I Said HR Should Go First. The Comments Told Me What's Really in the Way.

A community summary to last week's piece. Seven themes from more than 120 comments, and what we should do about each.

A week ago, I argued that HR can't credibly redesign everyone else's work until it has redesigned its own. I expected some agreement and some pushbacks. The usual.

What I didn't expect was for the comment section to be better than the article.

More than 120 of you showed up and argued it out. Sitting CHROs and fractional ones, people analytics leaders, org-design practitioners, consultants, and founders building the tooling. You sharpened the case, broke it where it needed breaking, and handed me at least three ideas I wish I'd written myself.

Net net: this piece isn't mine. It's ours. Seven themes emerged, and rather than just summarize them, I took a run at the question that really matters: what do we do about it?

Key Themes and What to Do About Them

1. The bottleneck isn't tools. It's will, and exposure.

This was the loudest theme by a wide margin. Kyle Forrest opened with a distinction half the thread ended up borrowing: the "CHROs of yesterday" versus the "CHROs of tomorrow." Same role, two completely different reactions to the ground shifting underneath it. Cliff Jurkiewicz was blunter, arguing that leaders unwilling to take on work transformation should step aside for those who will.

Olivier Vidal named the mechanism underneath the reluctance, and it's the sharpest diagnosis in the thread. This work is almost front-office. Highly visible. Exposed to being wrong in a way HR rarely has been. That is the opposite of the risk profile the function has selected for over decades.

What to do about it: You can't buy your way out of a courage problem. No tool makes it feel safe to stand in front of the business and make a call that might be wrong. You get comfortable doing that the only way anyone does, by practicing where the stakes are low. The obvious play here is to practice on your own function first. Pick one workflow, redesign it, and let it be the place you're allowed to get things wrong. That's the real reason to run the first transformation on HR before you run it on the business: not to prove the model, but to build the nerve and the experience and the confidence to make these calls in public.

2. "If you don't own the work, what do you own?" landed harder than I expected it to.

The disintermediation question at the end of the original piece drew the strongest reactions of all. Kyle Forrest relayed a client now asking, function by function, whether each one even needs to exist. Paul Rubenstein and others circled the old provocation that HR nirvana is a world where the function disappears because the workforce data finally lives with the leaders making the decisions.

What to do about it: Don't defend the current remit. Move upstream, fast, to the definition of the work itself, before someone else claims that ground. Ownership of the upstream work – distillation of business strategy into work and roles strategy, reimagining and redefinition of work itself in light of automation, and developing scenario models and gap filling strategies for continuous workforce planning – is defensible and durable. Everything downstream of it, compliance and systems administration and disconnected talent tech, is execution. And execution is exactly what gets absorbed into the functions around you.

3. Orchestrate the work. Don't try to own it.

The most useful pushback was also the most pointed. Jake Mcdonald argued that HR should orchestrate this transformation without becoming its sole authority, and that the deeper need is a shared, continuously maintained representation of work rather than a new HR power grab. Chris Davis went further and called the whole idea an over-expansion of HR into command-and-control of the entire enterprise. Mark Levi doubted HR understands the actual work well enough to redesign it at all.

They're identifying a real failure mode, and they're right that "own" is the wrong verb.

What to do about it: Let’s switch out the verbs – from “own” to “orchestrate” – and treat orchestration as its own discipline. Think about how FP&A orchestrates financial planning without owning a single P&L line. The business still owns outcomes. Technology owns what's now possible. Finance owns the economics. HR connects those choices to roles, skills, capacity, and structure, through a small permanent core with a federated network around it. That isn't empire-building. It's growing connective tissue - the critical connective tissue that most organizations lack.

4. The scarce skill is work decomposition, and it's more learnable than it looks.

NINA HIROTA, MA, MBA named the capability everyone skips: taking a process as it actually runs and breaking it into the tasks, decisions, handoffs, and exceptions that produce the result. Most functions have never looked at themselves at that resolution. Paul Rubenstein's prescription was the most-liked comment in the thread, and refreshingly unglamorous: go do some job shadowing, get good at activity-based costing, dust off the Six Sigma book.

What to do about it: Just get started – cannonball into the deep end – because it's more accessible than it looks. A couple of SME interviews to get the shape, an LLM or some disciplined human thinking to fill the gaps, and you have a credible first pass. A number of vendors, us included, are now doing a lot of this with agentic support, which collapses the timeline further.  

The genuinely hard part comes after decomposition, when what's left over isn't a whole job anymore. That recombination problem, reassembling job fragments into new roles, is where the real design work is going to live for the next ten years.  The good news is that supporting tech is now emerging to enable this at scale – we’ll have something in market this fall.

5. Stop scaling the old lens faster. Mind the analytical tax.

Dr. Timothy J. Giardino, BS, MBA, PhD, SHRM-SCP added the idea I keep returning to: if we simply bolt AI onto HR's existing lens, we scale HR's existing blind spots faster. And a cluster of you pushed hard on the foundation problem, that most HR functions can't even agree with themselves on what a "role" or a "skill" is. The HRIS says one thing, the learning platform another, the talent marketplace a third, and external labor market data speaks a fourth language entirely.

When that's true, every transformation begins with reconciliation instead of analysis. That reconciliation cost has a name. It's the analytical tax: the hidden toll a fragmented stack charges you, where your best people spend their time reconciling data instead of reasoning over it. It compounds every time you add another tool with its own private definition of work.

What to do about it: Pay down the tax with a shared semantic layer. One definition of process, role, task, and skill that every system, every planning model, and every AI agent can reason over. It's the least glamorous investment on this list (somehow even less sexy than Paul Rubenstein's advocacy for job shadowing and Six Sigma), yet it’s easily the highest-leverage one, both in terms of impact but also in terms of problem scope and depth.  We have yet to find a non-client that has this dialed in which suggests this is a universal, root cause issue in just about every organization.

6. Redesign the work, AND redesign how you decide.

Crystal Wigfall, MBA added a layer that my structural model underplayed. As AI reshapes workflows, leaders are making higher-consequence judgment calls with less precedent and worse information. Workflow architecture alone isn't enough; you also need decision architecture, a disciplined way to decide where human judgment intentionally stays. Jordana Cole, MAPP, PCC reframed the whole thing: the CHRO living "additional duties as assigned" is experiencing exactly what her people experience every day. AI is just the catalyst. The story is human. Alok Vermaa reminded us that psychological safety is the substrate all of this actually runs on.

What to do about it: Build the human checkpoints into the redesign deliberately, and state plainly where judgment stays by design rather than by accident. Jake Macdonald's line belongs on the wall for this one: efficiency gained is not the same as human value preserved.

And make sure you hit JESS VON BANK's substack – esp “The Five Wounds of Change” – for a treatise on how to start thinking more deeply about human impacts from automation advances.

7. This is continuous now. The annual exercise is finished.

Rachael Kelly put it cleanly: workforce planning can no longer be an annual HR ritual. It needs to become a continuous operating discipline that connects directly to operating strategy and capital allocation. Natalia L. reminded us we're in the first inning of this. Beverly Carleton grounded it in real-time run-rate cost and scenario modeling.

What to do about it: Treat the model of work as a living system, not a document. A static map starts decaying the day it's published, which is why every job architecture that ends its life as a spreadsheet was doomed at launch. You need both the map and the mapmakers, a standing capability rather than a one-time project. Plan continuously, or plan to be out of date.

The part I can't shake

If the comment section made you wince, take it as a good sign. It means you're in the ring, duking it out and fighting for the future of HR and the future of the CHRO role. Nobody's job description a year ago included agent orchestration or the economics of human-machine work allocation. The role changed - foundationally - right under our feet. That isn't a personal failing. It's the condition we're all now working in.

But here's the observation I keep coming back to, and it makes me a bit unsettled. Look at who showed up in those 120 comments. Consultants, founders, fractional CHROs, thought leaders, a few "leading edge" CHROs. Largely the already-convinced. The sitting CHROs this piece is really about, the ones running billion-dollar functions on last decade's operating model, either read it and didn’t engage, didn’t read it, or weren’t moved enough to jump into the dialog.

That silence is the thesis proving itself. And that's the unsettling bit.

At TalentNeuron, we're furiously building toward the version of this where the map, the shared semantic layer, and the continuous plan all live in one place.  We’re already implemented and deployed across dozens of happy Fortune 2000 clients with an ever-growing pipeline of new prospects and a rapidly evolving capability set.

That said, all the tooling in the world isn’t going to matter much if HR organizations lack to courage to step up and own predictive models; refuse to move upstream to the true strategic work; refuse to reinvest in old-school skills like Six Sigma, job design, and time / work studies; fail to invest in or learn new skills like work decomposition, financial modeling, AI and consulting; and lack the imagination to reimagine the remit, the roles, the processes, and the organizational structures. Absent this investment and commitment, all the tooling in the world, no matter how advanced, is just noise.

So let me hand the question back to the room, since the room turned out to be collectively much sharper than the article itself:

"If you've already started running this transformation on your own HR function: what broke first, and what do you wish you'd stood up before you began?"