If you're running a $40 million revenue company and you've been told "you should look at what Deloitte is doing with AI," nobody told you the part where Deloitte will not sell to you.
It's not malice. It's structural economics. The Big Four (Deloitte, PwC, EY, KPMG) and the Accenture-class consultancies built their AI platforms around an enterprise sales motion that requires a six-figure annual engagement floor just to clear the unit-economics math. Below that floor, the partner-track delivery model, the global tax compliance overhead, and the SI-grade implementation team don't fit.
This piece is the audit. Public sources, analyst estimates, and what the firms themselves publish in their press releases. Pricing is not always transparent (one of the structural problems we're documenting), so where the firm refuses to publish, we cite the market intelligence and label it estimated.
The four firms, and what they're selling
Zora AI
Deloitte's enterprise agentic AI platform. Sold into IT through Deloitte Consulting's managed services contracts. Targets the Fortune 1000. Implementation timelines run six to eighteen months. Sold by an army of Deloitte consultants at the standard partner-track $300-500 hourly rate.
Floor pricing is not published. Analyst estimates put a mid-sized Zora engagement at $500,000 to $3 million for the first year, depending on integration depth and number of business processes wrapped in the agent layer. Multi-million-dollar engagements are normal.
agent OS
PwC's agentic AI platform. Sold by 50,000+ consultants globally. Targets the Fortune 1000. Same shape as Zora: implementation-led delivery, enterprise IT buyer, multi-quarter rollout.
No public pricing. Industry contacts in Q1 2026 reported initial engagement quotes starting in the high six figures, with the typical agent OS deployment landing in the seven-figure range over an 18-month deployment window.
EY.ai
EY's "AI ecosystem" platform integrating Microsoft, IBM, Dell, and Nvidia stack components. Sold into enterprise IT departments alongside EY's existing audit and consulting services. EY publicly committed $1.4 billion to the platform.
EY does not publish engagement prices. The platform is sold as part of larger consulting engagements, which means the AI line item is rarely separable from a multi-million-dollar consulting contract.
AI Refinery
Accenture's NVIDIA-partnered AI platform. Sold into the same enterprise IT buyer as the Big Four offerings, but Accenture's platform sits more in the systems-integration tradition: heavy compute, custom model fine-tuning, integrated with existing Accenture-managed cloud infrastructure.
No public pricing. Accenture engagements that include AI Refinery as a component are typically multi-million-dollar systems integration contracts spanning 12-24 months.
Why the $500K floor exists
The floor is not arbitrary. It comes from the unit economics of how Big Four firms staff and deliver work.
Every Big Four engagement is delivered by a team that includes a partner (charged at $500-1,200 per hour), a senior manager, several managers, and a stack of senior consultants and analysts. The blended hourly rate sits between $300 and $500 once you weight the team correctly.
The minimum staffing for an enterprise AI engagement is roughly 4-6 people working part-time across a 6-month rollout. That's 2,000-4,000 hours of billed time at $300-500. Do the multiplication and the floor lands at $600,000 to $2,000,000 just for the people. Software, licensing, and integration costs sit on top.
If the firm tried to deliver the same engagement at $50,000 (which is what a mid-market company can actually allocate), the math doesn't close. The partner doesn't get billed. The compliance overhead alone burns more than that.
This is why every Big Four AI offering is structurally locked to enterprise-only buyers. It's not strategy. It's arithmetic.
What the mid-market gets quoted instead
If you call Deloitte at $40 million in revenue, the conversation goes one of three ways.
One: they politely decline, point you to a Deloitte-affiliated boutique consultancy with a smaller floor, and disengage. This is the most common outcome.
Two: they offer you a "scaled" engagement at $250,000-500,000, which is a stripped version of what enterprise gets, but still a budget-breaker for most mid-market CMOs.
Three: they bundle the AI piece into a larger advisory engagement (strategy, transformation, etc.) that pushes the total contract above the firm's floor. The AI gets delivered, but you're now paying $750,000 for $200,000 of value.
None of these outcomes are good for the mid-market buyer. What you actually need (a working AI surface for your marketing function, integrated with your existing data, with a strategist in the loop) is a small, focused, monthly engagement. Big Four firms can't deliver that. The unit economics don't work.
The vendor categories that fill the gap
The mid-market AI category is being filled by four kinds of vendors right now:
Boutique consulting firms. Smaller shops with leaner overhead can deliver enterprise-grade work at mid-market floors. Engagement quality varies wildly; the best ones are excellent, the worst are people who learned consulting from a McKinsey alum's LinkedIn course.
SaaS-with-AI tools. Tools like HubSpot AI, Salesforce Einstein, and various marketing AI copilots that bolt AI onto existing software. Cost is low ($25-100 per seat per month), but the AI is shallow because it's confined to that vendor's data.
AI-native agencies. Players like Suits.ai showed the wedge: $8,000 per month for competitive intelligence that used to cost $25,000-30,000 in hourly billings from a strategy boutique. The shape is right; the delivery is product-wrapped service.
Operator agencies with their own product. Agencies that built software for their own clients and made it available externally. This is the lane Next Best Action sits in. The agency runs your marketing. The product is the surface. Pricing is engagement-shaped, not enterprise-shaped.
The real comparison
If the question is "should we buy Deloitte Zora or Next Best Action" the answer is almost always Next Best Action, because Deloitte will not sell to you. The real comparison is between the four mid-market vendor categories above.
What we built in Next Best Action sits closest to "operator agencies with their own product." Pricing is $3K-$20K per workspace per month. The agency (Innovative Group) is the delivery team. The product is the surface that converts the engagement from quarterly decks into a real-time conversation. The strategist is in the loop on every meaningful claim.
Compare that to a Deloitte engagement at $500K-$3M with 6-18 month implementation, sold to your CIO, delivered by a rotating cast of consultants who'll be off the project in 8 months. The economics aren't even adjacent. They serve different markets.
"The Big Four built AI for the Fortune 500. We built it for the company the Fortune 500 ignores."
What's coming next
Two trends will compress the gap.
One: the boutique consulting firms are getting better at productizing their delivery. Suits.ai is the early example; expect to see ten more in the next eighteen months. Productized boutique delivery at $5K-$15K per month is the natural pricing band that mid-market companies can absorb.
Two: the Big Four firms will eventually build mid-market product offerings, but their structural cost base means those offerings will land at $50K-$150K annual contracts, which is still 3-5x what mid-market companies will pay for an AI tool. The gap won't close from the top.
The next 24 months belong to the operator agencies and the productized boutiques. The Big Four are not coming for the mid-market. The mid-market is coming for itself.
If you have a Big Four quote in front of you right now and you're trying to compare it apples-to-apples with a mid-market alternative, the comparison isn't about features. It's about whether the engagement floor matches the size of your actual problem. For most mid-market companies, the answer is no.