The Mushroom Effect: What AI Really Means for the Future of Consultancy
The conversation around AI in professional services tends to follow a familiar arc. Excitement about efficiency. Reassurance that humans aren't going anywhere. A few slides about "augmentation, not replacement." Then everyone goes back to work and not much changes.
We think that framing misses something important. Not because AI isn't genuinely transformative, it is, but because the real disruption isn't coming for the work people think it's coming for. It's coming from underneath.
The Mushroom Effect
Mushrooms don't grow from the top down. They push up from the ground, quietly, in the dark, until suddenly they're everywhere and the landscape looks completely different.
AI in consultancy works the same way. The partners aren't going anywhere. The senior analysts, the domain experts, the people who sit across from a client and translate ambiguity into a brief — that work is still deeply human, still irreplaceable, still years away from being meaningfully threatened.
But the ground floor? That's already shifting.
Entry level roles — the graduate analysts, the junior data wranglers, the associates who spend their first two years cleaning datasets, building slide decks, running standard reports, and formatting dashboard. Those roles are being quietly hollowed out. Not all at once. Not with a bang. But steadily, and with increasing pace.
The tools are simply too good now. A mid-level consultant with access to the right AI stack can produce in an afternoon what used to take a junior a week. That's not speculation, it's already happening in firms across New Zealand, Australia, and globally.
Why This Is a Bigger Problem Than It Looks
Here's the part that doesn't get talked about enough: entry level roles aren't just the bottom rung of a consultancy. They are, for most firms, the engine of the revenue model.
The traditional consultancy structure is built like a pyramid. A small number of senior people win work, direct strategy, and manage client relationships. A much larger number of junior and mid-level staff execute that work and get billed out at rates that generate the margin the business runs on. The pyramid only works because the base is wide.
If AI compresses the base, if the work that used to require six junior staff can now be done by two with better tooling, the revenue model doesn't just shift. It breaks. Fewer billable hours at the bottom means less total revenue, even if the quality and speed of output improves. Efficiency gains don't automatically translate to commercial gains. Not under the old model.
And there's a second-order problem too. Entry level roles exist not just to generate revenue, but to develop people. The junior analyst grinding through data cleaning and report building is also learning, building pattern recognition, developing commercial instincts, understanding how data connects to decisions. If that pipeline dries up, where do the senior consultants of 2035 come from?
What Has to Change
The consultancy model was priced and structured for a world where human hours were the primary input. That world is ending. The firms that survive, and thrive, will be the ones that restructure around the new reality rather than defend the old one.
A few shifts we think are inevitable:
Billable rates will have to rise. If fewer hours are being billed to deliver the same or better outcomes, the value of each hour has to be repriced upward. Clients will push back — but the argument is straightforward. You're paying for the outcome, the expertise, and the accountability. You're not subsidising a training programme anymore.
Value-based pricing will accelerate. Time-and-materials billing was always a proxy for value. AI makes that proxy look increasingly arbitrary. Firms that move to outcome-based or retainer models, where the price reflects the result, not the hours, will be better positioned to capture the commercial upside of AI efficiency without eroding their own margins.
Specialisation becomes the moat. Generalist analytical work, the kind that used to fill junior timesheets, is precisely what AI does well. What it doesn't do well is deep domain expertise, nuanced stakeholder management, and the kind of contextual judgement that comes from years in a specific field. Consultancies that double down on genuine specialisation will be harder to commoditise.
The junior talent pipeline needs reinventing. If the traditional entry level role disappears, firms need new on-ramps for developing talent. Apprenticeship models, embedded client roles, research functions, and AI-augmented graduate programmes are all worth exploring. The goal is the same, build the next generation of senior practitioners, the path just has to change.
Smaller, sharper teams become the norm. The wide-base pyramid model gives way to something leaner. Boutique firms with highly skilled, senior-weighted teams and strong AI capability will be able to compete with, and often outperform, much larger consultancies on quality, speed, and agility. The overhead advantage of size shrinks when the tooling levels the playing field.
The consultancies that will struggle are the ones that treat AI as a productivity add-on while leaving the underlying business model untouched. The firms that will flourish are the ones willing to ask harder questions — about pricing, about structure, about what value really means in a world where the analytical heavy lifting is increasingly automated.
The mushroom is already growing. The only question is whether you're tending the ground or standing on it.