Jack Dorsey Fired 4,000 People and Called It AI. Here Is What Actually Happened, and Why Agency Owners Should Pay Attention.

In early 2026, Jack Dorsey cut nearly half of Block's workforce. The company went from roughly 10,000 employees to under 6,000. His reason, stated publicly: "AI can do their jobs now."

Block's stock surged 24% the same day. Wall Street loved it. The narrative was irresistible: visionary founder embraces AI, trims the fat, stock soars.

But that narrative is missing some important details.

What Actually Happened at Block

Block did not surgically replace 4,000 roles with AI systems. They quietly eliminated diversity roles, policy teams, and entire departments that had nothing to do with artificial intelligence. The "AI" framing gave the layoffs a futuristic sheen that made them palatable to investors and the press.

Even Sam Altman, the CEO of OpenAI, the company that arguably created the current AI wave, publicly stated that companies are using "AI" as a cover story for standard cost-cutting.

When the CEO of the world's leading AI company tells you that "AI layoffs" are often not about AI, that is worth paying attention to.

The Pattern Is Spreading

Block is not alone. A clear pattern has emerged across the technology and professional services industries:

Salesforce fired 4,000 people and replaced them with AI agents. The AI tools were not ready. Workflows broke. Remaining employees were stretched across two or three roles. Morale collapsed. This is a trillion-dollar tech company with some of the best AI engineers in the world. If they could not make wholesale AI replacement work, the odds for smaller organisations are significantly worse.

Klarna replaced 700 customer service agents with AI. The CEO celebrated on every stage he could find. Response times improved 82%. Then quality tanked. Customers started getting wrong answers to complex queries. Now Klarna is quietly hiring humans again at $41 per hour through a freelance model. The CEO's new line: "Cost was too predominant a factor."

WPP cut 7,000 jobs. Dentsu cut 3,400. IPG cut 3,200. Omnicom killed DDB, FCB, and MullenLowe and announced $1 billion in labour reductions. Over 17,000 agency jobs gone in 12 months.

In each case, AI was cited as the reason or the enabler. In each case, the reality was more complicated.

The Two AI Playbooks

What the data actually shows is that there are two fundamentally different approaches to AI in business. One works. The other creates chaos.

Playbook 1: AI as replacement. Fire people. Tell investors AI will do their jobs. Cut costs. Hope the stock price holds long enough to figure out the operational gaps later. This is the Dorsey playbook. It is also the Klarna playbook. And the early Salesforce playbook. The results are consistent: short-term stock bumps followed by operational problems that cost more to fix than the layoffs saved.

Playbook 2: AI as augmentation. Keep the team. Automate the manual, repetitive work that eats 60-70% of their time. Redirect those hours to revenue-generating activities. The team costs the same. The output doubles. Margins climb. This is the playbook that PwC's research found delivers 340% ROI within 18 months.

The difference is not subtle. It is the difference between a company that cuts costs and crosses its fingers, and a company that builds a structural advantage that compounds over time.

Why This Matters for Agency Owners

Right now, recruitment agency owners and marketing agency owners are hearing the same pitch from every direction: "Just replace your team with AI. It will be cheaper."

Vendors are selling AI tools with the promise that you can cut headcount. Consultants are advising restructures built around AI replacement. LinkedIn is full of posts celebrating companies that fired people and automated their work.

The pitch is seductive. And the maths seem to work, on paper.

But here is what the pitch leaves out:

The relationship cost. In recruitment, the placement fee is earned through relationships: with candidates and with clients. When a top candidate calls and gets a chatbot instead of a human who knows their career history, they call the next agency. That is an $8,500 placement fee gone. Bullhorn's data shows 54% of candidates abandon recruiters who move too slowly. Replacing recruiters with AI is the ultimate speed reduction for the interactions that matter most.

The quality cost. AI is excellent at pattern-matching and terrible at judgement calls. It can score a candidate against criteria. It cannot read the room in a client meeting. It can generate a brief. It cannot tell when a brief needs a human touch because the client is going through a restructure and needs to be handled carefully. Klarna learned this the hard way when AI-generated customer service responses started creating more problems than they solved.

The morale cost. When you fire half the team and tell the survivors to "use AI," what you get is not a leaner, more efficient organisation. What you get is a team of people doing three jobs, constantly anxious about being the next to go, and too exhausted to do any of those jobs well. Salesforce's experience confirmed this.

The Right Way to Use AI in an Agency

The agencies that are actually winning, not the ones making headlines for layoffs, but the ones quietly doubling their margins, are doing something very specific:

They are automating the work nobody was hired to do.

Nobody became a recruiter because they love formatting CVs. Nobody became a marketing strategist because they enjoy copying data between spreadsheets. Nobody became a client manager because they wanted to spend three hours a week on reporting.

That is the work AI should handle. The manual, repetitive, time-consuming administrative tasks that consume 60-70% of a recruiter's week (Randstad/LinkedIn data) and produce zero direct revenue.

When you automate that layer, and keep the humans doing the work that earns the fee, something interesting happens:

• Recruiters spend more time on relationships and placements
• Candidates get faster responses and better experiences
• Clients get higher-quality briefs delivered in hours instead of days
• Time-to-fill drops from 44+ days to under 10 (Bullhorn data for AI-embedded firms)
• Revenue per recruiter increases because the same person is doing more revenue-generating work
• Margins climb from 15-25% to 40-60% (StoryChief data for AI-native agencies)

That is not a cost-cutting story. That is a growth story. And growth stories do not require firing anyone.

The Test for Agency Owners

Next time someone pitches you on AI, whether it is a tool, a consultant, or a LinkedIn thought leader, ask one question:

"Does this replace my people, or does it make my people faster?"

If the answer is "replace," remember Dorsey. Remember Klarna. Remember Salesforce.

If the answer is "faster," look at the data: 340% ROI (PwC), 4x revenue growth likelihood (Bullhorn GRID 2026), 40-60% higher margins (StoryChief).

There is a massive difference between "AI replaced my team" and "AI made my team dangerous."

Dorsey chose the first option. His stock went up for a day. The agencies choosing the second option are building advantages that compound for years.