7 principles for redesigning work in an AI-powered enterprise

From my live broadcast with Josh Bersin last week. What to do, not just what to know.

"This isn't like buying Workday and hoping everyone becomes more productive. This is full business reinvention." - Josh Bersin

Let us stop treating AI like a tech initiative. Let us start treating it like what it actually is. A total restructuring of how work gets done.

Here are 7 tactical shifts to redesign your workforce model. Built from my live session with Josh Bersin.

1. You need critical infrastructure for work, not a static org chart

Not a static job architecture. Not an org structure. Something dynamic that reflects the work being done.

The problem:

Org charts tell you who reports to whom. They tell you nothing about how work actually flows. What is broken. Where value gets stuck.

Josh shared an example of a company with 100,000 employees and 65,000 job titles. After we mapped the work: just 3,000 unique jobs. The rest was redundant noise, legacy titles, and work no longer done.

You manage people by outdated roles. But people do not do roles. They do work. That work evolves daily. You need a live operating system. One that maps tasks, skills, and outputs in real time.

What to do:

• Do not start from scratch. Start with real data. Reejig automatically delivers a live feed of tasks and skills.
• Use that data to update your outdated job architecture. Build what we call a Work Architecture. A dynamic system that reflects how work actually happens.
• This is not about removing structure. It is about redesigning it. To manage risk, regulatory requirements, and real-time change.
• Focus on what is getting done. Track how work moves across teams, roles, and systems. Regardless of titles.
• Choose one high-impact function to get started. Customer support, marketing, or claims.

This is your Work Operating System. A real-time map of how your organization runs. Where you find velocity, automation, and reinvention.

2. Use tasks as your source of truth

The problem:

You are trying to automate work based on roles and skills. AI does not work that way.

AI automates tasks. It needs granularity. "Skills" taxonomies do not tell you what the person is doing or how often. Roles are too vague. Job descriptions are often pure fiction.

"People have skills. Jobs and work have tasks." - Siobhan Savage

What to do:

• For each job, list 10 to 20 repeatable tasks
• Tag each with effort level, business impact, and cost per task
• This gives you the basis to identify automation opportunities, spot task duplication across roles, and score AI readiness

Task-level visibility equals transformation visibility.

You will not need to do this manually. Reejig automatically gathers this data. Your only job is to validate what is already there.

3. Define what AI-ready work actually looks like

The problem:

Most "AI pilots" start with demos. Not with real use cases grounded in work.

Josh shared the story of a bank. Its biggest bottleneck was not in tech or headcount. It was in account opening. A process nobody had flagged.

Every enterprise is deploying AI. Almost none can see the work they're deploying it into.

What to do:

• Start with one problem domain. Onboarding, approvals, or reporting.
• Break it into tasks.
• Use this filter:
• Does this cost us money or cause pain?
• Is this task very repetitive?
• Is AI mature enough to take this task?
• What is the ROI when we reinvent this task?

Now you are not asking "Can we use AI?" You are asking "Where does AI deliver value today? What work do we redesign to make that happen?"

4. Advance beyond assistance by reengineering design

The problem:

Most organizations are stuck at "AI as Copilot" (Level 1). They think they are innovating.

Josh's AI Maturity Model is clear:

1. Personal Productivity
2. Task Automation
3. Workflow Reengineering
4. Autonomous Agents

The real value, 100% to 300% gains, lives in Level 4. You cannot leap there with fragmented processes and siloed technology.

What to do:

• Reejig automatically maps your workflows end to end. Sourcing to hiring to onboarding.
• It identifies all handoffs, delays, and duplicated tasks.
• Reejig tells you exactly where to eliminate friction. Not just speed up parts of the process.

AI maturity is not a technology investment. It is a design discipline.

5. Appoint someone to own the work

The problem:

CHROs own people. CIOs own systems. COOs own throughput. No one owns the work itself.

Without a single point of accountability, your AI efforts stay siloed and tactical.

What to do:

• Appoint a Work Design Lead (title optional for now)
• Their job:
• Own task-level design and governance
• Bridge HR, Ops, and Tech
• Run the AI-readiness map across the organization
• Plug them into your transformation office. Or report them to your CEO.

This is the start of the Chief Work Officer role. The organizations who create it first will lead the field.

6. Rethink career progression: from fixed paths to fluid pivots

The problem:

Traditional career ladders no longer reflect how work evolves.

Every time AI reshapes or removes a task, the role itself transforms. Instead of mapping fixed career paths, we now look at reskilling or pivoting. Based on the changes to the role we just redesigned.

I used the "Jenga metaphor" in our call. Remove one task, and the entire structure changes.

What to do:

• Stop mapping roles. Start mapping task adjacencies.
• Example: A marketing coordinator starts doing campaign tracking. AI automates that task. So what is next? Instead of climbing a traditional ladder, we pivot. Toward adjacent skills like campaign design or customer analytics.
• Every redundancy becomes an opportunity to reskill, redeploy, or redesign work in real time.
• Just-in-time readiness is not just efficient. It is how we build workforce adaptability.

7. Reinvention over efficiency

The problem:

Too many leaders still treat AI as a way to do the same work faster.

True transformation happens when you redesign the work itself. And rebuild the system to continuously evolve.

As I said in the broadcast:

"If I were a CEO and you couldn't help me do this, I'd find someone who could."

If I automate chaos, I just scale the chaos.

What to do:

• For any AI use case, also ask:
• What tasks are being removed?
• What new tasks will emerge?
• What will this mean for team capacity, structure, and skills?
• Build a reskilling plan that aligns with that evolution. Not just generic upskilling.
• Reward teams for removing work. Not just performing it.

AI does not just change productivity. It changes purpose.

Final word

Jobs still exist. They are no longer enough. We need new infrastructure built for the AI-powered workforce.

If you cannot see it, map it, and orchestrate it:

1. You cannot lead it
2. AI will amplify your problems. Not solve them.

But if you can, you will reinvent how your business runs. On a foundation designed for now.

Map. Analyze. Build. Run. Measure. Log. Update. That's Reejig.

Book a demo to see how the Work Operating System redesigns work at the task level.

Siobhan 💜