Stop Managing AI. Build Your AI Chief of Staff.

What I shared at DESHRM26 about applying familiar HR principles to a new way of working with artificial intelligence.

By Leilani Decena, SHRM-SCP | Founder, AI x HR Studios

What if I told you that HR professionals already know how to build an AI Chief of Staff?

No coding required. No complicated automation. And certainly no need to master every new AI tool that launches before you've finished your morning coffee.

Think about it.

We've spent our careers identifying workforce needs, defining roles, writing job descriptions, onboarding employees, setting expectations, assigning responsibilities, and evaluating performance.

So why are we treating AI differently?

That was the question I brought to the Delaware SHRM Conference on October 7, 2026, when I presented Stop Managing AI. Build Your AI Chief of Staff.

My goal wasn't to teach everyone how to build a sophisticated AI agent. It was to show how the skills HR professionals already use could help them approach AI more intentionally.

Instead of opening a chatbot whenever we need something, what if we created a defined AI support role that understood our priorities, followed our instructions, and helped us manage recurring work?

Not an employee. Not a replacement for human judgment. And not necessarily an autonomous agent.

A practical, organized way to work with AI.

Same HR discipline. New AI role.

HR Already Knows How to Do This

When an organization needs to hire someone, we don't begin by choosing a desk or buying software.

We start by identifying the need.

What work needs to be done? Why does the role exist? What responsibilities should it carry? How will we know whether it's successful?

Then we define the position, establish expectations, select someone with the appropriate capabilities, and provide onboarding.

And the process doesn't stop there.

We assign work, provide feedback, address problems, and evaluate results.

These are familiar HR practices, and they're surprisingly useful when applied to AI.

Consider the parallels:

Familiar HR Practice	AI Chief of Staff Equivalent
Identify a workforce need	Identify one task or bottleneck AI could help address
Write a job description	Define the AI support role and responsibilities
Recruit and select	Choose the appropriate AI capability
Onboard	Provide approved context, instructions, and examples
Assign work	Give AI a clear, bounded assignment
Coach and develop	Review outputs and refine instructions
Evaluate performance	Assess accuracy, reliability, usefulness, and time saved

Of course, AI isn't a human employee. It doesn't possess professional judgment, understand organizational culture as people do, or accept responsibility for its decisions.

But the underlying principles of clear expectations, appropriate resources, supervision, and feedback still matter.

That's where the AI Chief of Staff concept begins.

What Is an AI Chief of Staff?

Think about what a human Chief of Staff does.

They help coordinate priorities, prepare information, keep projects moving, and bring important matters to a leader's attention.

An AI Chief of Staff can support selected parts of that work, depending on the tools, permissions, and information available.

It might help you prepare for meetings, organize project updates, draft communications, structure recurring reports, or identify questions that need your attention.

But here's an important distinction I emphasized at DESHRM26:

The AI Chief of Staff is the role. A project can be its home base. Different AI capabilities help it do the work.

Your AI Chief of Staff doesn't have to be a complex autonomous system.

It may begin with a simple chat and a set of well-written instructions.

Over time, you might add a dedicated project, recurring tasks, defined workflows, or carefully governed agent capabilities.

The technology should fit the work.

Not the other way around.

First, Figure Out What You're Hiring It to Do

Before building anything, I encouraged participants to ask themselves a question:

Where am I the bottleneck?

Maybe you're spending hours preparing for leadership meetings.

Perhaps you're constantly following up on missing information.

You might be managing several projects, organizing research, drafting communications, or trying to turn scattered notes into meaningful action.

Start there.

And here's the part that can be surprisingly difficult.

Pick one responsibility. Not five.

We tend to get excited about what AI might do and immediately imagine everything we'd love to hand over.

But an unclear role with too many responsibilities is difficult to manage, whether we're talking about people or technology.

Instead, choose one manageable task with a useful, reviewable outcome.

For example:

"My AI Chief of Staff will help me prepare for my weekly HR leadership meeting."

That's specific enough to begin defining expectations and evaluating results.

Once the process works reliably, you can consider expanding its responsibilities.

Give Your AI Chief of Staff a Home Base

During the workshop, I introduced participants to the idea of creating a dedicated project or workspace called My AI Chief of Staff.

Think of it as the role's home base.

Depending on the AI platform, a project can help organize standing instructions, approved reference documents, examples, working preferences, and ongoing conversations.

Instead of repeatedly explaining your responsibilities or preferred communication style, you can maintain relevant context in one place.

But creating the workspace is only the beginning.

The next step should feel familiar to anyone who has ever onboarded a new team member.

You have to explain how things work.

Onboard Your AI Chief of Staff

During the DESHRM26 presentation, I introduced four areas to address when onboarding an AI Chief of Staff.

1. Know Me

Explain your role, responsibilities, priorities, and what success looks like.

What kind of support do you need? What tasks occupy your time? What should AI help you accomplish?

2. Know How We Work

Describe relevant processes, terminology, communication preferences, recurring responsibilities, and expectations.

If you prefer concise executive briefings instead of lengthy reports, say so.

If your team uses specific templates or workflows, provide approved examples.

3. Know What to Trust

Identify approved documents, reliable reference materials, and information sources.

Tell AI when it must rely only on supplied information and when it should flag missing or uncertain details.

Only provide information that your organization's policies and the tool's permissions allow.

4. Know When to Stop

Set boundaries.

What should AI avoid? Which matters require approval? When should it flag concerns or escalate a question?

For HR professionals, this is particularly important when work involves employee information, workplace policies, compliance issues, or decisions affecting people's opportunities and livelihoods.

A job description defines the role.

Onboarding explains how your world works.

And once that foundation is in place, it's time to assign meaningful work.

Give It Its First Assignment With the DECENA Method™

One of the most important skills when working with AI is learning how to communicate the assignment.

That's why I developed the DECENA Method™, a six-step framework for briefing AI.

It provides a repeatable structure:

D — Define the Role: Who should AI act as?

E — Explain the Task: What exactly should it accomplish?

C — Context: What background does it need?

E — Evidence: What information should it rely on?

N — Name Boundaries: What must it avoid, protect, or escalate?

A — Ask for Action: What should the result look like?

Let's apply it to our AI Chief of Staff.

Imagine you're an HR manager preparing for Monday's leadership meeting.

Instead of typing, "Help me prepare for my meeting," you might provide this briefing:

Define the Role: Act as my AI Chief of Staff supporting HR operations.

Explain the Task: Help me prepare for Monday's leadership meeting.

Context: Our discussion will focus on staffing updates, manager concerns, and priorities for the coming week.

Evidence: Use only the approved agenda, previous meeting notes, and current updates I provide.

Name Boundaries: Flag missing information. Don't invent facts, interpret legal requirements, or make employment decisions.

Ask for Action: Prepare a one-page briefing organized into key issues, decisions needed, questions to ask, and recommended follow-ups.

Now AI has a clearer assignment.

It knows the role it's supporting, the information it should use, the boundaries it must respect, and the result you're expecting.

The goal isn't to have AI make leadership decisions.

It's to help prepare information so you can make those decisions more effectively.

And the DECENA Method™ isn't limited to one platform or a single prompt. It can help you structure assignments within projects, scheduled tasks, workflows, and appropriately configured agents.

Brief the work clearly before you delegate it.

Don't Build an Agent When a Prompt Will Do

There's a lot of excitement surrounding AI agents, and some of it is warranted.

But more sophisticated technology isn't automatically better.

Sometimes you simply need a well-written prompt.

During the workshop, I walked participants through five ways of working with AI.

AI Capability	When It May Help
Chat	You need help with one task right now
Project	You need ongoing context and approved reference materials
Workflow	You have a repeatable process with defined steps
Scheduled Task	You want supported work to happen at a particular time
Agent	You need AI to work through multiple approved steps toward a goal

These capabilities differ across platforms, and their availability depends on the tools, plans, permissions, and integrations involved.

An agent may be appropriate when work requires several steps, permitted tool use, and defined escalation procedures.

But it also introduces additional considerations around access, accuracy, security, and oversight.

Before choosing a capability, ask:

What's the simplest reliable way to accomplish this task?

That question can save an organization considerable time, complexity, and unnecessary expense.

And it brings us back to a central lesson from the presentation.

Don't build an agent when a prompt will do.

Coach Before You Automate More

One of my favorite analogies from the presentation involved a piano lesson.

Imagine sitting a beginner at a piano and expecting a flawless performance of a complicated composition.

You wouldn't.

You'd start with a manageable exercise, listen to the result, identify mistakes, provide guidance, and gradually increase the difficulty.

Working with AI also requires review and refinement.

Before expanding your AI Chief of Staff's responsibilities, evaluate the work it's already doing.

Was the information accurate?

Did it follow the instructions?

Did it respect the boundaries?

Was the result useful?

Did it actually save time or improve the process?

If the output falls short, examine what happened.

Maybe the instructions were unclear. Perhaps the source material was incomplete.

Maybe the task requires a different approach, or simply isn't appropriate for AI.

Adjust the process and test it again.

And don't confuse an impressive-looking output with a reliable one.

Sometimes the best improvement isn't adding automation.

It's fixing the underlying workflow.

Don't automate a bad process faster.

Try It Yourself: Your First AI Chief of Staff

You don't need an elaborate system to get started.

Here's a simple instruction you can adapt in an approved AI tool.

Your AI Chief of Staff Starter Instruction

You are my AI Chief of Staff. Your role is to help me organize priorities, prepare work, coordinate recurring responsibilities, and identify matters that need my attention.You support my professional judgment rather than replacing it.My role is: [Describe your position and responsibilities.]My first priority is: [Describe one task you want help with.]Success looks like: [Describe the result you want.]Use only information I provide or sources I explicitly approve and that you are authorized to access. Flag missing information and uncertainty. Don't invent facts, make decisions on my behalf, or take external actions without my approval.Present your work in a concise, actionable format.Before beginning, ask me up to five questions necessary to understand my first assignment.

Start with fictional, public, or appropriately sanitized information while you're learning.

Follow your organization's requirements before using confidential or restricted materials.

The instruction gives you a starting point.

From there, you can refine the role, add appropriate context, and evaluate how well the AI supports your work.

What DESHRM26 Reinforced for Me

My career has taken me through television newsrooms, communications, human resources, and organizational learning.

Those environments may seem very different, but they share something important.

Clear communication and coordination matter.

In television, producers, writers, reporters, editors, and technical teams work together to get a broadcast on the air. Everyone needs to understand the assignment, the deadline, and their responsibilities.

In HR, we define roles, develop people, establish expectations, and design processes that help organizations function.

And now, as I work with professionals exploring artificial intelligence, I see those same principles becoming valuable in a new way.

The technology has changed.

The importance of clarity, accountability, communication, and thoughtful delegation hasn't.

That's what I wanted participants at DESHRM26 to recognize.

We aren't starting from zero.

We already bring valuable skills to this work.

I didn't want attendees to leave simply understanding the concept of an AI Chief of Staff.

I wanted them to have the beginnings of a practical system: a defined role, an onboarding approach, a first assignment, and a way to evaluate the results.

Because successful AI adoption isn't measured by how many tools we introduce.

It's measured by whether those tools help people do meaningful work more effectively and responsibly.

The Future of Work Starts With Better Decisions

AI will continue to evolve.

New tools will emerge. Capabilities will expand. The possibilities for automation will grow.

But we don't have to use every capability simply because it exists.

We can make thoughtful decisions about what deserves our attention, where technology adds value, and which responsibilities must remain with people.

That's the approach I bring to AI x HR Studios.

Helping leaders, teams, and organizations move beyond experimenting with technology toward creating better ways of working.

So before you ask AI to do more, consider one question:

If you could hire someone to support you tomorrow, what would you want that person to take off your plate first?

Start there.

Define the role. Explain the work. Provide the context. Establish boundaries. Review the results.

You don't need a complicated AI system to begin.

You need a clear purpose and a manageable first assignment.

Because the future of work isn't simply about giving AI more to do.

It's about designing better ways for people and technology to work together.

People lead. AI supports. Better work follows.

Bring Practical AI Training to Your Organization

At AI x HR Studios, I help leaders, professionals, and organizations develop practical AI capabilities, improve workflows, and redesign work for a human + AI world.

My interactive workshops, speaking engagements, and consulting services focus on helping people move from AI curiosity to confident, responsible application.

Whether you're exploring an AI Chief of Staff, strengthening your team's AI skills, or identifying opportunities to work more efficiently, the goal is the same: meaningful improvements that support people and organizational outcomes.

Learn more at www.aixhrstudios.com.

Leilani Decena, SHRM-SCP, holds the SHRM AI+HI Specialty Credential and is the founder of AI x HR Studios. Her background spans award-winning television journalism, human resources, communications, and organizational learning.