THE REPLACEMENT REPORT

Weekly Intel on the
AI Workforce Shift.

No hype. No augmentation theater. Just straight analysis of what's happening, what it means, and what to do about it.

FEATURED

The Dorsey Signal: What Block's 4,000 Layoffs Mean for Your Company

Jack Dorsey just showed the entire business world what AI-first restructuring looks like in practice. Not in theory. Not in a McKinsey slide deck. In a tweet, followed by a stock price that went up 24% in 24 hours. Here's what it means for companies that can't afford to wait.

On February 26, 2026, Jack Dorsey fired 4,000 people. That's 40% of Block's total workforce. He announced it in a tweet. The stock market's response: a 24% surge. Goldman Sachs raised its price target. Wells Fargo called it "chock full of positive surprise."

Let that sink in. A CEO announced he was eliminating 40% of his employees. And the market cheered.

This wasn't a scandal. It wasn't a crisis. It was a strategy. And the market recognized it as the right one.

What Dorsey said that matters

The key quote isn't "we're firing people." It's this: "Within the next year, I believe the majority of companies will reach the same conclusion and make similar structural changes."

He's not predicting layoffs. He's predicting a structural shift in how companies are built. Smaller, flatter teams. AI handling the work that used to require headcount. Humans doing the things that actually require being human.

The math your competitors are already doing

Here's the calculation every mid-market CEO is running right now: If an AI agent can do the work of 3 data entry clerks, and that agent costs $3,000/month while the clerks cost $180,000/year collectively — why would you keep paying the clerks?

The answer used to be "because the AI isn't good enough." That answer stopped being true sometime around December 2025. Dorsey said it himself: "Something happened in December of last year where the models just got an order of magnitude more capable."

What to do right now

Start with an honest audit of your workforce. For every role in your organization, ask: is this person doing work that requires genuine human judgment, creativity, or relationship-building? Or are they executing a repeatable, documentable process that an AI could follow?

You'll find more of the second category than you expect. That's where Human Replacer starts.

Get Your Free Replacement Report →
ANALYSIS

24/7, No PTO, No Drama: The True Cost of an AI Agent vs. a Human Employee

Let's do the math that HR doesn't want you to do. When you fully load the cost of a human employee — salary, benefits, training, turnover, sick leave, performance management — against the cost of an AI agent, the comparison is not close.

The average fully-loaded cost of a mid-level office employee in the United States is $85,000-$120,000 per year when you include: base salary, employer payroll taxes (7.65%), health insurance ($7,911/year average employer contribution), retirement matching (3-5% of salary), paid time off (15-20 days), sick leave, workers' compensation, unemployment insurance, management overhead, training, and turnover costs (1.5-2x annual salary when someone quits).

For a $70,000 salary, the true annual cost is often $110,000-$130,000.

What an AI agent actually costs

An AI agent handling customer service triage, data processing, or document review typically runs $2,000-$5,000/month for compute, API calls, and monitoring — $24,000-$60,000 per year. That agent works 168 hours a week, not 40. It doesn't take vacation. It doesn't get sick. It doesn't quit in March and force you to hire and train a replacement.

On a per-hour-of-output basis, the AI agent is 80-90% cheaper than the human doing equivalent work.

The catch

AI agents aren't right for every job. They're exceptional at high-volume, rule-based work: data entry, email triage, invoice processing, compliance checking, report generation, appointment scheduling. They're poor at work requiring genuine creativity, complex negotiation, or deep human empathy.

The audit process exists to figure out which is which. Before you automate anything, you need an honest assessment of what's actually being done and whether a machine can do it better. That's the Replacement Report.

Learn About the Audit →
PRACTICAL

The Five Jobs AI Will Replace at Your Company This Year

Not in 10 years. This year. The capabilities are already here. The only question is whether you deploy them first or whether your competitors do. Here are the five highest-value automation targets in most mid-market businesses.

The model capabilities Dorsey referenced — the order-of-magnitude jump he said happened in December — are real. Multimodal AI can now read documents, understand context, make decisions, and take actions in software systems. That changes the calculus on what's automatable.

1. Customer Service Tier 1

80% of customer service tickets are the same 20 questions. An AI agent can handle tier-1 support, routing complex issues to humans while resolving the routine ones. Most companies can eliminate 60-70% of their customer service headcount or redeploy them to higher-value relationship work.

2. Data Entry and Reconciliation

Invoice processing. Expense reporting. CRM data entry from emails and calls. Any role that involves reading data from one place and entering it in another is a candidate for elimination. These agents pay for themselves in 3-6 months.

3. Report Generation

If someone on your team spends 10+ hours per week pulling data from multiple systems and assembling reports, that's an agent job. Automated data aggregation and report generation is one of the fastest-ROI automations we deploy.

4. Lead Qualification

The top of your sales funnel — initial outreach, qualification questions, scheduling discovery calls — is highly automatable. AI agents can qualify leads 24/7, respond to inbound inquiries instantly, and only escalate to human salespeople when there's genuine purchase intent.

5. Compliance Checking

Contract review for standard terms. Invoice validation against purchase orders. Expense report policy compliance. Document classification. These are high-volume, rule-based processes that create enormous liability if done wrong by a human, and that AI does with greater consistency.

Start With a Free Audit →