This is a constructed example, not a real company or a documented case study — I'm applying the task/orchestration/judgment framework from the last piece to a plausible scenario. Where I mention Copilot or agents, I'm describing generic, illustrative capabilities rather than asserting specific verified product features.
Call it Kordana IT Services — a fictional global IT service provider, roughly 500 people in HR supporting an 18,000-person workforce across a dozen countries. The HR function is organized the way a lot of enterprises look right now: a shared service center (outsourced, handling tier-1 Q&A — leave balances, policy lookups, payroll questions), and an internal central HR organization split into People Strategy, Talent Development, Talent Acquisition, HR Technology & Analytics, HR Operations, and a dedicated Employee Relations & Change team — plus a layer of HR business partners embedded with the business units.
Mapping the tasks, not the org chart
The instinct in most transformations is to ask which of these six areas "gets automated." That's the wrong question, for the same reason it was the wrong question at the team level in the last piece. Each of these areas is a bundle of tasks with a different mix of green-light, orchestration, and judgment work — and the mix is genuinely different function to function.
HR Operations (contract changes, onboarding paperwork, data entry into the HRIS) is where the green-light zone is largest. A Copilot-style agent handling structured data entry, drafting standard contract letters, and pre-filling onboarding checklists is a reasonable fit — this is the closest thing to the customer-support example from the last article. Talent Acquisition has a large green-light layer too (screening against defined criteria, scheduling, drafting job ad variants) but a judgment core that doesn't shrink: deciding who's actually right for a role involves reading things a resume-matching tool can't. Talent Development splits similarly — content curation and administrative tracking are automatable; deciding what capability the organization actually needs to build next is not.
People Strategy and the Employee Relations & Change team look almost the opposite of HR Operations: thin green-light layer, most of the work sitting in judgment and, especially for the Employee Relations & Change team, in relationship and trust-building that no tool touches. HR Technology & Analytics is unusual — it's the team that builds and maintains the orchestration layer for everyone else, so its own task mix shifts toward designing and monitoring rather than manually pulling reports.
What the three layers look like, function by function
Execution absorbs most of HR Operations' current headcount and a meaningful slice of Talent Acquisition and Talent Development admin work — this is where the shared service center's remit could plausibly expand, taking on some of what central HR currently does manually, precisely because it's already structured to handle high-volume, well-specified work.
Orchestration is the layer this case study makes concrete in a way the smaller team example couldn't. HR business partners are a natural fit for this layer almost by design — they already sit between policy and the specific mess of a real business unit, which is exactly the job of catching when an AI-drafted answer is technically correct but wrong for this team's situation. The redesign here isn't "replace HRBPs with a chatbot." It's formalizing what a good HRBP already does — reviewing AI-drafted guidance before it reaches a manager, knowing which cases need escalation to People Strategy or legal — and giving them the tools and training to do it faster across more of the business, rather than spreading their time thin across manual case handling.
Judgment and accountability stays smallest and gets more concentrated: People Strategy decides what the organization's actual talent and skilling priorities are; the Employee Relations & Change team owns the relationship and the negotiation, full stop — nothing about AI assistance changes who's accountable in that room. This is worth being explicit about, particularly in a German or EU context: co-determination rights around the introduction of systems that could monitor or evaluate employee performance are a real legal and negotiated matter (in Germany, under the Betriebsverfassungsgesetz, works councils typically have co-determination rights over the introduction of such technology) — which means this team's role in the transformation isn't just internal change management, it's the entity that has to negotiate the terms of the rollout itself. That's judgment-layer work in the most literal sense: someone is accountable, and it isn't a system.
Where it actually gets hard
The honest complication, using this case: the same person can sit in different layers depending on the task. An HRBP doing orchestration on routine policy questions might sit squarely in judgment when a business unit is going through a reorg and the "policy" genuinely doesn't cover the situation. That's not a flaw in the model — it's why task-mapping, done at the level of "what does this person actually do this week," produces a better redesign than mapping at the level of "what is this person's job title." Kordana's version of the exercise the Human Agency Scale researchers ran wouldn't be 844 tasks across 104 occupations — it'd be a few hundred tasks across six functions and one HRBP role, done with the people doing the work, the same principle at a fraction of the scale.
The other complication is sequencing. HR Operations and Talent Acquisition admin are the easiest wins and the most tempting to automate first, precisely because the green-light zone is largest there — but they're also where the people currently doing that work have the least obvious next role unless the orchestration layer is built out in parallel, with real training, ahead of the automation, not after it. Do it in the wrong order and you've cut execution capacity before you've built anywhere for those people to go.
Where would this exercise land differently in your own organization — and which function would turn out to have a bigger judgment core than the org chart currently assumes?
If you're working through a transformation like this, I coach HR and business leaders through exactly this kind of transition — or start with my book on self-coaching.
Want to explore this further?
Start with a free 30-minute discovery session — no commitment required.
Book a Discovery Call