Nova: Welcome to Aibrary. I'm Nova, and today we're diving into a Harvard Business Review article that's been making waves across the consulting world. It's called "AI Is Changing the Structure of Consulting Firms," and it argues something pretty radical: the entire pyramid structure of consulting — the model that's defined firms like McKinsey, BCG, and Bain for decades — is crumbling. Not in a decade. Right now.

Nova: That's exactly the debate the authors — David Duncan, Tyler Anderson, and Jeffrey Saviano — tackle right at the top. They say the conversation swings between two extremes. One side says AI will render consultants completely obsolete. The other says AI will only make them more indispensable. And here's the twist: both sides are wrong.

Nova: Their argument is that consulting isn't disappearing and it isn't being saved — it's being fundamentally reshaped. The structure itself is morphing into something new. And if firms don't adapt, they won't be killed by AI directly. They'll be killed by smaller, faster, AI-native competitors who figured out the new model first.

Nova: Let's start with what's actually breaking. For decades, the consulting industry has operated on what the authors call the pyramid model. Picture this: at the bottom, you've got a huge base of junior consultants — fresh MBAs by the hundreds. Their job? Research, data gathering, number crunching, building slide decks. At the top, a narrow point: the senior partners who manage client relationships and set strategy.

Nova: Exactly. And here's a staggering stat the follow-up HBR article cites: at prestige consulting firms, roughly one or two out of every hundred junior hires eventually becomes a partner. The rest leave — they burn out, move to industry, or go to clients. The entire economic model depends on this churn. Junior consultants bill lots of hours at relatively high rates while earning relatively lower salaries, and that spread generates the firm's profit.

Nova: That's exactly right. And this pyramid model has shaped everything: compensation, promotions, staffing, even the mental model of what "good consulting" looks like. The article points out that incentives are wired around headcount and leverage. More juniors equals more billable hours equals more profit. It's been a beautifully effective money machine for the big firms.

Nova: Let's get specific about what AI is actually doing. Generative AI tools, predictive algorithms, and synthetic research platforms are now automating the tasks that once filled junior consultants' weeks. And we're not talking about trivial stuff. We're talking about research, data analysis, scenario modeling, and — here's the big one — creating presentation decks.

Nova: It absolutely is. BCG developed a tool called Deckster that can spit out presentation decks in minutes. McKinsey built a proprietary AI assistant called Lilli that's now used by over 72% of its workforce and has reduced research and synthesis time by about 30%. Bain deployed Sage, an AI copilot trained on its internal intellectual property. Deloitte has Zora AI agents. PwC has an entire agent operating system platform.

Nova: That's the uncomfortable question. The article argues that if AI takes over the work that used to justify thousands of billable junior hours, the pyramid collapses under its own weight. You can't justify staffing a team of ten when AI plus three people can do the same work faster and arguably better.

Nova: The key phrase they use is that AI enables them to "staff assignments with smaller, more senior teams and spend more time with clients where it matters most." That's the promise. But it fundamentally breaks the old staffing model.

Nova: So if the pyramid is crumbling, what replaces it? The authors propose something they call the consulting obelisk. Picture a tall, narrow monument — like the Washington Monument. Fewer layers, smaller teams, but more leverage at every level. Instead of relying on sheer scale, the obelisk is built around three specific human roles.

Nova: Completely different. The first role is the AI Facilitator. These are early-career consultants, but their job isn't grinding Excel models. They're trained in the latest AI tools and data pipelines. They design and refine AI-driven workflows so the team can generate insights at speed. The authors call it a new kind of apprenticeship that emphasizes technical fluency and applied judgment from day one.

Nova: The second role is the Engagement Architect. These are experienced consultants who lead projects — but their work changes too. They define the problems to solve, interpret AI outputs with human judgment, and translate those outputs into actionable strategies. They're not managing an army of analysts. They're orchestrating how work gets done, adapting as conditions change.

Nova: The Client Leader. This is the senior partner who cultivates deep, trusted relationships with senior executives. The authors say this role is about helping clients "make sense of change" and staying close enough to advise them on staying ahead of disruption. Interestingly, this is the one role that looks most similar to the old model — the top of the pyramid — but now it's supported by a much leaner structure underneath.

Nova: Here's where the article gets really concrete. The authors highlight several firms that are already operating on the obelisk model — and most of them aren't the big names you'd expect.

Nova: Exactly. Take Unity Advisory. It was launched by former partners from Big Four professional services firms and backed by 300 million dollars in private capital. It positions itself as conflict-free — no audit-advisory entanglements — and AI-native by design. Unity doesn't hire large cohorts of entry-level analysts at all. Instead, it uses agile pods of senior consultants who work in close coordination with proprietary AI tools.

Nova: Then there's Monevate, a firm focused exclusively on pricing strategy. They combine deep expertise with AI-enabled playbooks and modeling tools, delivering advice without a traditional analyst layer. And SIB, which specializes in cost reduction, uses AI agents to scan invoices and vendor contracts for savings opportunities — they deploy human experts only when the AI flags something that needs a judgment call.

Nova: The article is clear on this. Firms that cling to old structures risk becoming slower, more expensive, and irrelevant. And the authors are blunt: "This is not a moment for incrementalism. The winners will be those who move first and reimagine the industry before someone else does it for them."

Nova: The article leans heavily on Clayton Christensen's Innovator's Dilemma to explain why the big firms won't easily adapt. And the reason is simple: the pyramid model is still printing money.

Nova: That's exactly the tension. The authors point out that even as firms invest in AI tools and announce AI innovation labs, those capabilities often remain siloed from core delivery. You get flashy demos for clients, but the underlying engine — large project teams staffed with junior talent — remains largely untouched.

Nova: There's also a talent realignment challenge. Traditional firms are built to recruit and train generalist MBAs by the hundreds. But the obelisk model demands smaller cohorts fluent in AI tools, data workflows, and systems thinking. PwC has committed a billion dollars to AI training, but the article notes that culture and incentives will naturally lag. Systems still favor hours billed over insight delivered.

Nova: The compensation model is another huge barrier. The article argues that compensation must evolve to reward strategic contribution and client outcomes over billable hours. But that's a massive cultural shift for organizations where the billable hour has been sacred for decades.

Nova: There's a fascinating section in the article about AI governance and ethics — and this is where Jeffrey Saviano, the third co-author, brings deep expertise. He's a former EY partner who now leads a research team developing new models for AI governance and ethics at Harvard's Edmond and Lily Safra Center for Ethics.

Nova: Absolutely. In the traditional pyramid model, client deliverables passed through layers of review — analyst, senior consultant, manager, partner. Each layer caught errors and assigned responsibility. But in the obelisk model, small teams move fast, and AI plays a much larger role in decision-making. There are fewer human checkpoints.

Nova: That's exactly the question the article raises. Saviano's work at Harvard emphasizes that business leaders need to take responsibility for governing AI themselves rather than waiting for government regulation. The article calls for ethical guardrails to be built directly into how AI is used, not relying solely on centralized compliance teams or after-the-fact reviews.

Nova: The article argues this is especially important because small expert teams are being empowered by tools capable of influencing high-stakes decisions. You're giving a tiny team enormous leverage through AI. That's powerful, but it's also risky if the guardrails aren't there.

Nova: So let's bring this together. The HBR article makes a compelling case that AI isn't just a tool for consulting firms — it's a structural earthquake. The pyramid model that's defined the industry for decades is being replaced by the obelisk: leaner, faster, and built around three roles — AI Facilitators, Engagement Architects, and Client Leaders.

Nova: The article leaves us with three big implications. First, for consulting firms: evolve your delivery model or risk becoming irrelevant. Not incrementally — structurally. Second, for consultants: double down on judgment, framing, persuasion, and leadership. Those are the skills AI can't replicate. Third, for clients: expect faster, leaner, and potentially cheaper consulting engagements, but also demand clarity on AI governance and accountability.

Nova: Whether you're a consultant, a client, or just someone watching the AI revolution unfold, the message is the same: the structure of expertise is changing. And the obelisk is rising.

Nova: This is Aibrary. Congratulations on your growth!