Megan Skiendel: Hey — you have a good week? Because I have been sitting with something weird since Monday and I genuinely cannot decide if it's a story or just a very well-timed press release.
Zara Reyes: Oh no, what dropped?
Megan Skiendel: Okay — picture this. A founder hits a billion-dollar valuation, like, literally just closed a hundred-million-dollar round. And then, the same week, he walks up to a microphone and tells the entire consulting industry its hiring model is broken. Publicly. On the record.
Zara Reyes: Wait — on the record? Not a leaked deck, not an internal memo?
Megan Skiendel: Business Insider, August twenty-fourth. His name is Nicolas Kopp. Co-founder and CEO of a company called Rillet — AI-native accounting software, basically the infrastructure that would replace the junior analyst grunt work. And the day his company hits unicorn status, he's telling McKinsey and Deloitte and every firm like them: stop mass-hiring junior generalists. The era is over.
Zara Reyes: So the car dealership owner just announced nobody needs to learn to drive.
Megan Skiendel: That is almost exactly it. And the question is whether that disqualifies the argument or whether he's actually right.
Zara Reyes: Because he could be both. That's the uncomfortable part.
Megan Skiendel: And the market data that landed after he said it is genuinely uncomfortable. Adzuna tracks UK graduate job vacancies. As of July 2026, they're at the lowest level since Adzuna started recording in 2016. One year, almost fifty percent down.
Zara Reyes: Wait — fifty percent in one year?
Megan Skiendel: One year. And look, at their 2017 peak, Adzuna had over fifty-five thousand graduate roles listed. By July 2026 you're looking at roughly one-sixth of that number. So picture a recent grad — offer letter from a Big Four firm, maybe rescinded, maybe just never came — refreshing that job board and finding half the listings gone versus twelve months earlier. That's not a vibe, that's a decade-low data point.
Zara Reyes: And it's not just vacancies. UK youth unemployment — sixteen to twenty-four — hit sixteen point two percent in the three months to March 2026. Over a million young people classified as NEET. So it's not theoretical.
Megan Skiendel: That number is real. But here's where I actually want to slow down, because — honestly — the causation story is messier than the headlines make it. Then you get DeciBio. Specialized management consulting firm. They cut their incoming entry-level class from fifteen hires to four. Despite double-digit revenue growth.
Zara Reyes: No but that's the one that lands for me. Revenue up, bodies down. That's the pyramid actually compressing, not just a prediction.
Megan Skiendel: It is the clearest example we have. One firm, though. One. And then KPMG Australia — three hundred sixty employees, twenty-seven partners cut, concentrated in consulting and business services. Five percent of the workforce, August 2026.
Zara Reyes: So that's the AI story, right? That's Kopp's thesis playing out?
Megan Skiendel: KPMG said it was government consulting demand. Difficult market conditions. Potentially lasting until 2028. They did not say AI. Which — frankly — is its own data point.
Zara Reyes: But that's — that's actually the thing that doesn't let me sleep on this cleanly. Because if KPMG Australia is blaming government demand, that's real. They reported a one percent revenue decline, A$2.257 billion down from the prior year. That's a macro story. So how much of the graduate vacancy collapse is AI and how much is National Insurance hikes and minimum wage floors all landing at once?
Megan Skiendel: Honestly — you can't separate them. That's the actual answer. The data doesn't let you. All three forces hit the same hiring window. And no one — McKinsey, BCG, Deloitte, EY, Accenture, IBM Consulting — not one of those firms has stood up and said 'AI is why we pulled back on junior classes.' Not one.
Zara Reyes: Wait, none of them?
Megan Skiendel: None of them. Which — look — could mean they're being careful. Or it could mean the AI causation story is being written by people with a product to sell. And then Harvard Business Review publishes analysis saying AI disrupts the classic professional-services leverage model, firms shift toward selective leadership-focused recruiting — which independently supports Kopp's framing. But HBR also has an incentive to publish the provocative frame.
Zara Reyes: So even the independent validation has a noise problem.
Megan Skiendel: And then Mercury.com surveys fifteen hundred founders — August 2026 — and finds that fears about AI replacing junior workers weren't fully translating into reality at early-stage companies yet. So at the startup layer, the apocalypse hasn't arrived. Which is the complication. The narrative is running ahead of the evidence.
Zara Reyes: Except DeciBio. Like — a consultant at DeciBio, supposed to get an offer letter this spring, revenue at the firm is growing double digits, and four people get called instead of fifteen. That hire who didn't get the call? That's not a narrative. That's a closed door.
Megan Skiendel: No, that one doesn't bend. Revenue up, class slashed — that's the exception that forces the real question. And look, the part that comes later actually makes this harder, because if Kopp is right that firms should be converting their expertise into agents layered onto platforms like Rillet — we need to ask who owns that asset when it's done.
Zara Reyes: Oh — that's the one I want to get into.
Megan Skiendel: And that ownership question is the one that actually cracks the whole historical parallel open. Because — look — offshoring didn't do this. ERP didn't do this. Every prior wave automated the execution layer. Junior analysts went to Manila or got replaced by SAP and the pyramid just... absorbed it. Redeployed the base. The work mutated, the bodies stayed.
Zara Reyes: Right — but the mutation worked because the intellectual core was still human. Like, you move the spreadsheet-building offshore, but someone still has to think through what the spreadsheet means.
Megan Skiendel: That's exactly where this breaks differently. AI isn't automating the geography or the keystrokes. It's automating the cognitive pattern-matching — the synthesis. That is the job. That's what McKinsey, BCG, Deloitte are actually paying junior analysts to do with their minds. Not type. Think in patterns across large data sets.
Zara Reyes: So the thing we were redeploying them to — that's the thing that's automated now.
Megan Skiendel: And then Kopp's thesis layers on top of that — I mean, the expertise-to-agent conversion piece. Firms package decades of accumulated domain knowledge into skill files, repositories, agents, layered onto platforms like Rillet. Which sounds like a great value proposition until you ask — wait, actually — who owns those agents when they're built?
Zara Reyes: Hold on — if a firm converts its most defensible knowledge asset into an agent that runs on Rillet's platform, doesn't Rillet now... hold that asset?
Megan Skiendel: That's the question nobody's answered publicly. And frankly, I don't think the firms have answered it internally either. The eighty-two percent figure from Mercury — eighty-two percent of AI-adopting startups already making changes to compensation or role structure — that tells you redefinition is real and it's moving fast. But redefinition toward what? Toward building agents on someone else's infrastructure?
Zara Reyes: No but that's — that's the commercial trap, right? A junior analyst at DeciBio who doesn't get the offer this year isn't just a casualty of automation. They're a casualty of the firm deciding its expertise is now an asset you package and hand to a vendor. That's a different kind of loss.
Megan Skiendel: And Kopp, whose company just hit a billion dollars on exactly that premise, is the one narrating it as progress. Which — honestly — doesn't mean he's wrong. It just means nobody should be taking that framing at face value without asking what he's selling.
Zara Reyes: So he's narrating the fire and selling the exits.
Megan Skiendel: Yeah. And maybe he's right. That's the part I can't resolve cleanly. I mean — at the start of this I kept thinking: is this a story or a very well-timed press release? And honestly, sitting here now... it might be both. The data behind it is real enough that the press release gets to call itself a story.
Zara Reyes: The thing to watch is just — do those entry-level titles come back? Or do they come back with names nobody's heard yet?
Megan Skiendel: That's it. Change management and implementation near-term, skill files and agents long-term — that's Kopp's map. Whether the person building those agents gets called a junior analyst or something else entirely, I genuinely don't know. And frankly, neither does he.
Zara Reyes: Good weird week, though.
