Manager Spans Hit 12.1 Direct Reports in 2026—Is AI Making Us Better Leaders or Just Cheaper Organizations?

General
1

I’ve been wrestling with something uncomfortable for the past 6 months. When I started as VP of Engineering at our EdTech startup 18 months ago, I had 7 direct reports. Today, I have 11. And I’m not alone—this is happening everywhere.

The data confirms it: average engineering manager spans grew from 10.9 direct reports in 2024 to 12.1 in 2025. Meanwhile, every piece of research I read says the optimal range is 5-10 direct reports, with companies like Netflix finding that 6-8 is the sweet spot for effective technical mentoring and career development.

So what changed? Two words: AI tools.

The “AI Makes It Possible” Narrative

Our exec team’s logic went something like this: AI-powered engineering analytics give you real-time visibility into every engineer’s work. AI-assisted code review tools speed up PR feedback loops. Automated standup bots collect status updates. Productivity dashboards surface blockers before they become crises.

Translation: “You can manage more people now because the tools do the heavy lifting.”

And honestly? There’s some truth to that. I can track 11 engineers’ sprint progress without manually checking in. I can spot patterns in code review turnaround times. I can see who’s blocked and on what.

But here’s what the dashboards don’t tell me:

  • That one of my senior engineers is quietly burning out because they’re mentoring 3 juniors while shipping a critical feature
  • That my tech lead is thinking about leaving because our 1:1s have become 25-minute status syncs instead of career development conversations
  • That my team’s culture is shifting from collaborative to transactional because I’m spread too thin to notice the small moments that build trust

What We’re Really Optimizing For

I’ll say the uncomfortable part out loud: We’re using “AI enables it” to justify not hiring more engineering managers.

At $180K-$250K per engineering manager, every additional layer of leadership is expensive. If AI tools let each manager handle 11-12 reports instead of 7-8, that’s millions in savings at scale. I get it. I sit in the budget meetings. I see the pressure.

But I also see the cracks:

  • My 1:1 quality has dropped. I used to spend 30-45 minutes per person biweekly. Now it’s 30 minutes monthly and half of it is project updates.
  • My ability to mentor has evaporated. I barely have time to review career development plans, let alone coach through technical decisions.
  • My own manager (our CTO) went from 4 direct reports to 7. She’s drowning too.

The leadership shortage is real—55% of US engineering firms are recruiting internationally for executive roles because there aren’t enough qualified leaders domestically. Turnover among engineering executives is running at 15% annually.

So we stretch the managers we have. We call it “AI-enabled leverage.” We celebrate that a 12-person engineering team can now be restructured to 3 people using Cursor and Claude while missing the part where that manager becomes redundant when the org flattens further.

The Question I Can’t Shake

Are we building better organizations, or just cheaper ones?

Because if the answer is “cheaper,” we need to be honest about what we’re trading:

  • Long-term team health for short-term cost savings
  • Manager effectiveness for manager capacity
  • Leadership development for leadership efficiency

I don’t have the answer. But I do know that when my direct reports start looking for other jobs, it won’t be because their manager didn’t have good analytics. It’ll be because their manager didn’t have enough time.

So I’m asking: How are you navigating this? What’s your current span of control? What’s breaking first—your team’s performance, your own effectiveness, or your ability to develop future leaders? And are we all quietly accepting that “AI-enabled management” means “understaffed leadership”?

I’d love to hear what others are seeing, especially if you’ve found ways to push back on the “just manage more people” pressure while still hitting growth targets.

2

Keisha, this hits close to home. At our Fortune 500 financial services company, I’ve watched the same trend play out over the past 18 months. My own span went from 8 direct reports to 10, and some of my peer directors are now managing 12-13 engineers.

The financial pressure you’re describing is very real. When each engineering manager costs $180K-$220K (in our market), and you’re trying to scale from 200 to 300+ engineers, the math gets brutal fast. Adding 5 more managers to maintain an 8:1 ratio vs. stretching to 12:1 is a difference of $1M+ annually. Finance isn’t subtle about which option they prefer.

Where AI Actually Helps

I want to be fair to the AI tooling argument because some of it does work:

  • Code review velocity: GitHub Copilot and AI-assisted review tools genuinely sped up PR turnaround, which used to be a major bottleneck in my 1:1s
  • Sprint tracking: Engineering analytics dashboards (we use Jellyfish) give me real-time visibility that would’ve required 10+ manager-engineer check-ins per week
  • Incident management: AI-powered alerting and root cause analysis tools reduce the “firefighting” time that used to consume 20% of my calendar

Those tactical wins are legitimate. They do free up time.

What AI Doesn’t Replace

But here’s where your point about 1:1 quality really resonates. The time I gained from automated status tracking didn’t turn into more 1:1 time—it turned into managing more people with the same 1:1 budget.

What’s suffering:

  • Mentorship depth: I used to pair-program with my engineers quarterly to stay technical and coach them. That’s gone.
  • Career development: Performance reviews have become checkboxes instead of development conversations.
  • Early warning systems: I’m catching performance issues 2-3 months later than I used to because I don’t have the daily context.

The tools give me breadth, but I’ve lost depth.

The Hybrid Model I’m Testing

I’m trying something that’s showing early promise: tiered management responsibility.

  • Tech Leads (senior ICs, not people managers) handle: Daily standups, code review oversight, technical mentoring, sprint planning details
  • Engineering Managers (me and my team) handle: Career development, performance management, cross-team coordination, hiring, strategic decision-making

This lets me maintain a span of 10-11 while delegating the tactical parts that AI actually can assist with. Tech leads use the AI tools heavily. I focus on the human parts.

It’s not perfect—tech leads aren’t compensated like managers, and there’s an equity question there. But it’s better than pretending I can do deep mentorship for 12 people.

My Question Back to You

How are you maintaining culture at 11 reports? That’s where I’m struggling most. The small moments—hallway conversations, noticing when someone’s energy is off, building trust through consistency—those don’t scale. And the AI tools definitely don’t help there.

I’m also curious: Are your engineers feeling the difference? Or are they self-sufficient enough that the wider span isn’t impacting them yet?

3

Both of you are asking the right questions, but I want to reframe this conversation: It’s not about what AI makes “possible”—it’s about what’s sustainable.

I’ve been CTO at companies ranging from 50 to 250 engineers, and I’ve seen this movie before. The technology changes (first it was agile tools, then DevOps automation, now AI), but the pattern is the same: We use new tools to justify understaffing critical leadership roles, then act surprised when things break.

The Data Nobody Wants to See

At my last company, before I joined as CTO, they’d stretched manager spans from 7 to 11-12 during a growth phase. When I arrived and started pulling data, here’s what I found:

Manager Burnout Correlation:

  • Managers with 8-9 reports: 15% burnout/turnover rate
  • Managers with 10-11 reports: 28% burnout/turnover rate
  • Managers with 12+ reports: 42% burnout/turnover rate

Performance Issue Detection Lag:

  • At 7-8 reports: Performance issues surfaced within 4-6 weeks on average
  • At 10-11 reports: Detection lag increased to 10-14 weeks
  • At 12+ reports: Some issues weren’t caught until annual review cycle (6+ months)

Cultural Drift Indicators:

  • Teams under managers with 8-9 reports: 87% “feel connected to team culture” (quarterly survey)
  • Teams under managers with 12+ reports: 61% “feel connected to team culture”
  • The difference? Less face time per person = weaker culture transmission

The Hidden Cost Accounting Misses

Luis mentioned the $1M+ difference in maintaining lower spans. That’s real. But here’s what Finance doesn’t see in their spreadsheet:

When a manager burns out and leaves:

  • 3-6 months to backfill (during which their team is orphaned or redistributed)
  • $30K-$50K in recruiting costs
  • Loss of institutional knowledge and team relationships
  • Disruption to 10-12 engineers’ productivity during transition

When performance issues go undetected for 6 months:

  • Failed projects, missed deadlines, technical debt accumulation
  • Cost to remediate can be 10x the cost of early intervention
  • Impact on team morale when underperformers aren’t addressed

When culture erodes:

  • Attrition increases by 15-25% (we saw this directly)
  • Recruiting difficulty increases (employer brand suffers)
  • Lost productivity from disengaged engineers

Run that math and the “savings” from wide spans evaporates.

My Controversial Take

If you can’t afford enough engineering managers to maintain sustainable spans, you can’t afford that many engineers.

I know that’s harsh. I know CFOs hate hearing it. But it’s true.

Building a 100-person engineering team with 8 overloaded managers instead of 12-14 well-supported ones is a technical debt decision. You’re borrowing against future organizational health to hit short-term growth targets.

What I Actually Did

When I took over as CTO, I made two unpopular decisions:

1. Hard cap on manager spans at 8-9 reports
Even if it meant slowing hiring velocity. Even if it meant investing in manager development faster than we were comfortable with.

2. Tracked manager effectiveness metrics, not just capacity metrics:

  • Team retention rate (90-day and 12-month)
  • Performance review timeliness and quality (measured through skip-level conversations)
  • Career progression rate (are people growing?)
  • Manager burnout indicators (1:1 frequency, PTO usage, after-hours work patterns)

The result? We grew from 80 to 120 engineers over 18 months (slower than plan), but with:

  • 8% attrition vs. 22% industry average
  • 94% of engineers saying their manager “has time for me” in quarterly surveys
  • Zero manager burnout departures

The Path Forward

Keisha, to answer your question directly: You need to push back with data.

Start tracking:

  • How often you’re having meaningful (not status-update) 1:1s with each report
  • How long it takes you to notice when someone’s struggling
  • How many of your directs are actively developing new skills vs. just executing

When those metrics start declining, that’s your evidence that the current span isn’t sustainable. And if leadership won’t listen to the data, you have a decision to make about whether you’re willing to preside over the slow degradation of your team’s effectiveness.

Luis’s tiered model is smart—but be careful. Tech leads doing manager work without manager authority or compensation is a short-term patch, not a sustainable structure. Eventually they’ll either want the role officially (and the comp) or they’ll burn out too.

The real question: Are we optimizing for quarterly budget targets or for multi-year organizational health? Because those are different games with different rules.