AI work transformation is the leadership test of the decade, and most organizations are still treating it like a technology upgrade. The enterprises making real progress are the ones who have moved beyond AI deployment into a deliberate redesign of how work actually runs. Michael Fraccaro, former fellow and Chief People Officer at Mastercard, is working with CHROs, boards, and executive teams across industries on exactly this challenge.
This conversation, from Reejig's Work Blueprint series, covers the three strategic buckets every CEO should be deciding between, why HR is now in the front row of AI transformation, how to protect the junior workforce pipeline, and why building internal capability is not optional.
The biggest mistake organizations made in the early wave of AI was treating it like a technology upgrade. Buy the licences, give people access, wait for productivity to go up. That framing is now visibly broken.
The organizations making real progress have moved to a different question: are we using AI to cut costs, to improve our products and services, or to fundamentally reimagine how value is created? These are not the same question, and the answer determines everything from who owns the initiative to how success is measured. Every enterprise is deploying AI. Almost none can see the work they're deploying it into.
For the past two years, AI transformation sat with the CIO. That has shifted. The work of redesigning how an organization operates belongs to the people function, and most executive teams have now recognized it.
What changed the conversation is the realization that deploying AI without work redesign produces chaos, not productivity. When everyone is given access to AI individually and told to figure it out, organizations do not move faster. They move slower.
Fraccaro described the shift he has seen: HR is now a "front-row seat" in developing and implementing workforce transformation. That means owning the question of how work happens today, which tasks should be AI-led versus human-led, and what the organization needs to look like on the other side.
Job architecture was built for a world of stable, siloed roles. It is not built for agents, dynamic workflows, or task-level AI deployment. From Job Architecture to Work Architecture is not a cosmetic update. It is the foundational shift that makes AI transformation possible.
Work Architecture maps every department, job, level, role, workflow, task, and subtask, structured for both humans and agents. It is dynamic by design, because the work keeps changing as agents get stronger. Job architecture describes what people are called. Work Architecture describes what actually happens.
The task-level exercise Fraccaro uses with leadership teams, taking a single role and classifying every task as AI-only, human-only, or hybrid, illustrates the scale of the problem. Do that across every job family in a large enterprise, and the scope of redesign becomes clear. You cannot deploy AI into work you cannot see.
One of the most urgent and underappreciated risks in AI work transformation is what happens to junior workforce pipelines when entry-level analytical work is automated. Many organizations have already started pulling back on graduate and intern programs. Most have not thought through the five-year consequence.
The Reejig perspective is that the question is being framed too narrowly. Rather than deciding whether to hire entry-level people at all, the better question is: what if agents were wrapped around those people to make them productive faster? Instead of three to five years of structured training, a new hire could operate at a higher level much earlier, because agents are handling the repetitive execution while the person develops judgment, craft, and business context. The goal is to protect the people bench while accelerating how quickly people become genuinely valuable.
The organizations that are outsourcing their AI work transformation to consultants are solving the wrong problem. Work redesign is not a one-time project. It is an ongoing operational capability that compounds over time.
Fraccaro framed this as an oxygen mask problem for HR leaders: fix your own function first. If HR is advocating for work redesign across the business but has not redesigned its own workflows, the credibility is not there. Leading by example means HR demonstrating task-level AI deployment in its own processes before asking the business to follow.
The internal team model that is emerging in large enterprises includes a small workforce innovation group, working alongside HR business partners, that maps AI opportunity, models agent cost against people cost, and drives redesign at the department level. This is not a consulting engagement. It is a permanent capability.
|
CHRO Focus |
CIO Focus |
Shared Outcome |
|
Task-level work design: what is AI-led, human-led, or hybrid |
Agent inventory and governance: what is approved and deployed in the environment |
A unified map of which workflows to redesign, in what order, with what guardrails |
|
Junior workforce pipeline strategy and entry-level role redesign |
Agent capability assessment: what analytical tasks agents can reliably perform today |
Entry-level roles rebuilt around human judgment and agent support, not replacement |
|
Internal HR capability building and workforce innovation team |
Engineering and integration support for AI workflow deployment at scale |
A permanent, internal work redesign function that does not depend on external consultants |
|
Employee communication, trust, and adoption at the role level |
Change logging and audit trail for every workflow and agent modification |
AI adoption that employees understand, trust, and can operate within |
Why is AI transformation described as a leadership test, not a technology problem? AI work transformation requires decisions about value creation, risk tolerance, workforce strategy, and organizational culture, none of which are technology decisions. The organizations treating AI as a technology upgrade are deploying AI into unchanged workflows and finding no return. The leadership test is deciding what work to redesign, in what order, and how to bring the organization through continuous change without losing people or expertise.
What are the three strategic buckets every CEO should decide between? The three buckets are: using AI for efficiency and cost reduction, using AI to improve products and services, and using AI to fundamentally reimagine how value is created. Each requires a different program, different leadership ownership, and different success metrics. Organizations that conflate them end up optimizing for the wrong outcomes, often short-term cost cuts at the expense of long-term capability.
What is Work Architecture and why does it replace job architecture? Work Architecture is the entity model that replaces static job architectures: every department, job, level, role, workflow, task, and subtask, structured for both humans and agents. Job architecture describes what people are called. Work Architecture describes what actually happens, at the task level, dynamically updated as work changes. It is the foundational map that makes AI deployment decisions possible.
Why is the junior workforce pipeline an AI transformation risk? When entry-level analytical work moves to agents, organizations stop hiring for those roles. Within five years, the pipeline of people with the foundational experience needed to become senior leaders and domain experts has dried up. The answer is not to keep hiring for roles exactly as they were, but to redesign those roles so that new hires build judgment and craft faster, with agents handling execution while people develop the capabilities that cannot be automated.
Why should enterprises build internal work design capability rather than hiring consultants? Work redesign is not a bounded project with an end date. As agents get stronger, workflows change again. An organization that relies on external consultants for each wave of redesign will always be behind, always paying for the same capability, and never building the institutional knowledge of how its own work runs. Internal capability compounds. External dependency does not.
How should organizations measure whether AI transformation is working? The right measures are changes to actual work: time saved per task, throughput per workflow, decisions accelerated, capacity redirected to higher-value activity. Licence adoption, prompt volume, and usage metrics tell you whether people are using a product. They do not tell you whether work changed. AI ROI is proved through outcomes, not consumption.
AI work transformation is not a project with a completion date. It is a permanent shift in how enterprises operate, and the organizations building internal capability to manage that shift continuously are the ones that will compound their advantage. The starting point is always the same: understand how work actually runs before deploying anything into it.
Book a demo to see how Reejig's Work Operating System gives your team the task-level visibility to start redesigning work with confidence.