A Conversation about the Strategic Shift: AI-Enabled Insourcing and Corporate Capability Building

Speaker A: I want you to take a second and, uh, just visualize your organization's budget.

Speaker B: Oh, yeah, that massive, intimidating spreadsheet.

Speaker A: Right. Think about that spreadsheet and just try to locate the line items for outside vendors.

Speaker B: There are usually quite a few of those.

Speaker A: Exactly. Picture the, uh, the monthly retainer for your external marketing agencies. Or think about the specialized legal counsel you have to call in for contract overflow.

Speaker B: Or that massive offshore software development consultancy your IT department always relies on.

Speaker A: Yes. Perfect example. Just hold that staggering figure in your mind for a moment. Now, what if I told you that you could bring all of that outsourced power, like all of that specialized output back inside your own building and presumably

Speaker B: for a fraction of the cost?

Speaker A: An absolute fraction.

Speaker B: I mean, it sounds completely counterintuitive.

Speaker A: Yeah.

Speaker B: For decades, the foundational rule of corporate strategy has literally been the exact opposite.

Speaker A: You shed those complex functions because maintaining them internally was always seen as just a bloated, expensive liability.

Speaker B: Right. They were cost centers you wanted off the books.

Speaker A: Well, today we are diving into a foundational text that argues that decades old rule is just completely dead. We're looking at this really fascinating research paper titled the Strategic AI Enabled Insourcing and Corporate Capability Building by Dr. Jonathan H. Westover.

Speaker B: It is a phenomenal, rigorous piece of research. Dr. Westover looks at how major enterprises are essentially just ripping up their traditional outsourcing playbooks.

Speaker A: Yeah. Totally redesigning how work gets done. And our mission in this deep dive is to explore exactly how and why they are doing it.

Speaker B: Because it's not just about saving money.

Speaker A: No, not at all. We're going to unpack why bringing these capabilities back in house using artificial intelligence is, uh, it's actually a critical, aggressive move to build a sustainable competitive moat around your business.

Speaker B: A moat that your competitors really can't cross.

Speaker A: Exactly. Okay, let's unpack this. Because to really grasp why this new trend is so disruptive, we first have to understand the old rule. It's completely replacing.

Speaker B: Right. So if we connect this to the bigger picture, we really have to look at the traditional paradigm of the make or buy decision.

Speaker A: The classic Economics 101 stuff.

Speaker B: Exactly. Since, like the 1990s, the logic of outsourcing has been treated almost like a law of physics. The golden rule is just stick to your core competencies.

Speaker A: Right. So if your company manufactures medical devices, your core competency is engineering medical devices.

Speaker B: Yes. You don't build a massive internal advertising agency, you don't maintain a giant legal firm inside your headquarters, and you definitely don't house an army of Software developers. Just for internal tools, you shed those

Speaker A: non core functions to specialized vendors because those external vendors have something you don't, which is massive economies of scale.

Speaker B: They have the specialized talent pool.

Speaker A: Yeah. And the operational flexibility to scale up or down whenever they need to. Plus they spread their massive overhead across dozens of different clients.

Speaker B: So you let the specialists do the specialized work.

Speaker A: Right. And you pay them a premium for it because the math dictates it's still cheaper than building an entire department from scratch inside your own four walls.

Speaker B: But Dr. Westover's paper points out that artificial intelligence is fundamentally breaking that old mathematical equation.

Speaker A: Breaking it completely.

Speaker B: AI tools are amplifying individual and small team productivity so intensely that those massive external agencies are just, well, they're no longer required to generate high volumes of work.

Speaker A: The cost of generating a contract, a marketing campaign, or even just a piece of code has plummeted.

Speaker B: It's basically approaching zero in some cases.

Speaker A: Yeah. And as I was reading through the mechanisms of how this works in the paper, a very specific analogy came to mind to explain, like the problem with the status quo.

Speaker B: I'm curious.

Speaker A: Well, if you keep outsourcing your work in this new landscape, you are essentially paying a premium for a taxi driver who is now just using a free GPS app on their smartphone.

Speaker B: Oh, I love that framing. It perfectly captures the economic misalignment going on right now.

Speaker A: Right. The taxi driver, which represents your external agency, is getting all the benefits of the AI efficiency.

Speaker B: The app is making their road faster, it's making their job easier.

Speaker A: Yeah. And it requires significantly less specialized geographical knowledge on their part. But you, sitting in the backseat as the passenger, you're still paying the exact same high fare.

Speaker B: And to make matters worse, that taxi driver is using that exact same GPS app to drive your direct competitors around.

Speaker A: Yes. Getting them to their destinations just as fast as you.

Speaker B: The paper actually frames that exact dynamic through the lens of organizational economics, so specifically focusing on, uh, where productivity gains are captured.

Speaker A: Because right now, the vendors are capturing it all.

Speaker B: Exactly. When your external vendors use AI, and let's be totally clear, they absolutely are doing this, those massive efficiencies diffuse broadly across their entire client base.

Speaker A: Right. So if an agency uses generative AI tools to draft your marketing copy in, half the time, they're also using those same tools for your industry rival.

Speaker B: Or if an external law firm uses AI to review your contracts, they just use that same platform for everyone else on their roster.

Speaker A: So your competitive advantage is completely neutralized. It just becomes a commoditized service.

Speaker B: Yes, you are Basically funding their internal efficiency. But you don't own any of it.

Speaker A: Wow. That is a tough pill to swallow for a lot of executives.

Speaker B: But when a company insources and uses AI internally.

Speaker A: Mhm.

Speaker B: They capture those gains directly. They start building proprietary workflows.

Speaker A: They train the AI models on their own specific localized historical data.

Speaker B: Right, Exactly. That creates differentiated capabilities that competitors literally cannot access. So you aren't just saving the taxi fare in your analogy.

Speaker A: Right? You are building your own hyper efficient transportation network that only your company has the keys to.

Speaker B: Well said.

Speaker A: Here's where it gets really interesting though. Because the math on this make or buy decision has changed so drastically, we are seeing this shift happen in some very specific high stakes areas.

Speaker B: Yes. The research highlights three key domains where AI enabled insourcing is aggressively taking over.

Speaker A: Let's look at marketing first.

Speaker B: Marketing is arguably the most active domain right now because generative AI tools, the platforms that can instantly generate text, images and video from simple prompts, have matured so rapidly.

Speaker A: Yeah. The paper shares this wild example of a mid sized financial services firm. Historically, they relied heavily on a sprawling external agency for all their content production.

Speaker B: We're talking blogs, social media posts, massive

Speaker A: campaign assets, white papers, everything. They decided to bring it all in house. They built a lean internal studio, equipped just a handful of internal creators with advanced generative AI tools, and the financial results were staggering.

Speaker B: What were the numbers on that again?

Speaker A: They slashed their external agency spend from $800,000 down to $300,000.

Speaker B: Wow. And that $300,000 figure isn't just what they paid a smaller agency. That actually covers the internal salaries for their new Lean team and and the enterprise subscriptions for all the AI tools.

Speaker A: But the cost savings aren't even the most impressive part. Like they didn't just save half a million dollars, their actual content output increased by 30%.

Speaker B: Because the AI fundamentally changes the workflow.

Speaker A: Exactly. Instead of a human copywriter drafting a white paper and then taking three days to manually chop that white paper up into 20 different social media posts and like five localized email newsletters, the AI handles that adaptation instantly.

Speaker B: The human strategy remains, but the heavy lifting of execution is automated.

Speaker A: Fewer people, less money, dramatically more output.

Speaker B: We see a very similar mechanism playing out in the legal sector, actually.

Speaker A: Oh, legal is fascinating, right?

Speaker B: Routine contract review, due diligence, regulatory compliance. This is work that traditionally gets shipped out to external law firms at incredibly high billable hours.

Speaker A: Thousands of dollars an hour sometimes.

Speaker B: Exactly. The paper discusses how internal corporate legal teams are now using AI powered document analysis tools to handle these Massive volumes themselves.

Speaker A: The text mentions a tech company that used AI to slash their standard contract turnaround time from five days down to under two days, which is huge. And if you've never worked in legal, you might wonder how AI actually does that without making massive errors. It's because the AI isn't reading the contract the way a human does. Line by agonizing line.

Speaker B: Right. The software pattern matches.

Speaker A: Yeah, it takes a 50 page non disclosure agreement agreement and instantly compares it against thousands of previous contracts and the company's established risk playbook.

Speaker B: It basically ignores the standard boilerplate.

Speaker A: Exactly. And it only flags the three specific clauses that deviate from company policy.

Speaker B: So the internal lawyer only has to review those three flagged clauses, not all 50 pages.

Speaker A: Which raises an important question. Why does a three day reduction in contract review actually matter to the broader business?

Speaker B: It matters because a legal department is not just a cost center, It's a bottleneck for revenue.

Speaker A: Oh, uh, that's a great point.

Speaker B: When you cut contract review from 5 days to 2 days, you are directly boosting your sales velocity.

Speaker A: Deals close faster, revenue hits the books faster. The AI doesn't replace the lawyer's deep expertise. Right. It just allows one internal lawyer to handle a caseload that previously required an entire external team of associates.

Speaker B: Exactly. Now, the third domain the research points to is software development.

Speaker A: Yes, the IT departments.

Speaker B: Companies that used to hire huge offshore teams or expensive external consultancies to build their internal applications are now utilizing AI coding assistance.

Speaker A: These are the tools that act like incredibly intelligent autocomplete for programmers, right?

Speaker B: Exactly. A developer can write a plain English comment saying, uh, create a secure login screen and the AI, uh, generates the 40 lines of foundational code instantly.

Speaker A: And internal developers are seeing 30 to 50% productivity gains from this. They are just bypassing the need for those offshore teams entirely.

Speaker B: It's completely changing the talent acquisition strategy.

Speaker A: But hold on, I have to push back here for a second. If I'm listening to this and I own a marketing agency or like I'm a partner at a law firm, I'm probably sweating right now.

Speaker B: I would imagine so, yes.

Speaker A: This sounds almost apocalyptic for the B2B service industry. Like, are, uh, companies just firing all their agencies overnight and letting the bots run the show entirely.

Speaker B: It is a totally valid concern. But the research is very clear that the answer is no. We are not witnessing a wholesale apocalyptic replacement of vendors.

Speaker A: Okay, good.

Speaker B: Dr. Westover calls this movement a strategic recalibration.

Speaker A: So they aren't just burning their vendor contracts in the parking lot. They are splitting the work. Right.

Speaker B: They bring the standardizable everyday execution inside the building.

Speaker A: But I imagine you still need the heavy hitters like the deep specialists on the outside.

Speaker B: That is exactly the dynamic companies are insourcing, the high volume work, the heavy lifting. That AI is exceptionally good at accelerating, but they are absolutely keeping highly specialized vendors on retainer.

Speaker A: The paper gives a brilliant example of a tech company that did this right.

Speaker B: Yes, they managed to cut their total vendor spend by 70%. They brought all their standard everyday software development m in house. But they deliberately kept three specialized external consultants on the payroll.

Speaker A: What were the three specific areas they kept outside?

Speaker B: They retained one consultant for deep machine learning optimization, one for highly complex accessibility compliance, and one for sensitive security architecture.

Speaker A: Okay, that makes perfect sense. You keep the surgeons on spade dial, but you bring the general practice inside your own building.

Speaker B: That's a perfect way to put it.

Speaker A: You don't need to pay an expensive agency to write basic boilerplate code or to draft a standard vendor agreement anymore. You keep your external partners for the true edge cases that require deep human specialized expertise.

Speaker B: The kind of rare knowledge that wouldn't make economic sense to hire full time anyway.

Speaker A: Exactly. Now, the ROI on all this is undeniably incredible and the underlying strategy makes perfect economic sense. But we have to address the implementation reality check.

Speaker B: Yes, because executing this transition is incredibly complex.

Speaker A: If you rush it, you will fail and you will likely break your existing operations in the process.

Speaker B: The research warns heavily against the abrupt rip and replace strategy.

Speaker A: You cannot just fire your entire marketing agency on a Friday, buy a couple of enterprise AI subscriptions and expect your internal staff to seamlessly replace them by Monday morning.

Speaker B: The companies that actually succeed use a highly structured phased approach. They allow their vendor contracts to wind down naturally while simultaneously building and testing their internal muscle.

Speaker A: If you are listening to this and panicking because your boss read an article about AI and wants to replace your external agency next week, you can point them to the healthcare organization case study in the text.

Speaker B: Oh, uh, that's a great example.

Speaker A: It proves that a methodical transition is the only financially responsible route. They wanted to phase out their marketing agency, but they took a full 12 months to do it.

Speaker B: You started with a hybrid model, right?

Speaker A: Yeah. In quarter one, the agency was still handling 75% of the workload, while the newly formed internal AI team handled 25%.

Speaker B: And that phase is crucial because it allows the internal team to stumble, figure out the tools and learn without grinding the company's output to a halt.

Speaker A: Exactly. Then over the quarters, they carefully adjusted the Dials it went to 50 50, then 25% agency and 75% internal, until

Speaker B: finally, after a full year, they hit 0% agency reliance.

Speaker A: And at the end of that year, their costs had dropped by 55% and their content output had risen by 40%.

Speaker B: Incredible numbers.

Speaker A: But they only achieved those numbers because they gave themselves a Runway to figure out how these tools actually fit into their specific messy real world workflows. Which brings up a huge trap that the paper highlights. I like to call it the tool trap.

Speaker B: What's fascinating here is the vital distinction between having a tool and having a capability. The paper points out that simply providing your staff with access to AI platforms is woefully insufficient.

Speaker A: It's like buying everyone in your marketing department a concert grand piano on a Tuesday and demanding a flawless symphony by Friday.

Speaker B: Exactly. The tool itself is totally useless without the underlying human capability to play it.

Speaker A: You can't just drop an AI prompt box in front of a copywriter or a paralegal and say, go, be 100% more productive.

Speaker B: No. And there was a professional services firm mentioned in the research that understood this intuitively.

Speaker A: What did they do?

Speaker B: They were building an internal AI augmented content team, and they made the radical decision to intentionally delay their production ramp up by three entire months.

Speaker A: Wow. Three months of just waiting while they

Speaker B: used that entire quarter purely to train their staff.

Speaker A: I want to break down what that training actually looked like, because it wasn't just like watching tutorial videos. They taught their team prompt engineering, which

Speaker B: is such a critical skill right now.

Speaker A: For anyone unfamiliar, prompt engineering is essentially learning how to speak the AI's language to get a highly specific, usable result.

Speaker B: It's the difference between asking an AI to write a blog post about finance, which gets you generic, unusable robot speak, and giving it a highly structured set of parameters.

Speaker A: Right. A trained employee knows to say, like, act as a senior wealth Advisor, write a 500 word piece targeting millennials, use an approachable tone, and incorporate these three specific data points from our Q3 report.

Speaker B: And beyond just prompt engineering, they ran workflow redesign workshops. They established clear quality assurance frameworks so the AI didn't publish hallucinations or errors.

Speaker A: Oh, that's huge. You can't have the AI making up facts in a financial report.

Speaker B: Definitely not. And they also developed ethical AI use guidelines to ensure compliance. It was a calculated, heavy investment, and it paid off massively.

Speaker A: Because of that intensive capability building, they vastly, uh, outperformed another division in the same company that had just handed out the AI tools with zero training, right?

Speaker B: Yes. The trained team had higher quality outputs and hit their productivity goals much faster. Because they weren't just using AI to do their old manual processes slightly faster. They had fundamentally redesigned how the work was executed from the ground up.

Speaker A: Part of that redesign is also realizing there is no single magic bullet tool that solves every problem.

Speaker B: This is a really common misconception.

Speaker A: Yeah. The paper detailed a financial firm that was trying to build their internal legal capability. They didn't just go out and buy the biggest, most famous AI software on the market and call it a day.

Speaker B: They piloted six different legal AI platforms over several months.

Speaker A: Six. Because the AI landscape right now is incredibly fragmented. Different underlying models are optimized for entirely different tasks.

Speaker B: They discovered they actually needed a multi tool setup.

Speaker A: They adopted one platform that was incredibly good at analyzing complex regulatory compliance documents and a completely different AI tool that was optimized for just churning out standard non disclosure agreements.

Speaker B: It requires real technological maturity to realize you are building a custom ecosystem of tools, not just installing a single app on your desktop.

Speaker A: We've covered the dramatic cost savings and we've analyzed how to survive the implementation phase without breaking your company. But if we connect this to the bigger picture, what is the ultimate strategic goal here?

Speaker B: Like, why go through all this effort?

Speaker A: Exactly. The answer is the ultimate moat.

Speaker B: Yes. The paper relies heavily on an academic concept called resource based view theory, which basically says the core idea is that a sustainable competitive advantage cannot come from something that anyone with the checkbook can just go out and buy.

Speaker A: It has to come from valuable, rare resources that are deeply difficult for your competitors to copy.

Speaker B: Think about the old UM model. When you outsource work to an external agency, that agency learns.

Speaker A: They learn what specific marketing messaging works best in your niche industry.

Speaker B: They learn what legal clauses consistently get pushback from your vendors. But all of that valuable knowledge lives in their heads.

Speaker A: It's their institutional memory. It literally walks out the door when your contract ends.

Speaker B: But when you do this work in house with AI, your organization accumulates that knowledge. Your internal AI tools get trained on your specific historical data.

Speaker A: You build custom templates that perfectly match your organizational voice and your unique risk tolerance.

Speaker B: And that knowledge becomes embedded within the DNA of your company. It becomes proprietary institutional memory.

Speaker A: A competitor can easily go by the exact same base AI tool you use. But they absolutely cannot buy the years of fine tuned organizational context you have built into it.

Speaker B: Because that knowledge is entirely internal, it unlocks a level of operational agility that is honestly unmatched by the old vendor model.

Speaker A: The research had this striking example of A retail company that got hit by a surprise highly aggressive marketing campaign from a major competitor.

Speaker B: Under the old traditional model, responding to that is an absolute nightmare.

Speaker A: Oh yeah, you have to call your agency, schedule a, uh, briefing, draft a formal statement of work, get the new budget approved by finance, wait for their

Speaker B: creative team to brainstorm, wait for legal review. It takes weeks to get a response into the market.

Speaker A: And by which point the competitor has already captured the narrative entirely.

Speaker B: Exactly. But this retail company had an internal AI augmented team. They were able to redirect their resources immediately.

Speaker A: They just sat in a room, used AI to instantly brainstorm counter messaging, use generative tools to build the creative visual assets in hours instead of days.

Speaker B: And they launched a full scale, highly polished responsive campaign in just 72 hours.

Speaker A: 72 hours. From surprise attack to full scale response. That is the exact definition of strategic agility.

Speaker B: You don't have to negotiate a scope change with an external vendor when the market shifts. You just pivot your internal resources instantly.

Speaker A: It's not just about speed either. It's about how these internal teams actually interact with each other. Because AI is amplifying individual output so much, you no longer need a marketing department of 50 people and a legal department of 50 people to get things done.

Speaker B: You could achieve the exact same output with much smaller teams.

Speaker A: And when teams are smaller, the traditional organizational silos start to naturally collapse.

Speaker B: This is one of the most profound secondary effects of this entire shift. AI is fundamentally flattening organizational structures.

Speaker A: The cross functional magic. There was a healthcare tech company in the techs that completely reorganized how they build and launch new products.

Speaker B: Historically, the developers would build the software, then they would toss it over the metaphorical wall to the legal department for a lengthy compliance review.

Speaker A: And then legal would eventually toss it to marketing to figure out how to sell it. It was incredibly slow.

Speaker B: Instead, they put everyone on the same small team from day one.

Speaker A: You have a software developer, a legal compliance expert, and a marketing specialist all sitting at the exact same table, all

Speaker B: armed with AI tools that multiply their individual capabilities.

Speaker A: And because they were deeply integrated from the very beginning, the legal compliance and the marketing positioning were baked into the software as it was being written.

Speaker B: They didn't have to go back and rewrite code because legal found a flaw.

Speaker A: Three months later, they sped up their entire product development cycle by 40%. So what does this all mean?

Speaker B: It means we are looking at a fundamental redesign of organizational boundaries. AI is not merely a productivity hack to help you write emails faster or summarize a PDF.

Speaker A: It is a catalyst for rethinking what your company is actually capable of doing on its own.

Speaker B: And Dr. Westover's paper makes it very clear the window to act proactively on this shift is rapidly closing.

Speaker A: The early movers are already compounding their advantages.

Speaker B: Every single month a company spends relying on external vendors for standardizable work is a month their competitor is spending training

Speaker A: internal AI models, refining their proprietary workflows, and building an institutional memory that cannot be replicated. Which brings us right back to you, the learner, listening to this deep dive. I want to challenge you to look closely at your own daily workflows and the budget of your own department.

Speaker B: Where are you currently relying on external specialists?

Speaker A: Where are you paying that premium taxi fare for tasks are that that a small, internally trained team armed with the right AI ecosystem could execute faster, cheaper, and with far deeper alignment to your company's actual goals?

Speaker B: It requires taking a hard, honest look at what is genuinely a core competency in this new era. Because the definition of core has just expanded dramatically.

Speaker A: It absolutely has. And I want to leave you with one final, slightly provocative thought to mull over as you go about the rest of your day. We've talked about the incredible benefits for the companies doing the insourcing, but if the research is right, every major sophisticated enterprise out there successfully insources its core functions into these lean, hyper efficient AI driven internal teams. What happens to the massive global ecosystem of B2B service agencies?

Speaker B: That is a million dollar question.

Speaker A: Are we witnessing the slow, inevitable death of the traditional mega agency model? Will the giant, sprawling ad agencies and massive legal consultancies eventually be replaced by a new fragmented economy of highly specialized micro consultants who only parachute in for the absolute most extreme edge cases? It's, uh, a fascinating, massive economic shift, and it's happening right now, hidden right inside your vendor budget.