The AI-Native Business: Why HR Is About to Become the Architect of the Enterprise

For the past few years, most conversations about AI in the workplace have followed the same script: AI as a tool that helps people do their existing jobs faster. Better forecasting. Faster drafting. Smarter search. Useful, but fundamentally additive — AI bolted onto an organization that otherwise runs exactly as it always has.

I recently had an opportunity to read an advanced copy of Jim Scully’s upcoming book The AI-Native Business (due out 2027) in which he argues that this framing badly undersells what’s actually happening. AI isn’t just automating tasks inside jobs. It’s dismantling the assumption that jobs are the right unit of organization in the first place. And because that assumption is baked into nearly every HR and HCM system in existence, Scully’s thesis lands squarely on HR’s desk: the function built to manage jobs, positions, and people is now being asked to design something else entirely — a system built around outcomes, capabilities, and signals.

This claim is bigger than it sounds. Let me walk you through it.

The Real Constraint Isn’t People or Technology — It’s the Operating Model

Scully’s starting point is that HR’s persistent struggles — the endless process redesigns, the module upgrades, the reorganizations that never quite stick — aren’t failures of execution. They’re symptoms of a deeper problem: the operating model itself was designed for coordination and control, not adaptation.

That model made sense in an industrial context. When intelligence — the ability to sense a problem, interpret it, and decide what to do — was locked inside human roles, you needed a hierarchy to route decisions to the right person, a job structure to define who was accountable for what, and a command chain to keep it all coordinated. Jobs, positions, and reporting lines weren’t arbitrary; they were the technology of organization for an era when intelligence lived only in people.

That hierarchy carried a second assumption, borrowed straight from the factory floor: two people in the same department, with the same job title, in the same position, are treated as interchangeable. Any manager could tell you that’s not really true. It’s why so much of the real work of running a team happens off to the side of the org chart, not through it. But the assumption stuck around because treating every person as a distinct case was more than the old systems could handle.

AI breaks that constraint. Machines can now sense, interpret, recommend, and even act, which means intelligence is no longer tied to a role or a seat on an org chart. That’s a genuine shift in what an organization is capable of, not just a productivity boost. And it exposes the old model’s weaknesses: rigidity, latency, poor visibility, and a chronic lack of agency at the point where decisions actually need to be made. A structure built to route intelligence through fixed roles doesn’t know what to do when intelligence shows up everywhere at once.

From Jobs to Outcomes

If jobs were the organizing unit of the industrial model, Scully argues the organizing unit of the AI-native model is the outcome. This is a genuine inversion, not a rebrand. An activity-based organization asks, “Is this person doing their job?” An outcome-based organization asks, “Is the outcome actually happening?” Those are different questions, and they demand different infrastructure to answer.

Underneath that distinction is a simpler, more practical one: are you supply-driven or demand-driven? A job-based organization supplies whatever people and capacity it happens to have, whether or not that matches what’s actually needed right now. An outcome-based organization flips that order: figure out what the moment requires, then pull together whatever mix of people, agents, and automation can deliver it. Work gets done because it’s needed, not just because someone was staffed to do it.

Outcome-based accountability doesn’t eliminate roles, but it stops treating the role as the thing being managed. Work gets organized around what needs to be true in the world — a customer retained, a product shipped, a risk mitigated — and the humans, AI agents, and automated systems that contribute to that outcome get assembled and reassembled as needed. The job description becomes a much less useful artifact than the outcome definition.

The Workforce Becomes a Hybrid System

This is where the “hybrid workforce” idea stops being a buzzword and becomes a design requirement. In an AI-native organization, the workforce isn’t composed of humans who happen to use AI tools. It’s humans, AI agents, and automation working together as a single system to produce outcomes, with capability drawn from whichever combination gets the job done.

That has real implications for anyone building or buying workforce systems. A platform that only has fields for “employee” and “job” has no way to represent an AI agent that owns part of an outcome, or a piece of automation that’s doing meaningful work alongside a person. If the workforce is genuinely hybrid, the system of record has to be hybrid too — which is a fundamentally different design brief than adding an “AI” tag to an existing employee record.

Signals Replace Activity Reports

Perhaps the most technically interesting idea in the book is what Scully calls signal architecture. If the organization is managing outcomes rather than activities, it needs a different kind of information flow to know whether those outcomes are actually being achieved. Traditional HR systems are built to track activity — hours logged, tasks completed, reviews submitted. None of that tells you whether the outcome is true.

This does not mean that HR was doing it “wrong” before AI entered our workplaces. Tracking activity and managing supply made sense when time, knowledge, and human bandwidth were the real limits on what an organization could do. As AI loosens those limits, the job changes too. When capacity was the bottleneck, managing capacity was the right call. When capability becomes the bottleneck instead, that’s what needs managing, and that’s exactly what signal architecture is designed to help with.

Signals are meant to answer that question directly, and in something closer to real time than an annual review cycle allows. That means real-time signals about what’s happening now, predictive signals about what’s likely to happen next, workforce risk signals that flag where an outcome is at risk before it fails, and capability signals that show whether the organization actually has what it needs to deliver. This is a genuinely new layer for most HR technology stacks, not some report you pull. It is a live nervous system for the workforce.

Authority Moves Sideways

One of the quieter but more disruptive ideas in the book is what Scully calls East-West authority: the notion that as work organizes around outcomes rather than jobs, coordination increasingly happens across teams rather than up and down a hierarchy. Mission-based teams form around a specific outcome, draw authority and resources from wherever they’re needed, and dissolve once the outcome is delivered — rather than living permanently inside a fixed department.

For a platform to support this, it needs to make temporary, cross-functional structures a first-class concept rather than a workaround bolted onto a permanent org chart. That’s a meaningfully different data model than most HR systems were built on.

Why HR, Specifically

Given all of this, it would be reasonable to assume IT or corporate strategy should own the redesign. Scully makes a different argument, and the logic holds up: HR is the function that already owns workforce design, capability deployment, performance logic, and governance. Those are exactly the levers this transition runs through. IT can build the infrastructure and strategy can set the direction, but the actual architecture of how work gets organized, how authority flows, and how performance is measured is HR’s domain — whether or not HR has historically thought of itself that way.

That reframes HR transformation as something bigger than an HR project. Scully calls it enterprise architecture. Operating models influence how businesses produce value at a given moment in time, and the operating model is now HR’s to define. It also means the HR leaders who treat this as “someone else’s technology problem” are handing away the most consequential design decision their organization will make this decade.

Two Models, Running in Parallel, for Years

It’s worth being honest about the timeline here, because Scully is. This isn’t a switch that flips. Organizations will run their legacy model and an AI-native model simultaneously, for years, as a structural reality rather than a temporary transition phase. That has quiet but important implications: workforce systems need to support both jobs and outcomes at once, both hierarchical and mission-based coordination at once. Both activity tracking and signal architecture need to take place all at once, because most real organizations will live in that overlap for a long time before any full cutover happens.

What This Means for the Systems We Build

The book’s most compelling claim isn’t really about AI at all. It’s this: the next generation of enterprise software isn’t a category of applications. Operating models are now implemented as platforms. The system stops being a tool that supports the operating model from outside and becomes the operating model, encoded directly in how the software works.

For anyone building HR or HCM technology, that reframes the whole design brief. It’s not enough to add AI features to a system still organized around jobs and positions. The system needs to manage capabilities, outcomes, work, agents, and capacity as first-class objects — not just employees, jobs, and departments with an AI layer on top. It needs to function as a Total Workforce system of record that treats humans, contractors, automation, and AI agents as parts of one coordinated system, not as separate categories bolted together. And it needs to support outcome ownership and tracking, live signal architecture, and mission-based team structures as native capabilities, not add-ons.

None of this happens overnight, and Scully’s point about parallel operating models is a useful check against overclaiming otherwise. But the direction is clear enough to act on now: the organizations, and the platforms, that start building toward outcomes, hybrid workforces, and signal-driven visibility today will be the ones equipped to operate when the industrial model finally runs out of runway. The question worth asking isn’t whether this shift is coming. It’s whether your workforce system is being built for the organization you have, or the one you’re about to become.