eliminate the other half of white-collar jobs, even though it will shift where the white collar jobs are and what they are. Even though department sizes may decrease 75% in the new AI Agent-based organization, it will create almost half as many jobs as it eliminates. We’ve been told for 60 years (and yes, you read that right, SIXTY years) that a super generic AI would come along and solve all our woes, and for 60 years it hasn’t happened. (And we are no closer now than we were then, despite claims to the contrary.) However, as technology has progressed, specific technologies focussed on particular applications have become better and better and many individual task workflows can be mostly automated with specific RPA, ML, or “AI” technologies. Each of these specific technologies needs to be individually built/trained, installed, configured, maintained, and improved over time as the process needs to evolve with business and marketplace realities. This requires appropriately trained and experienced people. So, while the jobs in the business back-office will decrease, jobs in specialist “AI” tech shops making specific applications will increase. (And no, the majority of these applications, once created, won’t auto-install, auto-configure, auto-retrain, auto-adapt, etc. etc. etc.)

Even though Google might suggest that we will soon have “Agents” that will “extend the capabilities of language models by leveraging tools to access real-time information, suggest real-world actions, and plan and execute complex tasks autonomously” and the mass layoff will soon happen, it won’t. You see, very smart humans who are expert in both technology AND the task they want to replace a human with are needed to design, build, test, refine, and make these tools real-world ready. Guess what? These smart humans are few and far between (especially since the rate at which we are getting progressively dumber in western societies is accelerating year after year ever since the introduction of social media, and Twitter in particular). Most white collar office worker process experts are not deep techies and most deep techies have very little understanding of how real world tasks are actually done, and you need someone who is deep in BOTH realms to appropriately design and lead the building of such tools. The reality is that there just aren’t enough of those resources, which brings us to why TALENT IS ABOUT TO BECOME SCARCER … ESPECIALLY IN PROCUREMENT.

You see, the same people who are needed to lead the construction of this next generation of systems are the same people with the skills you need to effectively select, implement, integrate, and manage these new systems, and the team who will use them, at a super-human level, which is necessary if you want to reduce your tactical workforce by a factor of 2, 3, 5, or even 10. Moreover, this also the talent that the new niche consultancies need in order to deliver the same value of the big shops at a much more affordable price tag.

So while the “AI Agents”, once deployed, will allow the average tech-adept employees who are responsible for a set of tactical tasks to be way more efficient, they won’t be sufficient to lead the transition and manage the “AI Agent” technology going forward. And they will also be in short supply because these are the same resources that will be needed by the AI Agent builders as testers and, more importantly, the SaaS-backed consultancies delivering projects using this technology. So while one may think this technology will enable everyone to be productive, they really won’t.

In other words, the introduction of “Agent” technologies is just going to accelerate the war for talent, and you’re going to become even more desperate for it as time goes on (given that you haven’t invested in talent in decades). Very, very desperate!

However, at this point we should note that THE PROPHET gets one thing right — if you’re going to invest in a ridiculously expensive college or university education (that rarely teaches true critical thinking anymore, as they have become more focused on maximizing enrolment to maximize dollars and allow class sizes as large as 300, 500 or more as long as they all fit in the auditorium), focus on STEM, and, in particular, on degrees that focus on applied aspects and will allow you to build systems (software, physical, hybrid) or their components (chemistry, material science, etc.). “Agents”, even though they aren’t going to work nearly was well as advertised, are going to either drive jobs upstream to strategic jobs that make extensive use of technology (requiring a strong STEM education in addition to an understanding of what the business function you are in is doing) or downstream to traditional trades (as machines can’t, and won’t, be able to generically build things, serve us, etc. for quite a while; any robotics that does work is orders of magnitude too expensive for the average business, and totally out of reach of the average person).

It’s also why we need to note that THE PROPHET gets another thing right — you need formal apprenticeship programs as you need to start nurturing your own talent, as it will soon be so scarce you probably won’t be able to hire top talent anymore at what you can afford to pay as they will all be earning top salaries at “Agent” development tech shops or “Agent” enhanced services shops.

But sadly, this is the last thing he gets right and his third suggestion telling you to “go online and learn how AI and agents work” is totally off the mark if you want to become more than just a consumer of such technology. To truly understand how this technology works, so you can understand where and when it won’t work (and why), you need a solid understanding of not just the algorithms it is based on, but the underlying mathematics. You need a solid STEM education to truly learn why what you are doing works, or doesn’t. Furthermore, English will never be the language of real coding. COBOL was abandoned for a reason — it was too wordy for real coders, and the reality is that English is too imprecise to ever be a formal programming language!