Listen to the episode

  • 13 min

The Accidental Manager: Why Everyone in Your Company Is About to Lead a Team (of AI Agents)

Org Design Podcast

Play

🎧 Spotify, Apple, 📺 YouTube 

Please hit subscribe wherever you consume our podcast ✌️

About the guest

Amy Springer is the Chief of Staff at Functionly, specializing in organizational design and workforce planning. With a background in botany and extensive experience in consulting, including roles at a Big Four firm, global Mining and a health tech company, she applies systems thinking to enhance organizational structures. As the host of the Org Design Podcast, she bridges the gap between consulting theory and practical application, showcasing insights from leaders and practitioners in the field. Learn more about her on her expert page.

Summary

In this episode of the Org Design Podcast, host Amy J Springer explores the transformative impact of AI on organizational structures. As AI tools and agents become integral to daily operations, everyone in the organization is starting to manage teams of these agents. The discussion emphasizes the importance of organizational design thinking as a critical skill in the AI era, highlighting four key principles for effectively structuring hybrid teams. Listeners will learn how to define roles, measure outputs, and ensure clarity in responsibilities to optimize the collaboration between humans and AI. The episode also outlines actionable steps for organizations to prepare for this shift, making it essential listening for leaders navigating the evolving landscape of work. Tune in to understand how to leverage AI for effective team management and organizational agility.

Transcript

[00:00:00] Amy J Springer: Here's something your organization probably isn't preparing people for:

[00:00:03] Amy J Springer: In the next 18 months, everyone in your company is going to become a manager.

[00:00:07] Amy J Springer: Not a manager of people—a manager of agents.

I’ve spoken to five leaders in the last couple years of interviews that are implementing AI at scale, and they all said the same thing in different ways: AI isn't just augmenting individual work. It's fundamentally changing the structure of how work gets done.

[00:00:26] Liz Jamieson: It's not AI replacing humans, it's AI augmenting humans so that humans can do the really important critical human work, have taste and judgment and keep that work that's really close to your competitive edge with a human.

[00:00:41] Amy J Springer: But here's what she told me next that changed how I think about this: 

[00:00:45] Liz Jamieson: So org design is everybody's job and it will become that point where humans are overseeing agents. We are the human bosses of the agents, and everyone is an org designer in your own little space, even if you're an individual contributor, you'll need to design your organization of agents.

[00:01:04] Amy J Springer: The challenge…?

Most people don’t have experience in how to structure that team, delegate to it, or…. measure its performance.

That's not a technology problem. That's an organizational design problem.

Today, we're talking about why one of the most overlooked skills in the AI era isn't coding or prompt engineering—it's org design thinking. And why every person in your company needs to learn it.

I am Amy Springer, and this is the Org Design Podcast.

Let's start with what's actually happening inside organizations right now.

You've got a marketing manager who used to write social posts herself. Now she's managing a team of AI agents all with different tasks in research, planning, writing and scheduling.

You've got a financial analyst who used to spend 60% of time pulling data—and now spends his time on interpretation and recommendations based on the analysis from his AI team.

You've got a customer service lead who's orchestrating a team of chatbots, escalation agents, and knowledge base agents.

They're all managing teams now. And the skills that make someone good at their individual craft don't automatically make them good at managing a team—whether that team is human or AI.

[00:02:12] Liz Jamieson: It's that rule of 90% planning, 10% execution with creating an agent as well. 

[00:02:18] Amy J Springer: That 90%? 

That's not a tech tip. That's a leadership principle.

Because here's what I realized, and maybe you're realizing it too as you're listening:

Planning how an agent should work—that's the same skill as planning how a team should work. Delegating to an AI? That's delegation. Reviewing an agent's output and giving it feedback? That's management.

We're not just getting new tools.

We're stepping into new roles.  

And here's the thing: most individual contributors have never had to think this way. They've never designed a role. They've never created a delegation framework. They've never built performance metrics.

But now they have to.

Because if you don't intentionally design how your AI agents work, they'll create chaos. They'll duplicate effort. They'll work at cross-purposes. They'll create more work than they save.

Just like a poorly designed human organization.

The difference is, AI moves faster. So a poorly designed AI team creates chaos at scale.

[00:03:13] Amy J Springer: So what does good organizational design look like when you're managing both humans and agents?

I found four principles that leaders are using to structure these hybrid teams.

Remember Liz's 90/10 rule? Here's how it applies to org design.

Before you bring in an AI agent, you need to map your organizational structure:

1. What work exists? (Not what's in job descriptions—what actually happens)
2. Who's responsible for what? (Clear ownership)
3. Where are the handoffs? (Human to human, human to AI, AI to AI)
4. What decisions require human judgment? (Escalation paths)
5. How do we measure value creation? (Not activity, but outcomes)

The 10%—selecting and implementing the actual AI tool—becomes straightforward once you've done the design work.

Most organizations do this backward. They buy the tool, push it out, and then wonder why it doesn't integrate cleanly. It's because they never designed the role it would play.

Josh Epperson from Amazon said:

[00:04:11] Josh Epperson: encourage daily use of it. Make it really personal,

[00:04:16] Amy J Springer: Daily. That's the key word.

You don't hire someone and then check in quarterly. You give daily direction, you review work, you provide feedback, you adjust course.

The same applies to your AI agents.

[00:04:27] Mike Montague: There's no barriers. The variable cost of creating content has essentially dropped to zero. 

[00:04:33] Amy J Springer: In org design terms, this means the traditional constraints on span of control have changed.

A manager could typically handle 5 to 10 direct reports. But if those "reports" are AI agents? You might manage 20, 30, or 50 agents simultaneously.

That changes everything about organizational structure.

[00:04:52] Mike Montague: AI is gonna hit marketing first or the hardest, but we're prepared for that part of the change management. My advice for people that are not is just marketing and AI needs to be a team sport.

[00:05:07] Amy J Springer: They're thinking through:

- Which roles become primarily agent management roles?
- Where do we need deeper human expertise vs. broader agent coordination?
- How do we structure teams when output capacity increases 10x?

These are classic org design questions. But they require rethinking all your assumptions about team size, hierarchy, and role definition.

[00:05:28] Josh Epperson: I think a more holistic set of metrics are gonna help us stay in better step with who we're working back from, and it's gonna ultimately make us quicker because we're gonna build machine learning algorithms that we can trust based on the signals that we're seeing.

[00:05:47] Amy J Springer: Holistic metrics for hybrid teams look different than traditional productivity metrics.

You need to measure:

- Team output: What's the combined human + AI team producing?
- Human value-add: Where are humans spending their time? Is it n high-judgment work?
- Agent effectiveness: Are agents handling routine work autonomously?
- Handoff quality: How smooth are the human-AI transitions?
- Customer outcomes: Is the end result better?

Notice what's not on that list: Individual productivity metrics.

In a hybrid organization, measuring individuals in isolation makes no sense. You're measuring the performance of the team—the human-AI unit.

[00:06:31] Gary Cohen: So now with AI, I think it's lowering the cost to put out software. It helps in efficiency with lots of other things, too. But, if it lowers the cost of doing it you're going to see more people doing it. And therefore, I think it raises the bar on what customers will expect

[00:06:51] Amy J Springer: In an organization where agents handle execution, humans need crystal-clear roles.

But here's where most organizations fail: They don't explicitly define these roles. They assume people will figure it out.

That creates role ambiguity. And role ambiguity destroys organizational effectiveness.

Every person needs to understand:

- What am I uniquely responsible for?
- What are my agents responsible for?
- Where do we collaborate, and how?

Without this clarity, you get:

- Humans doing work agents should do (inefficiency)
- Agents making decisions humans should make (quality problems)
- Gaps where no one—human or agent—is responsible

Traditional organizations are rigid. Restructuring takes months. Adding headcount takes quarters.

[00:07:34] Rhonda Frith-Lyons: AI is an enabler of being able to pivot. Being able to understand in the marketplace where your unexpected disruptors are going to come around the corner.

[00:07:44] Amy J Springer: when everyone manages teams of agents, you can reorganize rapidly:

- Need more capacity in customer service? Redeploy agents.
- New market opportunity? Spin up specialized agent teams.
- Competitor launches new feature? Reallocate AI resources within days.

This organizational flexibility is only possible if you've designed your human-AI teams with flexibility in mind.

That means:

- Modular agent design (reusable, not custom one-offs)
- Clear role boundaries (easy to reassign)
- Documented workflows (transferable knowledge)
- Standardized management practices (consistent across teams)

The organizations that win won't be the ones with the most AI. They'll be the ones that can reorganize their human-AI teams fastest in response to market changes.

That's organizational agility powered by good org design.

Alright, you understand the principles. Now let's talk about actually building and leading hybrid organizations.

The Management Development Gap

Here's an uncomfortable truth: Most people have never managed anyone. Now they're managing teams of agents.

This requires a massive organizational development effort. 

These are all management competencies. 

Now? Everyone needs them. Fast.

The change management challenge isn't just "people are afraid of AI." It's "people don't know how to be managers, and now they have to be."

Let's make this concrete. Three things you need to do this week to prepare your organization for the reality that everyone is becoming a manager.

ONE

Take your org chart.

Now add every AI tool or agent your team is currently using.

Draw the reporting relationships.

Who's managing which agents? Where are there gaps? Where is there duplication?

You'll immediately see:

- orphaned tools, no one is managing them
- People managing agents without realizing it (no intentional design)
- Work that's falling through the cracks between human and AI

This is your baseline. You can't redesign what you can't see.

TWO

Pick one team or workflow. Sit down and explicitly define:

For each human:
- What are you uniquely responsible for?
- What decisions are yours alone to make?
- What's your desired output?

For each agent:
- What's the specific role?
- What inputs are needed?
- What decisions can be made autonomously?
- Where does human escalation occur?

Write this down. Make it explicit.

This is a RACI exercise for the AI era. And it will reveal all the assumptions and ambiguities that are currently slowing you down.

THREE

In your next all-hands or team meeting, introduce this concept: "Everyone is becoming a manager of agents."

Then teach basic org design principles:
You don't need to be an org design expert to start this conversation. You just need to acknowledge that this is now a core competency for everyone.

Consider bringing in resources—workshops, training, reading materials—on organizational design fundamentals.

And here's what to STOP doing:

Stop treating AI adoption as an IT initiative.
Stop assuming people will naturally figure out how to manage agents. 
And stop thinking about AI tools in isolation.

Think about them as team members that need to be integrated into your org structure.

[00:10:47] Amy J Springer: Here's my final thought.

The organizations that succeed in the AI era won't be the ones with the most advanced technology. They'll be the ones with the best organizational design.

Because when everyone's managing a team of agents, organizational clarity becomes your competitive advantage.

Clear roles. Clean handoffs. Intentional structure. Rapid reorganization capability.

And it all starts with recognizing that we're not just augmenting individual work—we're fundamentally redesigning how organizations function.

[00:11:18] Liz Jamieson: agentic employees need to be treated almost the same as human employees in terms of all of that HR support. I think there's a movement now for that function or a new function to be _intelligence resources_. And that is the mix of intelligence of humans and also intelligence of agents.

[00:11:41] Amy J Springer: The question is: Are you designing intentionally, or are you letting it happen accidentally?

If this reframed how you're thinking about AI implementation, share it with your leadership team. This is a conversation you need to have together.

We hope you've enjoyed this episode, bringing together some of our previous guests on AI and what it means for org design in your organization. We look forward to sharing another episode like this with you soon.

I'm Amy Springer. See you next time.

Functionly is designed to empower organizations to effectively manage their evolving structures as they integrate AI into their workflows. By providing a clear framework for mapping roles and responsibilities, Functionly enables teams to define what work exists, establish ownership, and ensure smooth handoffs between human and AI agents. This clarity is crucial for avoiding role ambiguity and enhancing organizational agility.

With tools for visualizing your organizational design, Functionly supports leaders in creating a hybrid workforce where both humans and AI can thrive, allowing for rapid reorganization in response to market changes and ensuring that teams remain aligned and effective in achieving their goals.