The AI Consulting Playbook: How OpenAI is building the Mckinsey of AI
How OpenAI and a 15-Person Startup Are Quietly Capturing a Multi-Billion Dollar AI Market
Hey everyone, on this article you’ll find:
Why OpenAI is moving from pure product to enterprise consulting
How “AI-native consulting” works (and why it looks like Palantir)
The playbook startups like Every are using to scale past $1M in revenue
The 5 key lessons from OpenAI and Every:
Consulting as GTM
Services force upfront commitment
Embedding into workflows
Escaping platform lock-in
Deploying experts fast
Don’t forget to subscribe to stay on top of the opportunities AI is bringing to the world
Want to get a personal newsletter tailored to your needs?
Subscribe to Alfred, the AI newsletter that sends you resources, news and connections so you build your company.
1. Why OpenAI is moving from pure product to enterprise consulting
You may remember that a couple months ago I talked about why I thought that an AI-native consulting firm could become the next McKinsey, and capture a multi-billion dollar gap in the mid-market.
If you’re in tech, it feels like AI is everywhere. But that’s a bubble. Most large companies still haven’t done anything with it.
According to Elad Gil (the investor behind Airbnb, Stripe, Coinbase and Anduril), “AI is dramatically under-hyped because most enterprises have not done anything with it, yet. And that's where all the money is.”
OpenAI knows it. And they’re moving fast to be the ones who capture that wave.
According to Ramp’s transaction data: 40.1% of U.S. businesses were already paying for AI tools as of April. And from those, 32.4% of spend goes to OpenAI. That’s up from 18.9% in January. A massive jump. Their competitors haven’t kept up.
However, Sam A and his team realised that, to sell into big companies, your product alone isn’t enough. Your company needs to know how to navigate the “enterprise maze.”
Enterprise AI isn’t a product. It’s a services business in disguise.
You can build the best model in the world. But if your customer’s internal systems are chaotic, you’re not handing them a solution. You’re handing them a problem.
That’s why OpenAI launched its AI Enterprise Consulting unit
2. How “AI-native consulting” works (and why it looks like Palantir)
OpenAI is embedding engineers directly inside customer orgs to fine-tune, build custom tools, and ensure GPT-4o actually works in real workflows.
But this wasn’t just a pivot. It was a quiet admission that models alone aren’t enough to win in enterprise.
What matters is deployment, making AI actually work inside the chaos of big orgs.
And it looks a lot like Palantir: FDEs, tailored work, long-term lock-in. Mira Murati’s Thinking Machines Labs is running the same blueprint, Scale AI did as well, and it seems every big AI company will end up following this playbook.
But this isn’t just for billion-dollar platforms like OpenAI, there are some startups already capitalising this opportunity, like Every.
3. The playbook startups like Every are using to scale past $1M in revenue
Every is a 15-person AI-native startup that ships internal AI tools, publishes a daily newsletter, and now runs a fast-growing consulting arm.
They are already generating over $1M in revenue through their consulting arm. In less than nine months.
And it’s on track to double this year, according to their CEO Dan Shipper. All while their engineers write almost no code.
4. The 5 key lessons from OpenAI and Every:
1. Consulting as GTM: enter with services, stay with product
This is the loop:
Start with consulting
Map internal pain points
Train teams with tailored AI playbooks
Build internal automations on top
Stay embedded as infra
Services don’t just bring revenue. They drive adoption, which turns into long-term ARR later.
Palantir made $1.1B in 2023 with this model. OpenAI is now using the same playbook through its new enterprise consulting arm.
The Startup Every did it the other way around.
They started by helping teams explore how to apply AI. But over time, their offering has evolved.
After the initial training phase, teams kept coming back, asking for help with small automations, workflow fixes, and internal tools
It quickly became clear this wasn’t a one-off service. They were getting pulled deeper into core operations, becoming part of how the company actually worked.
That demand revealed a bigger opportunity, and became the foundation of a $1M+ consulting business.
They’d stay on to build internal tooling, refine automations, and help maintain the systems they set up; turning what began as training into long-term operational support.
Not a traditional SaaS product, but repeatable, sticky, and hard to rip out. The more embedded they get, the harder they are to replace.
2. Services force upfront commitment
Startups live in fear of endless enterprise pilots that never convert.
OpenAI found the solution: a $10M entry fee that kills “pilot purgatory”.
It forces commitment from day one. No internal debates. No workshops that lead nowhere. Just paid access and a clear path to real deployment.
Every doesn’t charge $10M, but the logic holds.
Their playbook starts with a paid engagement: they go into the company, run in-depth interviews with teams, and map where AI can actually help.
Then they deliver a detailed report, a custom dashboard, and even a chatbot trained on those internal conversations.
That’s what turns vague curiosity into concrete action, and what lets them skip months of selling and stay rooted from day one.
3. Embed your product into your client’s workflows to become irreplaceable
Remember what we talked about earlier this week? OpenAI’s real moat isn’t just model quality, nor having the best API. It’s memory and context.
The deeper your system understands a customer’s workflows, data, and decisions, the harder you are to rip out. Churn isn’t a risk when removing you breaks the way work gets done.
That’s the game OpenAI is playing, and Every plays it too.
Through their consulting work, Every integrates directly into clients’ day-to-day operations.
They don’t just deliver insights, they turn those into internal dashboards, AI playbooks, and custom chatbots trained on company-specific interviews.
These tools stay behind and get used in real work: surfacing pain points, guiding prompt usage, and helping teams adopt AI where it actually matters.
The result? They aren’t just offering advice. They’re wiring themselves into how the company works.
4. Services bypass platform lock-in
OpenAI’s product revenue sends 20% straight to Microsoft through Azure until 2030. Services don’t.
Going back again to my last newsletter, this is how OpenAI escapes Microsoft’s moat: they package fine-tuning and workflow workarounds as “consulting.”
OpenAI keeps the margin and, more importantly, bypasses platform dependencies and keeps the direct line to the customer. Classic channel-partner judo.
Every isn’t beholden to any infra partner either. Their consulting work is fully direct. They keep 100% of the revenue—and more importantly, they own the full customer relationship.
5. Deploy experts to ship fast and stay in the loop.
Palantir pioneered this. Their Forward Deployed Engineers (FDEs) didn’t just write code. They embedded inside client teams, built tools on-site, and turned messy workflows into software.
OpenAI is following the same play.
The services arm isn’t about selling access to GPT-4o. It’s about deploying engineers who can ship inside the client’s chaos—fast.
Every does it too, just leaner. Their consulting team runs four-week engagements, with just one hour per week per team.
But it’s not off-the-shelf content. They tailor everything based on what they learn in the discovery interviews.
Their framework is simple: 10% of employees are naturally curious, 10% will resist, and 80% just need to be shown how AI helps with their job. So the team builds hands-on playbooks: “Here are the prompts for your role. Here’s where and when to use them.”
If you enjoyed this article don’t forget to subscribe and share it with other investors and founders.
FAQs
1. Why is OpenAI moving from pure product to enterprise consulting?
OpenAI realized that selling large enterprises an API or model isn’t enough. Most companies still haven’t implemented AI because their internal systems are chaotic. By launching an enterprise consulting unit, OpenAI can embed engineers inside client organizations, fine-tune models, and build custom tools that work in real workflows. This makes adoption faster and more reliable.
2. What does “AI-native consulting” mean?
AI-native consulting is when a company doesn’t just provide an AI product, but also integrates directly into a client’s workflows to ensure adoption. Consultants and engineers act like extensions of the client’s team, embedding AI into daily processes. This approach is similar to Palantir’s forward-deployed engineers who build custom software inside client organizations.
3. How is OpenAI’s enterprise strategy similar to Palantir’s?
Like Palantir, OpenAI deploys engineers inside client organizations to tailor solutions, build internal tools, and stay deeply embedded. This creates long-term lock-in because once AI systems become part of a company’s core workflows, removing them would disrupt operations.
4. How are startups like Every using this model to grow revenue?
Every, a 15-person AI-native startup, has built a $1M+ consulting business by helping companies adopt AI through training, workflow integration, and custom tools. Their playbook starts with discovery interviews, followed by tailored dashboards, AI playbooks, and chatbots. This approach has allowed them to double revenue in less than a year.
5. What is the consulting-as-go-to-market (GTM) strategy?
Consulting as GTM means using services to enter a client, map pain points, and build trust. Once embedded, companies can upsell automation, internal tools, and infrastructure. This creates recurring revenue and long-term relationships, rather than one-off consulting engagements.
6. Why do AI consulting services require upfront commitment?
Services like OpenAI’s enterprise consulting or Every’s discovery engagements prevent “pilot purgatory”—endless tests that never scale. OpenAI requires a $10M entry fee for enterprises, forcing commitment. Startups like Every use paid engagements that quickly turn curiosity into real adoption.
7. How do AI consulting firms make themselves irreplaceable?
By embedding directly into workflows, AI consulting firms ensure that their tools, dashboards, and automations become part of daily operations. This makes it difficult for clients to remove them without disrupting work, creating high switching costs and long-term retention.
8. How does consulting help OpenAI bypass platform lock-in with Microsoft?
OpenAI’s product revenue depends on Microsoft Azure, but consulting services don’t. By packaging workflow integration and fine-tuning as consulting, OpenAI keeps 100% of the margin and controls the customer relationship directly, without sharing revenue with Microsoft.
9. What role do forward-deployed engineers (FDEs) play in AI consulting?
Forward-deployed engineers are embedded experts who work inside client teams. They don’t just deliver AI models; they customize solutions, ship tools quickly, and ensure adoption. This model, pioneered by Palantir, is now being replicated by OpenAI and AI-native startups like Every.
10. What are the key lessons from OpenAI and Every for founders?
The five biggest takeaways are:
Use consulting as GTM to drive product adoption.
Require upfront commitment to avoid endless pilots.
Embed AI into workflows to create stickiness.
Use services to bypass platform lock-in and keep customer control.
Deploy experts quickly to ensure adoption and long-term value.