Streamline & Scale Newsletter: Issue 14: Podcast

If someone asked you right now, what does your company spend on thinking? Could you actually answer them?

I mean, most leadership teams would just completely freeze at that question.

Right.

Because, you know, you might point to payroll or maybe, the R&D budget, but compiling a single accurate number for the sheer cognitive effort driving a business, it's nearly impossible for most organizations today.

Because thinking isn't just a, you know, a neat little line item on a P&L statement.

No, it's everywhere.

So today, we are cracking open volume two, issue fourteen of Red Hawk Technology's monthly newsletter.

Streamline and Scale, and we're gonna explore this exact problem.

The mission for this deep dive is basically turning the invisible mechanics of how a business operates into a visible, optimizable machine.

Because the central argument presented in this issue, and it's introduced by Matt right at the top, is that cognitive labor in a modern business has completely splintered.

Yeah. If you don't know what you're currently spending to, solve problems, [00:01:00] analyze data, make decisions, you certainly cannot optimize it for the future.

The main articles in this issue are penned by Ryan Frederick, who is a Red Hawk's principal of AI and business advisory.

Mm-hmm.

And he introduces a concept that I found honestly just incredibly clarifying. He says, "The cost of thinking isn't one budget; it's three entirely separate budgets that almost never sit in the same department."

Right. They're completely siloed. So the first of those is the human budget.

Which is the obvious one.

Exactly, the obvious one. This is the traditional domain of HR and department heads. It encompasses salaries, benefits, training, essentially what you pay for human judgment, creativity and daily cognitive labor.

Mm-hmm.

Yeah.

But then we have the software budget-

which is usually guarded by IT.

Yeah.

And this covers the tools we use to automate analysis and reporting, the ERP systems, the data visualization dashboards, the project management suites.

And the tension historically has always been between those two, right?

Yeah.

Software was originally [00:02:00] brought in to reduce the burden on the human budget, but now we have this third rapidly expanding category, which is the AI budget.

And I think we really need to define the boundaries of that AI budget 'cause it feels like a lot of companies just lump AI under general software. Like, oh, it's just another tech tool.

Yeah, and lumping them together is a major mistake, primarily because the cost structure is just fundamentally different.

I mean, traditional software is usually a fixed subscription, right- or a licensing fee.

You just pay for access, and you're done.

Exactly. You pay for access. AI, however, introduces the cost of inference.

Inference being the actual compute power required every single time the AI model is asked to process data and generate a response, right?

Yes.

So it's like paying for the electricity to run a digital brain.

That is a highly accurate way to frame it, actually. Every single time an AI agent summarizes a document, writes a piece of code, or, analyzes a customer sentiment, it is drawing on compute [00:03:00] resources.

Wow!

So the more complex the thinking, the more inference it requires, and the higher the variable cost.

I have a theory about how these three budgets, human, software, and AI interact out in the wild.

Okay, let's hear it.

Well, if they are siloed under different departments, they must be cannibalizing each other.

Oh, yeah.

Like you don't realize you're spending a hundred and fifty bucks a month on five different streaming services, a premium news site, and like a cloud storage upgrade- Uh-huh. -until you finally sit down and put them all in one spreadsheet.

Yes. And the business equivalent of that subscription bloat is what the industry calls shadow IT.

Right. Meaning software or tech tools used by employees without the official approval or even the oversight of the IT department.

Precisely. When budgets are fragmented, visibility just vanishes. So you have a marketing team expensing premium ChatGPT accounts on a corporate card to write copy, while the sales team secretly buys an AI outreach tool out of their discretionary budget.

Wow!

IT [00:04:00] doesn't see it, HR doesn't account for it, and the leadership team has absolutely no idea what the company's true thinking spend actually is.

Which just completely sabotages any attempt to measure the return on investment for AI. I mean, you can't claim an AI tool saves you money if you don't even have a baseline for what the total cognitive load was costing you in the first place.

Exactly. Putting these numbers side by side in one unified cost of thinking spreadsheet, it forces a really vital conversation. You have to look at the columns and determine where an extra dollar yields the highest return.

Right.

Is it hiring another senior analyst? Is it buying a better data analytics platform?

Or is it deploying a custom AI agent?

Okay, I gotta play devil's advocate here for a second, though.

Sure.

Because whenever a company starts measuring human judgment against software and AI, it triggers immediate anxiety. I mean, it sounds like a thinly veiled efficiency hunt. Like we're just looking for human jobs to eliminate to satisfy the bottom line.

Oh, yeah, and that anxiety is a completely [00:05:00] natural reaction, but it really stems from a misunderstanding of how high-level human talent is currently being, well, wasted.

How so?

Well, Ryan's article specifically challenges leaders to ask if they are paying senior people to do work that software could handle just as well.

So we aren't talking about replacing the senior analyst. We're talking about the fact that the senior analyst is spending, like, four hours a day formatting data in Excel before they can actually do the high-value strategic thinking we hired them for.

That's the mechanical reality. Companies routinely pay a premium for high-level judgment, but they force their absolute best talent to perform low-level mechanical processing.

Which is crazy when you think about it.

By identifying the total cost of thinking, the goal is to shift the human budget away from rote execution and entirely towards strategic, creative, and complex problem-solving. So it is an optimization of human capital, not an elimination of it.

I love that, and that makes the barrier to entry surprisingly low, doesn't it?

You don't [00:06:00] need a multi-million dollar consulting firm to initiate this. You just need a spreadsheet and a management team willing to be, you know, honest about where the cognitive labor is actually happening.

Yeah, exactly. But once that baseline is established, the conversation naturally has to shift because you know what you are spending on thinking.

The next, and honestly infinitely more complex question is where and how that thinking operates within the company's infrastructure.

Right. And this is where we transition into the concept of the digital business twin, which I think is arguably the most fascinating part of this entire newsletter.

Oh, absolutely.

And it borrows a methodology directly from heavy engineering, doesn't it?

It does. In aerospace manufacturing, for instance, engineers do not simply build a physical jet engine and just hope it performs under pressure.

That would be a very bad idea.

Very bad. Instead, they construct a virtual replica, a digital twin that runs parallel to the real one.

Right. It's a simulated environment. They can feed it real-world data, test how it handles extreme temperatures, or see how a new part [00:07:00] affects fuel efficiency, all without risking a physical prototype.

Exactly. And applying that strict engineering discipline to a corporate structure is what Ryan Frederick proposes.

A digital business twin is a live functional model of how an organization actually operates.

Okay, let's dissect the mechanics of that because I can already hear the collective groan of middle management everywhere.

I'm sure.

Because when you say a digital map of how the business operates, it sounds dangerously close to a glorified corporate wiki or, you know, one of those static fifty-slide PowerPoints that maps out a workflow, gets presented once at an all-hands meeting, and goes completely stale by Friday.

Yeah. But a passive wiki relies on humans to update it, which is exactly why it always fails. A true digital business twin is an active integrated system. And to understand why this is different, you really have to look at Ryan Frederick's specific background.

Right.

He isn't approaching this purely as a management consultant.

His roots are in software architecture. He's an active angel investor and [00:08:00] author, and he has a deep background in workforce development, specifically through i.stars, which is a nonprofit he founded to train adults in digital skills.

Okay, so he has a totally different lens on this. He views a business like a complex code base.

You map it by connecting the actual systems.

Yes. The construction of a digital twin involves hooking directly into the company's infrastructure. It utilizes process mining, pulling log data from CRMs, ERPs, internal communication tools to build a real-time topology of data flows.

So it's seeing everything.

Everything. It maps where information enters the system, where a human has to make a decision, and where bottlenecks consistently occur.

Wait, monitoring API calls and tracing data flows?

Mm-hmm.

That sounds incredibly invasive. Are employees gonna feel like they're operating in a surveillance state?

That's a great question, and it is crucial to separate mapping the process from monitoring the person.

Okay.

The system tracks the journey of a purchase order, for example. It notes that a purchase order sat in a queue for [00:09:00] three days before moving to the next stage. It is diagnosing the structural friction in the workflow.

Okay, that makes total sense. So as a leadership team, you suddenly have a dashboard that shows the actual living heartbeat of the company. You can see that a specific approval process is creating a massive backlog, and just like the jet engine, you can simulate a change.

Right. You can route a new workflow through the digital twin and observe the downstream effects before you ever implement it in the physical office. It removes all the guesswork from organizational design.

But beyond just operational efficiency for human managers, this digital twin solves what I think is the single biggest barrier to enterprise AI adoption today.

The tribal knowledge problem.

Yes, the problem of tribal knowledge, which is just, you know, the undocumented, unwritten set of rules that only exists in the heads of employees.

Oh, we all know those rules.

It's knowing that if you want a specific budget approved, you have to email a certain director, but only on Tuesday mornings, and you [00:10:00] gotta format the request in a very specific way-

Yeah.

-or they ignore it.

Yeah. Tribal knowledge is annoying for a new human hire, but for an AI agent, it is an absolute death sentence.

Exactly, and the newsletter makes a definitive statement on this, which is: AI agents are only as good as the system they operate in.

They are.

I picture it like, dropping a world-class librarian into a massive library where someone has shredded all the books, mixed the pages on the floor, and turned off the lights.

That is quite an image.

But it's true. The librarian is brilliant, and they possess immense processing power, but without a structured catalog system, they are going to fail.

And that is a phenomenal visualization of how AI hallucination actually occurs in a corporate setting.

Right.

An AI model hallucinates when it is forced to connect dots in a structural void if it doesn't have a clear map of your data permissions, your hierarchies, your operational logic. It will just guess.

Hmm.

And in a highly regulated business environment, guessing is a massive liability.

Which brings us to the [00:11:00] technological solution that's highlighted in this issue.

Red Hawk Technology's proprietary methodology, TransformOS. Mm-hmm. And Scott Schultz, their AI lead staff engineer, he describes it as a system to build systems. Let's strip away the marketing terminology, though.

Mm-hmm.

How does TransformOS actually catalog the shredded library?

So TransformOS acts as the governance layer.

Okay.

Once the digital business twin is built, once the workflows, data silos, and decision trees are mapped out-

Yeah ...

TransformOS provides the operating environment where AI can safely interact with that map.

Ah, so it sets the boundaries.

It is entirely about boundaries. You are not giving a generalized AI model unrestricted access to your entire server.

TransformOS establishes strict read and write permissions. It dictates exactly which databases an AI agent can pull from, what format its output must take, and crucially, at what specific node in the process it must stop and hand the task over to a human for final review.

The text provides a [00:12:00] really concrete example of this in action with a healthcare client.

Yeah.

And we aren't just talking about a, you know, a customer service chatbot pinging a knowledge base. This client is running what they call agent fleets across six different business functions.

Yeah, and an agent fleet is a coordinated group of specialized AI models. Each one is designed for a narrow task, passing information back and forth entirely within the boundaries of TransformOS.

So you might have one AI agent that strictly parses incoming medical records to extract diagnostic codes. It cannot do anything else.

Right.

Then it passes those codes to a second agent, whose only job is to cross-reference those codes against insurance compliance databases.

Exactly, and if the second agent detects an anomaly, the governance layer forces it to kick the file out of the automated loop and place it directly onto the dashboard of a human auditor.

Wow.

So the AI operates with incredibly high velocity, but it remains entirely constrained within the structured logic of the digital twin.

It just fundamentally reframes how we think about AI transformation. It isn't about buying a shiny [00:13:00] new intelligence and just hoping it figures out your messy business.

No, not at all.

It's about cleaning up your business so that intelligence actually has a place to live.

Structure must precede scale. if you attempt to scale a broken, undocumented process by just applying AI to it, you will simply generate chaos at a much higher speed.

Yeah, chaos at scale. let's distill everything we've unpacked from Volume two, Issue 14 into a practical takeaway for the listener.

Yeah.

The era of treating human labor, software tools, and AI compute as isolated experiments That's over.

It has to be. To remain competitive, you must calculate your total unified cost of thinking. You must invest in mapping the reality of your day-to-day operations, making the invisible visible through a digital business twin.

Right.

Because only then can you deploy AI agent fleets with the governance and trust required to actually move the needle on productivity.

It is a profound shift in organizational architecture, and honestly, it provides a perfect setup for what the team at Red Hawk [00:14:00] Technologies is tackling in their next publication.

Matt mentions in his introduction that Issue 15 is going to address a vulnerability that every single company faces, regardless of its size.

Yes. They will be exploring the massive risk of knowledge attrition because every company relies on critical operational knowledge that is locked exclusively inside the heads of a few veteran employees.

Oh, man. Yeah. When those senior employees retire or get poached by a competitor or honestly, even simply take a two-week vacation, that knowledge just walks out the door with them.

Exactly. So the upcoming issue will detail methodologies for extracting that hard-won tacit expertise and encoding it directly into the digital twin.

That's amazing.

It's about transforming individual memory into institutional capability, ensuring that both your human teams and your AI fleets can leverage that wisdom long after the original employee has moved on.

Getting the unwritten rules out of human heads and into the structural code of the business, that is gonna be a really [00:15:00] vital follow-up.

It really is the ultimate test of whether an organization is truly resilient.

Well, as we conclude this deep dive, we want to leave you with a final thought to mull over as you analyze your own professional environment. We've talked extensively today about mapping the enterprise, but I want you to apply this to your immediate sphere of influence.

Right. Look at the tasks you execute every single day.

Yeah. If your company deployed a digital twin tomorrow and mapped the data flows of your specific job, how much of your daily work would show up as a clear, logical, beautifully paved highway? And how much of what you do relies entirely on a mysterious wilderness of workarounds, favors, and hidden spreadsheets that only exist on your local hard drive?

It's a tough question. Identifying your own blank spots on the map is really the first step toward true optimization.

Because if you can't see the wilderness, you can't build a road through it.

Mm-hmm.

Until next time, keep mapping.

From managing the grind to designing the machine. we'll see you next time.