And Why Mid-Market CEOs Should Be Paying Very Close Attention
McKinsey & Company QuantumBlack, AI by McKinsey OpenAI Anthropic
Something important happened over the past thirty days. Two developments — separated by a few weeks but connected by a single logic — just redrew the map of who gets to benefit from AI and who gets left figuring it out on their own.
If you run a mid-market company or lead a family-owned enterprise, these moves were not made for you. And that’s precisely the problem.
What Just Happened
Move #1: The AI labs locked arms with the world’s most powerful consulting firms.
In late February, OpenAI launched what it calls “Frontier Alliances” — multi-year partnerships with McKinsey, BCG, Accenture, and Capgemini. The consulting firms will embed certified teams alongside OpenAI’s own forward-deployed engineers to help enterprise clients redesign workflows, integrate AI agents, and deploy at scale. Days later, Anthropic answered with its own $100 million Claude Partner Network, bringing Accenture, Deloitte, Cognizant, and Infosys into a formal ecosystem to do the same thing with Claude.
These aren’t marketing announcements. Accenture is training 30,000 professionals on Claude. McKinsey is embedding OpenAI’s Frontier platform into its transformation practice through QuantumBlack. Deloitte is standing up a Claude Center of Excellence with 15,000 certified practitioners. The scale of investment — in people, in training, in co-developed IP — signals that the largest consulting firms on earth have decided that AI deployment is their next decade of growth.
Move #2: The AI labs went directly to private equity.
This one is newer and arguably more consequential. Reuters reported this month that OpenAI is in advanced talks with TPG, Bain Capital, Advent International, and Brookfield Asset Management to form a joint venture valued at roughly $10 billion. The PE firms would commit approximately $4 billion, receive equity stakes, and — critically — influence over how OpenAI’s technology gets deployed across their portfolio companies. OpenAI is reportedly offering preferred equity with a guaranteed minimum return of 17.5%.
Anthropic is running a parallel play, courting Blackstone, Hellman & Friedman, and Permira for a similar joint venture.
The logic is straightforward: private equity firms control thousands of mid-market and large companies. They set technology budgets. They influence strategy. If you want rapid enterprise adoption, you don’t just sell software — you partner with the people who own the companies.
The Divergence
Read those two moves together and a clear picture emerges. There are now three categories of companies in the real economy, impacted by AI — and they are pulling apart fast.
Category 1: The giants. Fortune 500 companies now have McKinsey, BCG, Accenture, and others with direct access to OpenAI’s forward-deployed engineers and Anthropic’s Applied AI architects helping them redesign their operating models. These are not pilot programs. These are multi-year transformation engagements backed by hundreds of billions in combined consulting capacity. The largest enterprises in the world will have AI wired into their operations in a way that is structural, not experimental.
Category 2: PE-backed companies. If your company is owned by TPG, Bain Capital, Brookfield, Blackstone, or the firms that follow them into these joint ventures, you’re about to get AI deployment handed to you from above. Your PE sponsor will have early access to the newest models, forward-deployed engineering support, and a financial incentive to accelerate adoption across their portfolio. For PE-backed companies, AI adoption isn’t a question of strategy. It’s increasingly a condition of ownership.
Category 3: Everyone else. Mid-market companies. Family-owned businesses. Regional enterprises, small and mid-cap public companies, and others. Companies with real competitive exposure, meaningful revenue, and no consulting alliance or PE sponsor to hand them a deployment playbook. This is the group I’ve been thinking about most. Some companies in this group are moving aggressively on AI — often faster than their larger or PE-backed peers precisely because they have simpler tech stacks, shorter decision cycles, and leadership teams that can commit without committee. The risk isn’t inherent to the category. It’s just that without a forcing function, too many companies in this group will move too slowly.
The Real Challenge Isn’t Awareness — It’s Architecture
Most leaders I talk to know AI matters. That’s not the problem. The problem is that knowing AI matters and knowing what to do about it are two entirely different things, and the gap between those two states is where competitive advantage is being won and lost right now.
Here’s what I see in that gap:
The copilot ceiling. Many organizations have invested in AI training. They’ve rolled out ChatGPT Enterprise or Microsoft Copilot. Some employees are using it daily. Leadership checks the box: “We’re doing AI.” But copilot adoption — teaching people to use AI as a personal productivity tool — is the floor, not the ceiling. The companies in Categories 1 and 2 aren’t stopping at copilots. They’re deploying AI agents that execute workflows, make decisions within guardrails, and fundamentally change how work gets done. The distance between “my team uses ChatGPT” and “AI is embedded in our operating model” is enormous.
Asking the wrong question. Too many companies are still asking, “What is our AI strategy?” That question sounds responsible. It is actually dangerous because it frames AI as a standalone initiative — something that lives in a strategy document next to the digital transformation roadmap and the innovation agenda, often owned by the CIO or a Chief AI Officer.
That’s a fundamentally different inquiry, and it forces you to think in three dimensions simultaneously.
First, the revenue and market dimension. How does AI change what your customers expect from you? How does it change how they find you, evaluate you, and decide to buy? If your competitors can personalize at scale, respond in real time, and reduce friction in ways you can’t match — that’s not a technology problem, it’s a market position problem.
Second, the cost and operating model dimension. AI doesn’t just automate tasks. It restructures the economics of how work gets done. When intelligence becomes an abundant input rather than a scarce one — when the cost of analysis, synthesis, content creation, code generation, and decision support drops toward zero — every assumption you’ve made about headcount, process design, and margin structure needs to be reexamined. Not because you should fire people, but because your competitors will use that abundance to deliver more value at lower cost. And if you’re still operating on the old cost curve, you won’t see the threat until it’s too late.
Third, the competitive reaction dimension. This is the one many leadership teams miss. It’s not enough to ask what AI can do for your company. You have to ask what AI enables your competitors to do — including competitors who don’t exist yet. When a critical input to business — whether it’s energy, compute, distribution, or intelligence — becomes dramatically cheaper, the companies that figure out how to exploit that abundance first don’t just get a marginal advantage. They redefine what’s possible in the category. Think about what happened when the cost of distribution went to zero because of the internet. The companies that treated it as a “distribution strategy” initiative got disrupted by companies that rebuilt their entire business model around free distribution. The same pattern is emerging now with intelligence.
What Mid-Market Companies Are Actually Up Against
Let me be specific about the challenges, because platitudes about “AI readiness” aren’t useful.
Challenge 1: Where to deploy. The conversation in many mid-market organizations is still abstract. “We should use AI in finance.” “We should look at AI for legal” “Maybe there’s something in marketing.” The companies in Categories 1 and 2 are past this stage. They have consulting teams running diagnostic assessments, identifying the highest-ROI use cases, building business cases, and sequencing deployment. Mid-market companies are trying to do that work internally, often with a CIO or IT leader who has deep technical competence but hasn’t been trained to evaluate AI’s strategic implications at the enterprise level. This isn’t a criticism of those leaders — it’s a structural problem. Nobody trained them for this because this didn’t exist two years ago.
Challenge 2: Competitive threat. When your competitor is a PE-backed company that just got access to OpenAI’s newest models, forward-deployed engineers, and a deployment team through a dedicated joint venture, you won’t see the impact in their next earnings report. You’ll see it in their pricing, their speed to market, their customer experience, their ability to operate leaner. By the time the competitive impact is visible, the gap is already structural. The hardest threats to respond to are the ones that manifest as gradually improving execution by your competitors rather than a single dramatic move.
Challenge 3: Translation layer. The consulting alliances and PE joint ventures solve a specific problem: they provide a translation layer between what AI can do and what a specific company needs to do. McKinsey’s value in the Frontier Alliance isn’t that they understand AI models better than Google. It’s that they understand how to take a general-purpose technology and map it to specific industry workflows, organizational structures, and competitive dynamics. That translation layer is exactly what mid-market companies don’t have — and it’s exactly what they need most.
Challenge 4: Internal vocabulary. Across the mid-market landscape, AI frequently shows up as “management’s update on AI initiatives” — a framing that treats it as a status report rather than a strategic conversation. The more productive framing is to ask whether the company’s strategy still makes sense in a world where intelligence is becoming nearly free. Many companies are actively building the vocabulary and frameworks to engage with that question. The pace of that learning matters.
What Needs to Change
I don’t have a silver bullet, and I’m skeptical of anyone who claims to. But I’ll offer a few principles I believe in.
Stop treating AI as a technology initiative and start treating it as a strategy question. Move it off the CIO’s plate and onto the CEO’s agenda. Not because the CIO isn’t capable, but because AI’s impact crosses every function — marketing, operations, finance, talent, product — and only the CEO can hold that cross-functional view.
Get honest about where you actually are. Not where your innovation team says you are. Not where the vendor who sold you a platform says you are. Where you actually are in terms of data readiness, leadership literacy, organizational willingness to change, and strategic clarity about what AI means for your competitive position. That honest assessment is uncomfortable, but it’s the starting point.
Don’t wait for the perfect plan. The companies in Categories 1 and 2 are not waiting. They’re moving now, learning as they go, and building capabilities through action. A mid-market company that spends six months building a comprehensive AI strategy will emerge into a market that has moved another six months ahead. Start with the highest-stakes question: Where is your business most exposed to a competitor who uses AI better than you do? That’s where you start.
Make AI a standing strategic agenda item. AI is no longer a topic that can be delegated entirely to the technology function. It is a strategic force on the order of the internet, globalization, and deregulation. Companies whose leadership teams engage substantively with AI’s implications for strategy, risk, and competitive positioning will navigate what’s coming. Those that don’t will need to close the gap — through education, outside expertise, or new perspectives in the room.
Mid-Market Firms Have Some Advantages, but the Clock Is Ticking
Mid-market firms often have a less complex tech stack and less complex organizational structure. Transforming them is often easier than transforming a multinational firm that grew through a dozen acquisitions (each that came with its own tech stack) over the last couple of decades. And unlike PE-backed firms, the model, targets, and timelines are not being thrust upon you by an operating partner (often a former consultant with scant operating experience). You get to decide and build your own journey in a way that is practical for your company and culture.
What I am saying is that the largest, most powerful institutions in the global economy — the AI labs, the top consulting firms, and the largest private equity houses — have just formed a series of alliances that will dramatically accelerate AI adoption for their clients and portfolio companies. That acceleration creates a widening gap. And the companies that will thrive are the ones that recognize the gap, take it seriously, and act with urgency — even without the same resources.
That question deserves an answer from the top of the house. And it deserves one soon.
I'm working on a diagnostic framework that helps mid-market firms assess their right strategic posture for AI, given their industry, and their internal capabilities. If you'd like early access, reach out.