The AI Barbell: How Small Operators Win the Middle

What is the AI barbell effect and why does it matter for small businesses?

The AI barbell effect means AI concentrates power at two ends of every market: giant platforms that build or own the infrastructure, and small focused operators who rent that infrastructure to punch above their weight. The businesses getting squeezed are the ones in the middle. If you run an SMB, this is the most important strategic frame you can apply to AI right now.

Kasa CEO Roman Pedan laid this out clearly in Fortune: the next technology wave produces a barbell, with scaled platforms at one end and focused operators renting their capabilities at the other. His context is hospitality, but the mechanic is identical in professional services, logistics, retail, and healthcare. The question is not whether this happens in your industry. It already is.

Who gets squeezed and why?

The squeeze hits companies that are neither fish nor fowl. Think of a regional hotel management company with 15 properties. They are too large to be scrappy and personal, but too small to build proprietary revenue management software, AI-driven dynamic pricing, or a loyalty platform. They compete against both boutique independents and major chains, and they win on neither dimension.

This is not a hospitality-specific problem. A mid-size staffing firm with 40 recruiters faces the same trap. LinkedIn and the large staffing conglomerates have the data and the AI budget. A solo recruiter with a sharp niche and access to the right AI tools can undercut on cost and outperform on fit. The 40-person firm is often worse positioned than either.

The companies that will get hurt are the ones too comfortable to become focused operators and too small to become platforms.

The pattern repeats because the economics of AI favor extremes. Building proprietary AI is expensive and takes years. But renting AI capability through APIs, SaaS tools, and automation platforms costs a fraction of what it used to. According to McKinsey's 2024 State of AI report, 65% of organizations are now regularly using generative AI, up from 33% just a year prior. The tools are commoditizing fast. That is good news for focused operators who move quickly.

What does a focused operator actually do differently?

A focused operator does three things that mid-market generalists do not.

First, they define the niche sharply. Not "boutique hotels" but "extended-stay properties in secondary markets for traveling healthcare workers." Not "IT staffing" but "DevSecOps contractors for Series B SaaS companies." The niche is what lets AI actually work, because AI needs clean, specific context to produce useful output.

Second, they rent platform capabilities instead of building them. Kasa does not build its own revenue management engine from scratch. It plugs into existing infrastructure and runs lean operations on top. A small operator today can access the same quality of AI-driven pricing, scheduling, customer communication, and analytics that a Fortune 500 uses, through tools that charge per seat or per transaction. The infrastructure advantage of large companies is eroding.

Third, they use AI to do the work that used to require headcount. This is where the real leverage lives. A focused operator running on AI-augmented workflows can handle a volume of work that previously required a team two or three times larger. That is not a hypothetical. Firms using AI for proposal generation, client communication, scheduling, and reporting consistently report saving 8–12 hours per employee per week, based on data from Microsoft's 2024 Work Trend Index.

How does this play out in practice outside hospitality?

Let's make this concrete with a few direct parallels.

| Industry | Platform winner | Focused operator play | Mid-market loser | |---|---|---|---| | Hospitality | Marriott, Airbnb | Niche operator using AI ops tools | Regional management company | | Recruiting | LinkedIn, Indeed | Specialist recruiter with AI sourcing | 40-person generalist agency | | Accounting | Intuit, large CPA firms | Niche CFO advisor using AI for analysis | 10-partner general practice | | Legal | LegalZoom, BigLaw | Specialist attorney with AI-drafted docs | Mid-size generalist firm | | Marketing | HubSpot, WPP | Solo strategist with AI content engine | 20-person full-service agency |

In each row, the focused operator wins by combining human expertise in a specific domain with rented AI capability. They do not need to out-invest the platforms. They need to out-focus them.

What should a small business owner do right now?

The mistake most SMB operators make is treating AI as a productivity tool rather than a strategic positioning tool. They use ChatGPT to write emails faster. That is fine but it does not change their competitive position.

The operators who will look back at this period as their inflection point are the ones who ask a harder question: what would my business look like if I ran it with half the headcount and double the output? Then they redesign around the answer.

This is not about cutting people. It is about not hiring the next person until you have automated what that person would do. Every workflow you automate before hiring is margin you keep permanently.

The barbell effect is not coming. It is here. The window to position as a focused operator rather than a squeezed middle is shorter than most operators think. AI capability is not a moat if everyone has access to it. Your niche, your client relationships, and your speed of implementation are the moat.

What we'd actually do

• Audit your "middle" exposure this week. List the three services or products where you compete against both larger and smaller players. Those are your squeeze points. Pick one to either sharpen into a defensible niche or exit.
• Map one core workflow to AI-augmented delivery. Pick the workflow that takes the most hours per week and costs the most in labor. Document every step, then identify which steps can be handled by AI tools you can access today. Implement one change before the end of the month.
• Join a peer group that is actually running this. The operators moving fastest are not figuring this out alone. The AI For Business Skool community is where SMB owners share what is working, what is not, and how to implement without burning cycles on tools that do not move the needle.