The Collapsing Middle: The Preventable Talent Crisis in Software Consulting
Your junior consultants can code.
They can prompt.
They can ship.
But many are not customer facing, not consulting.
And in three to five years, you might not have any consultants who can consult.
The Disappearing Middle
In our conversations with 15 software agency leaders at the end of 2025, we heard a consistent pattern:
Firms have stopped hiring junior consultants.
They are not replacing people when they leave.
Teams are running lean, citing AI efficiency and economic pressure.
On the surface, this looks rational. AI increases output. Fewer people can do more work.
But something subtle is happening underneath.
The few junior and mid-level consultants who remain are often skipping the work that builds judgment.
They:
Offload early-stage thinking to AI
Accept outputs without deep interrogation
Avoid the repetition that builds pattern recognition
Miss the small failures that teach recovery
They produce. But they are not accumulating the cognitive reps that create strong consultants.
Why AI Accelerates the Risk
Consulting judgment develops through:
Wrestling with ambiguity
Reframing messy problems
Making imperfect decisions
Recovering from mistakes
Seeing patterns across situations
AI reduces friction in execution. That’s useful.
But friction is where learning often happens.
When juniors never struggle through a poorly defined problem because AI takes care of it for them, they don’t build the muscle memory that previous generations rely on instinctively.
Layer on:
Fewer mentors (because seniors are overloaded)
Leaner teams
Remote disconnection
Fewer client-facing reps
And you have the ingredients for a collapsing middle.
What Happens in 3–5 Years
If this trajectory continues, firms will face a painful reality:
There will not be enough competent mid-level consultants ready to:
Lead projects
Manage client relationships
Run discovery
Make sound decisions independently
Seniors will burn out with no successors ready.
Clients will work with technically capable but strategically shallow teams.
Firms will discover too late that execution scale without judgment scale is fragile.
This Future Is Preventable
But only if you treat talent development as a strategic investment, not an afterthought.
1. Use AI 30% Less Than You Want To
Not on everything. Not always.
But deliberately.
Assign projects or phases where juniors must:
Frame the problem themselves
Draft solutions without AI assistance
Explain their reasoning before validating it
This short-term cost buys long-term capability.
2. Turn AI Into a Thinking Tool, Not a Replacement
Choose projects where AI is used as exercises and require juniors to:
Explain the output to a senior
Identify three alternatives AI didn’t suggest
Describe one context where the recommendation would fail
This builds interrogation skills, not dependency.
3. Protect Mentorship Time
If seniors are fully billable, mentorship disappears.
Block time. Make it visible. Make it measured.
A practical benchmark: 15–20% of senior time allocated to coaching and development.
If you can’t afford that now, you’ll pay later.
4. Track Bench Depth
Stop measuring only utilization and revenue.
Start tracking:
How many mid-level consultants can lead discovery alone?
How many juniors can manage a client conversation competently?
How many projects can run without senior rescue?
Those metrics predict whether your firm will still be strong in five years.
The Real Risk
The danger isn’t that AI replaces consultants.
It’s that it creates a generation who never become skilled consultants.
By the time you notice the missing middle, the seniors who could have trained them may already be gone.
The firms that win won’t be the ones that automate the most.
They’ll be the ones that develop judgment on purpose.