Anthropic’s Steve Corfield On Claude Frontier Academy, Adding 10,000 AI Engineers

‘We want to be that organization that is genuinely readying the world for what’s next—both customer and partner—as our duty of care to make sure this thing lands properly,’ says Anthropic’s Steve Corfield.

Claude maker Anthropic continues to see its ecosystem of skilled, credentialed practitioners grow, with the artificial intelligence vendor now counting more than 175,000 Claude certifications issued to professionals across 46,000 companies.

About 4,000 people have completed Anthropic’s Basecamp immersive program aimed at training participants on curriculum used by the San Francisco-based AI lab’s applied AI workers. And its new Claude Frontier Academy, launched with a $100 million investment in its first program, looks to add another 10,000 trained engineers by the end of 2027.

“As we move into that more complex enterprise business change piece, there is a skill set that customers are looking for to really help them drive that transformation agenda,” Steve Corfield, channel chief and global head of business development and partnerships at Anthropic, told CRN in an interview. “We want to be that organization that is genuinely readying the world for what’s next—both customer and partner—as our duty of care to make sure this thing lands properly. … Our aim is to be the gold standard for AI skilling.”

[RELATED: Anthropic Forms AI Services Company Amid Claude Partner Network Push]

Geoffroy Pajot, global head of the AI partnerships ecosystem at Capgemini—No. 4 on CRN’s 2026 Solution Provider 500—said in a statement that working alongside Anthropic engineers as part of the academy’s first program, which includes a 12-week Frontier Deployed Engineer (FDE) Residency, will result in improved delivery for clients.

“What I’m hearing from clients isn’t about certification counts or training volume,” Pajot said. “It’s about knowing firsthand how to solve real-world problems that create value to our clients.”

Anthropic’s ecosystem growth comes after reporting in August that it had about 85,000 total certifications issued across nearly 2,000 partner companies.

The vendor also had at the time about 75,000 certified consultants in its ecosystem, a major jump from the 36,000 across 1,300 organizations disclosed in July and the 10,000 reported in June.

Now, the Claude Frontier Academy will serve “as the home for how the world will learn to work with Claude,” Corfield told CRN. It includes Anthropic’s online learning platforms, certifications and hands-on expert programs, among other resources for partners and customers participating.

Here’s more of what Corfield had to say about building out Anthropic’s credentialing motion and partner ecosystem.

What do you want Anthropic solution providers to know about the Claude Frontier Academy?

[We’ve worked on] the Claude Partner Network. Then we got into things like our approach to enabling the partners and making sure that we have a rich ecosystem that can deliver against our customer needs.

So if you bring that up a level … we’re going to launch the Claude Frontier Academy. Imagine it as the home for how the world will learn to work with Claude.

Deep enablement, hands-on frontier AI work underpinned by credentialing and certification. It is multidimensional by design. It spans our online learning platforms, certifications. … Hands-on expert programs. And the first of a kind FDE—what we’re calling the Frontier Deployed Engineer Residency.

Is the residency only for solution providers?

[It] isn’t just focused on our partners. It’s [also] going to be focused on our customers. We think about … the speed in which this economy is growing … the need for the skills, the need for our customers to be able to adopt the technology.

We want to be that organization that is genuinely readying the world for what’s next—both customer and partner—as our duty of care to make sure this thing lands properly. … Our aim is to be the gold standard for AI skilling.

And I want that to be recognized internationally. We don’t say that with like, ‘We just want a badge.’ We say it because we’re making our exams hard. And we’re making them tricky to pass. And we want the requisite bar to be really, really high. … Our broad academy is going to be underpinned by some significant programs.

And the first program we’re going to invest in is our FDE residency. We’re putting $100 million into the frontier residency.

When you think about why now, why it’s important, as our customers are moving AI into production, it is absolutely incumbent on those that are working in that enterprise to make sure it lands best.

They know the systems. They know the data. They understand the constraints. And what we’re trying to create is a common set of standards for our customers and partners that they need. … We want this to be way more than just this partner program.

It’s customers, partners, internationally recognized. Every geography, every industry, every segment. And how you get the largest global enterprises and these fast-moving digital natives, regardless of the industry you’re in, to be able to go fast and adopt.

The skills gap isn’t a partner problem or a customer problem. It’s a market problem. So we want to build the market that can be sustained by this work that we’re doing.

How did you arrive at the 10,000 goal number of Frontier Deployed Engineers?

When you think about Frontier Deployed Engineers, it’s that intersection between the deep technical understanding, the business understanding and then that imperative to be able to take the frontier technology and do something transformational.

And those skills are not commonplace. So, as part of our qualification criteria, the prerequisite is to have that technical understanding, that business understanding.

And then how do we Claude-ify those other bits to be able to develop those frontier skills that we then bring back into this kind of iterative cycle?

This isn’t one and done. This isn’t just do the test and you’re an FDE. Which I worry about when I hear other announcements. ‘Oh, we’re doing 1 million of these and 5 billion of those.’ And, like, what are you actually doing?

And I actually want this to be the proper first program that does think about cohorts, and it thinks about it being an iterative process because the technology is iterative by nature.

If we can do that, then [we make] that deeper bench of engineers that we can really embed into our clients.

And when you see that somebody is a core Frontier Deployed Engineer, you go, ‘Great. I trust that that person is here to transform my business.’

And this is how when we look forward and we use terms like ‘customer success’ and ‘adoption,’ and we think more about as we move into the more agentic work that’s required, or people building technology and using Claude almost as a platform, you can see that these skills are becoming much more differentiated in that market space.

And because there’s not many of them, we have a duty of care to make sure that there are plenty of them so that this can be diffused across both the partner landscape and the customer landscape.

How does building a program like this for an AI vendor compare to building one for Software as a Service?

It’s not like standard SaaS world. You don’t go in and implement a product, typically. Some of this can be training. Some of this could be the use of Claude Code or how you enable Cowork. … As you get into more business process re-engineering and more complex end-to-end business transformation where, actually, it’s probably not enough to just take a process and make it faster.

Where’s the intelligence to look at that process and refine that process and take something that might be 20 years old, bring it into, A, 2026, but then agent-ify it such that it's not any quicker, but it's more efficient?

And that’s where having this level of competency around the FDE residency is really clear. … Airlines have flight simulators. You don’t get in a 747 unless you know how it works, right? … So how do we get these people to be able to practice deeply in a safe environment while learning from others and then going out and doing real work?

This is where we differentiate because if you think about AI transformation right now, everybody wants to go fast. Typically, it’s high-stakes work. But it also requires a lot of discipline and skill.

What we want to do is make sure that we give that discipline and skill. But also we give it at a scale that doesn't mean it’s prohibitive for partners or customers to be able to do it.

We have to offer scale. This is something that, as we’ve seen, is growing at the speed of most industrial revolutions that take a long time. We’re going to do something probably in 10 years that [other] things have taken 100 years.

For us, it’s, ‘How do we ready the world for that?’ That means that we have to have a full continuum of enablement from our associates to our practitioners to our developers right through to the FDE residency and beyond so that we can cover everybody in that.

Are you building a program for both global systems integrators and boutiques?

We’ve learned a lot from these boutiques. If you think about what got Anthropic this far, it was typically a lot of the boutiques. They developed fearlessly on the platform and learned a lot of things that others just didn’t. Everybody’s catching up to that.

We’ve used our own Ode organization, which … was part of a cohort of investments that came from that space to help develop that FDE program.

The FDE curricula have been through our own FDEs. Everything that we have built has been done with those specialist functions. … You’ll see some of the launch partners in here—people like Accenture and Deloitte and McKinsey—how do we then put it in their hands with their relationships, their scale to go and do that in a much more pervasive way?

Are you starting to see partner-to-partner engagements to bring clients complementary skill sets?

The demand is still great. And it continues to be. So we don’t see huge crossover necessarily.

But we are seeing in some of our bigger engagements now where customers are expecting more of a multidimensional partner play. It won’t just be one. They will bring in an industry or a solution specialist because they’re incredibly good at complex algorithms, should we say, in hedge funding [for example].

We are seeing a much more holistic view of partnering. It’s not one size fits all. We are definitely seeing that. … Dario [Amodei, Anthropic co-founder and CEO] talks about the exponential.

If we continue on this line, for us to be able to deliver that to our customers, we have a duty of care to make sure the world is ready. And one of the restricting things that holds the market back isn’t skills—it isn’t prohibitive in any way for model adoption or using our technology. So we run really hard at this.

As the Anthropic Claude Academy grows, you’ll see a number of pillars coming in. We’re talking about a number of different things we could add in here.

But one of the things we wanted to do was address what we felt was the one that was most game-changing, market-making, what the customers are really demanding. Which is, ‘Yeah, we’ve used Code. And we’ve used this. And we’ve seen some fantastic benefits. But what is it really to now drive full and true transformation in a highly complex environment that has a standard rate of business change? And, by the way, we want to transform to AI.’

So, how do you even think about service governance in those sorts of environments? Let alone have the technology to be able to do it. … If you listen to the market, this was the harder end of the market. And what we have come up with here is pretty unique.