Back to all posts
The Third Race to the Bottom
Consulting Practice • • 8 min read

The Third Race to the Bottom

Consulting has played this game twice before—with body shops, then offshore. Now AI is the new lever. But what if the pattern itself is the problem?

consulting ai industry strategy
NC

Nino Chavez

Principal Consultant & Enterprise Architect

I’ve watched this movie before.

The consulting industry is celebrating its shiny new efficiency lever—AI agents that can synthesize research, draft decks, and automate the grunt work that used to fall on analysts and associates. McKinsey has “Lilli.” Bain has “Sage.” PwC has “ChatPwC.” Every major firm is racing to deploy internal AI tools that cut delivery time and reduce headcount.

And I keep thinking: we’ve done this before. Twice.

But this time feels different. Not because AI is more powerful than previous efficiency levers—though it is. Because AI doesn’t just reduce cost. It changes who can do the work well.

The First Race: Body Shops

In the mid-1990s, the Y2K crisis created unprecedented demand for COBOL and mainframe expertise. The solution? Body shopping—recruiting IT workers in India specifically to contract them out at high utilization rates.

Between 1996 and 1997, Indian firms trained tens of thousands of graduates for Y2K remediation work. The margins were extraordinary. The model was simple: low-cost labor, high billable rates, maximum utilization.

After Y2K lost its urgency, the body shopping phenomenon didn’t disappear. It evolved. By 2000, roughly 1,000 body shops offered Indian developers to US clients alone.

The result was predictable: fierce price competition, compressed margins, and a race to the bottom where the only differentiator was how cheaply you could provide labor.

The Second Race: Offshore Leverage

The early 2000s brought the next wave. Knowledge-intensive work—R&D, engineering, software development—surged to India and the Philippines. Western firms that had dominated the market suddenly faced competition from Tata, Infosys, and Wipro on global bids.

The response? Accenture, IBM, and HP opened offices in Southeast Asia. They had to. The labor arbitrage was too significant to ignore.

This created the “offshore leverage model” that still dominates today. Senior partners in expensive markets, delivery teams in low-cost geographies. Margins preserved through geographic wage disparities.

But here’s what the triumphalist narratives don’t mention: by 2005, a quarter of companies that outsourced work ended up reversing the decision. Rising salaries in offshore locations, communication challenges, quality issues. The arbitrage eroded.

The model didn’t fail. But it commoditized. When everyone has the same offshore leverage, it stops being a competitive advantage. It’s table stakes.

The Third Race: AI

Now we’re watching the third wave unfold in real-time.

McKinsey has been running “Project Magnolia” since 2023, targeting non-client-facing roles. Their headcount dropped from roughly 45,000 to 40,000 over eighteen months—a 10% reduction through layoffs, attrition, and raised performance bars. PwC cut 5,600 staff in the twelve months ending June 2025—their first workforce contraction since 2010.

In the UK, graduate job postings in accountancy fell 44% year-over-year according to Indeed data. KPMG made the steepest cut, reducing graduate intake by 29%.

Alan Paton, formerly of PwC’s financial services division, predicted that 50% of roles in audit, tax, and strategic advisory could be automated within three to five years. Industry skeptics note the limits of current AI in areas requiring professional judgment—but the direction is clear.

The playbook looks familiar:

  • New efficiency mechanism discovered
  • Firms compete on price by deploying it aggressively
  • Margins compress across the industry
  • Quality becomes harder to differentiate
  • The “advantage” becomes baseline expectation

But here’s where I think the pattern breaks.

The Accelerator Problem

Body shops and offshore leverage were pure labor arbitrage. You replaced expensive hours with cheap hours. The work stayed the same—just cheaper hands doing it.

AI is different. In the right hands, it doesn’t just reduce cost. It amplifies capability.

I’ve seen this in my own practice. An AI-augmented engagement doesn’t just cost less—it moves faster, explores more options, catches edge cases that would have slipped through. The ceiling goes up, not just the floor.

But here’s the uncomfortable part: in the wrong hands, AI adds slop.

It generates confident-sounding deliverables that don’t hold up under scrutiny. It produces volume without insight. It lets teams ship faster without understanding what they’re shipping.

I used to think AI would commoditize consulting the same way offshoring did. Now I’m less sure. It might do the opposite—create a wider quality gap that’s harder for clients to evaluate until it’s too late.

The Pattern Is the Problem

Here’s what I keep coming back to: each wave promises transformation but delivers commoditization.

Body shops turned expertise into interchangeable labor units. Offshore turned geographic wage arbitrage into an industry-wide expectation. AI is turning knowledge work into automated pipelines.

The firms that “win” the race don’t actually win. They just survive long enough to compete in the next round.

The consulting industry keeps discovering new levers to reduce the cost of delivery. What it hasn’t figured out is how to escape the race itself.

I’ve built my practice on the assumption that judgment can’t be automated. That the messy human stuff—reading a room, challenging assumptions, building trust—will always require a person.

Am I right? Or am I just telling myself a story that keeps me comfortable?

What Gets Left Behind

The thing about efficiency gains is they’re zero-sum in a competitive market. If everyone can deliver a strategy deck 30% faster using AI, the client expectation resets. Now that’s just baseline delivery speed.

But something else happens in the race: the things AI can’t do become harder to value.

Trust. Strategic judgment. The ability to challenge a leader, listen to their doubts, and confront their choices without a hidden agenda. The capacity to read organizational politics and navigate human dynamics that don’t show up in data.

One analysis I read put it plainly: “Clients don’t buy slides—they buy clarity, confidence, and judgment. AI can process information. It can’t persuade a board, reframe a problem, or lead a transformation under pressure.”

That’s true. But try pricing it.

The Bifurcation

What I’m watching now is the market splitting:

On one side: large firms consolidating, deploying AI to compress delivery costs, competing on scale and speed. They’ll eat the mid-market, acquire AI specialists, and fight for transformation projects where volume matters.

On the other side: boutique firms specializing in the things AI can’t touch. Deep industry expertise. Trusted relationships. Strategic judgment that requires experience, not just synthesis.

The middle ground—the generalist consulting firm with moderate offshore leverage and no AI differentiation—is becoming untenable.

What I Think Today

I don’t have a clean answer here. That’s partly the point.

What I see is an industry that keeps finding new ways to do the same thing: extract more output from less labor. Each wave feels transformative at the time. Each wave ends in commoditization.

But AI might break the pattern—not by avoiding the race, but by splitting it into two different races.

Race one: Cost compression. Big firms deploying AI to automate everything automatable, competing on price and speed, eating the mid-market. This race goes to the bottom. Same as before, just faster.

Race two: Capability amplification. Practitioners using AI to do work that wasn’t possible before—not cheaper versions of old deliverables, but fundamentally better ones. More thorough analysis. More creative solutions. More rigorous stress-testing of assumptions.

The firms that break the pattern will be the ones that stop competing on cost entirely. Value-based pricing. Outcome-linked fees. Specialization so deep that price comparison becomes meaningless.

But here’s what I keep circling: you can only run race two if you know what good looks like before the AI starts generating.

The slop comes from people who can’t tell the difference between a confident-sounding output and a correct one. Who accept the first draft because it’s plausible, not because it’s right. Who mistake velocity for value.

The race to the bottom has a finish line. The question isn’t just whether you want to be standing there—it’s whether you know which race you’re running.

Sources:

Share:

More in Consulting Practice

The Kitchen I'm Looking For

I don't need more Salesmen. I need a Sushi Master, a Pitmaster, and a Molecular Gastronomist. AI lets us return to the Guild—craftspeople in their own lanes, augmenting their own mastery.

The Tragedy of the Timesheet

What if the billable hour isn't a business model—it's a coping mechanism? A way to avoid confronting that the thing we're selling might not be scarce anymore.