Work redesign for AI: Why transformation has to start with the work

The promise of AI is enormous. Across industries, leaders reference to it as a generational shift: a once-in-a-lifetime productivity opportunity, comparable to major advances in technology such as electrification and mechanization.

The productivity promise is also the basis of a huge investment case. Organizations are investing in AI because they expect it to deliver a step change in performance.

Capitalizing on a generational opportunity requires disciplined transformation action

As Kevin Breskvar, Lead Product Manager at Orgvue, puts it: “It’s not enough to do a “back-of-the-napkin” exercise, eliminate roles in the name of efficiency, and call it a day. That may create a short-term improvement in the P&L, but it’s unlikely to deliver the transformational outcomes organizations are hoping for.”

The challenge is more complex than that. AI impacts work, which in turn impacts roles, structure, and then workforce size. If organizations want to deliver the full value of AI, they first need to understand how it changes work, then redesign the roles needed to perform that work, and only then redesign the organization around it.

Understand the work to plan the future workforce

As Steve Kelly, VP User Growth & Community at Orgvue, says: “Workforce begins with work. It’s called a workforce for a reason.”

The work changes first, then the roles needed to perform that work change, then the structure supporting those roles changes. This matters because many organizations still start with job titles, headcount, or organizational charts rather than understanding the work itself.

Most modern roles consist of many different activities, accountabilities and judgments, and AI typically affects parts of the work rather than all of it. This is why organizations need to understand the fractional impact of AI on work and how that affects roles.

Another important consideration is that even when work is automated, accountability remains. Steve made the point that accountability for outcomes stays with human beings because “you can’t send AI agents to jail. They can’t be sanctioned by regulators, and they can’t be fired”. AI may execute parts of the work, but accountability for outcomes still lies with humans.

So, planning the workforce is becoming more complicated because organizations are no longer balancing just labor costs. They also need to understand the redesign of work and roles, the level of oversight required, and who will be accountable.

Without a process-led view of the organization’s current state and the ability to model how changing work impacts the workforce, transformational change will come with material consequences, significant financial cost, and excessive human impact.

In practice: Connecting work to the workforce and the skills needed to perform

Work redesign for AI requires a connected view of the work, the workforce, and the skills needed to perform. In practice, this means understanding the current work, modeling how that work changes, redesigning processes, redesigning roles and accountabilities, and then translating those changes into workforce planning and organizational design decisions.

Activity analysis: understanding the organization’s current state

The goal is to visualize the organization in a process-led way rather than a purely structural way. Activity analysis helps organizations understand how work is performed and distributed across individuals & teams.

That visibility matters because every organization is different. Organizations need to plan based on their own reality if they want to avoid unintended consequences.

A thorough activity analysis also informs the economic case for change by equipping business leaders with the ability to evaluate the cost of compute against the cost of people.

Activity modeling: seeing impact before making decisions

Once leaders understand today’s work, activity modeling allows them to model changes to this work and see how the workforce is impacted. Which roles are impacted? How many positions are affected? Which people are involved? 

This becomes particularly important when AI’s impact is fractional rather than complete. Viewing roles as a collection of activities rather than a single unit makes it possible to understand the impact of making changes to these activities (e.g. automation). It enables organizations to see impact before making decisions.

Value chain mapping and process redesign: redesigning workflows

Kevin highlighted another practical challenge: As organizations redesign work, it often becomes easier to think in terms of processes and sub-processes, as opposed to individual activities. The challenge is that many organizations don’t have that information readily available.

Work may comprise a list of activities, but that doesn’t make it easy to see how work flows through the organization. Value chain mapping groups activities into processes and sub-processes, so organizations can understand how work is carried out in a more intuitive manner.

It provides a way to see how activities can be grouped logically, and how processes should flow once AI changes the way work is performed.

The objective is not simply to identify what can be automated, but to redesign how work gets done.

Role redesign

Once work has been redesigned, roles need to be redesigned around it.

This means connecting roles to accountabilities and avoiding common role-design misfires: overloaded roles, fragmented responsibilities, duplication of effort, and inefficient oversight.

The practical questions are straightforward: Can the role be performed effectively? Will the work get done efficiently?

Visual role design helps answer those questions by showing all the roles holding accountability for each activity. This makes it possible to identify overloaded roles, duplicated effort, fragmented decision-making, and inefficient approval structures.

Role redesign is not simply about rewriting job descriptions, but about systematically assigning and redistributing accountabilities while ensuring the redesigned role is practical, coherent, and capable of delivering the required outcomes.

Workforce planning and organizational design: translating redesigned work into workforce decisions

Once work and roles have been redesigned, organizations can translate those changes into workforce demand. As Kevin explained, organizations can estimate how many of each role they will now require, based on the work that needs to be done, and use that information to inform organizational design and workforce planning.

This creates a link between work redesign and workforce decisions. Kevin and his team have long encouraged organizations to think about work before workforce design. But in reality, workforce planning and organizational design often begin with financial targets.

As an example, organizations will start with a cost reduction, or efficiency, objective and then attempt to redesign the workforce around that number. The danger is that the financial target may be achieved without fully understanding how the work will be performed, how customers will be affected, or how outcomes will be sustained.

Kevin hopes the fact that everyone agrees AI is changing work will force organizations to start with the work rather than jump straight to the workforce and organizational design.

Minimize the financial cost and human impact of change

There’s also a human reason to get this right. If the only way organizations can respond to AI-driven change is through a fire-and-hire approach, both the financial and human costs will be significant.

Some organizations are taking a different approach. Rather than default to replacing people, they begin by recognizing that employees accumulate knowledge, trust, credibility, and an understanding of how the organization works. And that experience has value. The cost of severance, recruitment, onboarding, and the time required for new employees to become effective is substantial.

If organizations reskill and redeploy people instead, they can manage change more efficiently with less disruption. But that requires visibility. Organizations need clarity on the skills they have today and the skills they’ll need tomorrow. They need a connected view of the work, the workforce, and the skills needed to perform it.

Kevin emphasizes the importance of prioritizing reskilling programs, which he believes have been largely overlooked so far. Understanding skills gaps shouldn’t be a side issue; it should be central to making AI transformation practical, sustainable, and responsible.

AI may have increased the urgency around workforce transformation, but the underlying discipline is not new

Orgvue conducted its first activity analysis in 2006, years before AI became a major workforce transformation topic. The company’s heritage is rooted in understanding work, connecting it to people and roles, and using that understanding to support data-driven organizational design.

Twenty years later, this discipline remains the same.

Organizations still need to understand how work gets done, how to allocate accountability, and how changes to work affect roles, people, and structure.

AI increases the urgency of those questions because it changes work at speed. But it doesn’t remove the need to answer them.

FAQ: Work redesign for AI

AI-enabled workforce

Understand how AI will change work, redesign roles and processes, and execute workforce transformation with confidence.