DX Was Done by People. AX Will Be Done by Agents.

---

Prologue: The Empty Conference Room

Everyone is talking about "adopting AI." But I think they are missing the real story.

The headline is not that the tools are changing. It is that the doer is changing.

On May 4, 2026, Anthropic announced a $1.5 billion joint venture with Blackstone, Hellman & Friedman, and Goldman Sachs to build a new enterprise AI services firm¹. On the surface, it looks like "another AI consulting company." But buried in that announcement is the obituary of an entire era.

I have spent more than a decade inside a school system, watching digital transformation happen at the ground level. Paper attendance sheets became databases. Whiteboards became smartboards. Conference rooms became video calls. So I know how DX really works — and I know this announcement is not just another funding round.

DX was done by people. AX will be done by agents.

---

I – The DX We Knew Was "System Adoption"

Corporate digital transformation in the 2000s was structurally simple.

A consulting firm walks in. Partners, PMs, industry experts, and data analysts run workshops. They interview frontline staff and map out processes. They bolt on ERP, CRM, and a data warehouse. Months later, slick slides appear.

"DX, at its core, was about adopting systems. Moving paper to SaaS."

Why did this model exist? The company knew its inside but not the technology. The system integrator knew the technology but not the floor. Human PMs filled that gap by working through the night.

That is why BCG was already earning 20% of its consulting revenue from AI in 2024. They project that share will double to 40% by 2026².

Here is the contradiction. AI consulting revenue is doubling at the same time the consulting industry says it is being disrupted. Why?

The answer is simple. The unit of work has changed.

---

II – AX Means the Unit of Work Itself Splits Apart

AX is AI Transformation. It is not "deploying AI tools."

The company's unit of work stops being a software feature. It becomes an agent. Instead of one person using one tool, multiple agents share a single workflow.

Take a hospital. Old DX meant improving the EMR, suggesting billing codes, partially automating insurance pre-authorization. All of it built better systems for humans to use.

AX is structurally different.

graph TD
    subgraph DXera ["🔵 DX Era (2010s)"]
        A1[Consultant interviews] --> B1[Process mapping]
        B1 --> C1[System adoption]
        C1 --> D1[Humans use the system]
    end

    subgraph AXera ["🟢 AX Era (2026~)"]
        A2[Documentation agent] --> B2[Coding agent]
        B2 --> C2[Compliance agent]
        C2 --> D2[Operations & audit agents]
        D2 --> E2[Human: orchestrator]
    end

AtlantiCare in Atlantic City is already living this shift. Their clinical AI assistant reached 80% adoption among 50 providers in pilot, and users saw a 42% reduction in documentation time — saving roughly 66 minutes per day³.

Here is the core point. This is not one doctor with a better tool. There is a documentation agent drafting visit notes, a coding agent reviewing billing, a compliance agent checking risk language, an approval agent preparing payer submissions. They hand off to each other.

"DX taught humans to use systems. AX teaches agents to hand off work to other agents."

Gartner projects that 40% of enterprise applications will embed task-specific AI agents by the end of 2026 — up from less than 5% in 2025. That is roughly an 8x jump in a single year⁴.

---

III – The Real Message of $1.5B: "Engineers Inside the Client"

The most important sentence in Anthropic's announcement is not the deal size.

"Engineers will be embedded directly inside companies, redesigning workflows."¹

What does that actually mean? Old consultants built decks in conference rooms. The engineers inside this new firm will sit inside the client and bake Claude into the actual flow of work.

Dimension	Traditional DX consulting	New AX joint venture
Output	❌ Slides, roadmap	✅ Working agents
Staffing	❌ Workshop in a meeting room	✅ Embedded inside client
Unit of change	❌ A process	✅ A task + an agent
Revenue model	❌ Billable hours	✅ Value or subscription
Doer	❌ Human experts	✅ Humans + AI teammates

McKinsey projects that AI agents and robots could generate $2.9 trillion in annual U.S. economic value by 2030⁵. But the same body of research warns: 79% of enterprises face real challenges adopting AI, and more than 40% of agentic projects are likely to fail by 2027⁶.

"The real gap in the agent era is not technology. It is governance."

In 2024, only 11% of enterprises had a named "AI agent owner" or "agentic ops lead." By 2026, that number jumped to 56%⁶. Companies are inventing new job titles in real time.

---

IV – So Where Does the Human Go?

Here is the paradox.

DX turned humans into system operators. What does AX turn them into?

66% of enterprises are reducing entry-level hiring as they deploy AI⁷. At the same time, 77% of employers say they will reskill their existing workers around AI. Roles disappear and new roles appear in the same breath.

I saw the same pattern in schools. Manual attendance entry vanished, but the ability to read NEIS data and translate it into a parent conversation became newly valuable. So what should humans actually do in the AX era?

Practical strategy

Strategy	What it looks like
🎯 Agent orchestration	Designing and supervising flows across multiple agents — workflow design over coding
📐 Context curation	Encoding domain knowledge, policies, and edge cases so agents perform well
🛡️ Governance & audit	Owning the "why" behind agent decisions, especially in healthcare and finance
🤝 Trust with humans	Translating agent output into a story a customer or colleague can act on
🔍 Meta-judgment	Deciding which work should be delegated to an agent in the first place

In other words, humans no longer run the system. Humans conduct the systems.

---

💭 Questions to Sit With After This Piece

1. What is the first unit of work in your job that you would hand to an agent — and which slice of it do you most want to keep human?

2. If a new role like "agentic ops lead" appeared in your organization tomorrow, who in your team is best positioned to own it, and why?

3. With 79% of companies stuck on AI adoption, what separates the 21% that succeed — better technology, better people, or better governance?

Share your thoughts in the comments.

---

Conclusion: DX Is Over. AX Is Already Here.

What Anthropic just funded with $1.5 billion is not "another consulting firm." It is a bet that the work consultants used to do will be done by AI agents inside the client.

In the DX era we learned how to adopt systems. In the AX era we have to learn how to delegate work across agents.

Asking whether AI can write your blog post is a small question. The bigger one is waiting underneath.

"In an era where agents transform the company, what is the human role worth keeping inside it?"

The person who answers that question first will define the next decade of work.

---

Sources

Further reading: Anthropic takes shot at consulting industry in joint venture with Wall Street giants | Fortune · Anthropic teams with Goldman, Blackstone on $1.5 billion AI venture | CNBC

---

If this resonated, share it with one friend.

Footnotes

1. Building a new enterprise AI services company with Blackstone, Hellman & Friedman, and Goldman Sachs | Anthropic ↩ ↩²

2. 2026 Consulting Trends: Turning Uncertainty and AI Disruption into Competitive Advantage | Deltek ↩

3. How Agentic AI Is Transforming Enterprise Operations in 2026 | Lyzr AI ↩

4. Gartner Predicts 40% of Enterprise Apps Will Feature Task-Specific AI Agents by 2026 | Gartner ↩

5. AI Agent Adoption 2026: What the Data Shows | Joget (citing McKinsey) ↩

6. Agentic AI Stats 2026: Adoption Rates, ROI, & Market Trends | OneReach.ai ↩ ↩²

7. Enterprise AI adoption in 2026: Why 79% face challenges despite high investment | Writer ↩