a16z: In the AI Era, Companies Compete for Talent by Redefining Job Titles

The value of the title FDE (Forward-Deployed Engineer) lies not in sounding more modern, but in redefining a type of work that was previously undervalued: on-site technical implementation for customers.

In traditional software companies, this type of work is often placed at the periphery of pre-sales, implementation, solutions engineering, or customer success. It is close to both the customer and the product, yet frequently occupies a marginal position in the organizational narrative.

Palantir saw this very early on.

Around 2011, it renamed the previously customer-site-focused and system-integration-oriented engineering role to FDE. This renaming reflected a clear insight: for large enterprises and government clients, the real challenge is not writing the software, but integrating it into the client’s actual business systems—encompassing permissions, data, workflows, legacy systems, and organizational responsibilities.

Someone who can pull this off shouldn't be simply categorized as after-sales support or project implementation.

They represent a new form of organizational capability.

a16z refers to this strategy as "title arbitrage," which can be understood as "title name arbitrage": when a particular skill rapidly gains importance within an organization but existing job titles haven't yet reflected its value, the first person to create a new title has the opportunity to secure talent, power, and market perception.

This approach is fascinating and highly worth emulating, especially for AI founders running B2B businesses.

Many companies underestimate the role of titles.

On the surface, a job title is just a line of text in the HR system. But within the company, it serves as a form of organizational language—it communicates to others what this person is responsible for, what skills they represent, and whether they are qualified to participate in certain decisions.

Titles such as CEO, CTO, and CFO are not merely descriptions of responsibilities—they are also markers of authority. The same applies to titles like Vice President of Manufacturing, Product Lead, or Growth Lead. Behind each name is an acknowledgment by the organization of a specific set of capabilities.

This is also why job titles continuously evolve with changes in the industry.

In the early days, people who wrote code were often categorized as IT. Later, they became known as programmers, and then as software engineers. This shift wasn’t just semantic—it reflected the growing importance of software within business systems. Coding evolved from a back-office support function into a core capability for building products, organizing processes, and shaping business models.

The data role follows a similar path: from clerk to data entry, to data scientist, and finally to machine learning engineer. Each change in title reflects an increasing strategic value of data work.

Google’s site reliability engineer is also a classic example. It redefines the traditional system administrator’s role as an engineering problem, reflecting the judgment that keeping a system stable is as technically demanding as developing new features.

Therefore, job titles are not merely superficial labels.

It reflects whether the value of a category of work has been transferred.

FDE has become a classic case because it transformed customer-site engineering from an underappreciated role into a high-impact position.

In many companies, the role of on-site technical work lacks clarity. It’s too close to sales to be seen as “pure” by engineering teams, and too close to delivery to be viewed as anything other than a cost center by management. As a result, truly outstanding engineering talent may not be inclined to take on this role.

Palantir's name change shifted the narrative.

The message it conveys is: You're not handling routine after-sales support or external project deliveries. You're resolving the most complex issues on-site at the client’s location, connecting real business systems with the company’s products.

This narrative appeals to a hybrid talent: someone who can code and interact with clients, understands systems and navigates organizational complexity, solves immediate problems, and brings on-the-ground customer insights back to product development.

Someone who sees “Implementation Engineer” or “Solution Engineer” might perceive the role as having limited growth potential. But if they see FDE, their perception will be entirely different.

This is the hiring advantage that naming brings.

To this day, when people hear the term FDE, their first thought is still Palantir—not because Palantir is the only company capable of this work, but because it was the first to associate the term with its own capabilities.

The person who names something first often secures mental dominance.

Of course, not all new job titles have value.

Some are merely instances of title inflation—for example, renaming a marketing specialist as a growth strategist without changing the actual responsibilities, or labeling an assistant as a manager without granting any additional decision-making authority. Such naming conventions offer only short-term superficiality and fail to create genuine talent appeal.

The original text provided a good criterion for judgment:

Would someone from five years ago find this new job title unfamiliar?

If the answer is yes, it may truly correspond to a new capability. For example, the GTM engineer proposed by Clay and the legal engineer proposed by Harvey are not merely renamed roles—they point to new combinations that emerge after AI: individuals who understand business processes and automation, who grasp professional contexts and can embed workflows into systems.

But prompt engineering is another example.

This term was once popular but quickly became outdated. The reason is that writing prompts never solidified into a standalone profession—it’s more of a foundational skill that all knowledge workers should master. If a title precedes actual work, its popularity will quickly fade.

Therefore, whether a new job title is valid depends not on how novel it is, but on whether it represents a real, new type of work.

No new work, just new packaging—that’s job inflation.

The most valuable part of this article is that it places job titles within the organizational context of AI transformation.

When many companies discuss AI transformation, the default assumption is that interfaces will become smarter, tools will become more automated, and processes will become more efficient.

These are all valid, but not enough.

A deeper change is that a new group of highly leveraged individuals will emerge within organizations. They may be young and initially hold low-level positions, but because they know how to use AI, build workflows, and turn ambiguous problems into automated systems, they are gaining influence that was previously unavailable to them.

A similar phenomenon occurs whenever large companies introduce critical software.

Those who first understand new tools are often not the highest-ranking individuals, but the fastest to act. They are the first to identify which processes can be restructured, which tasks can be automated, and which problems no one else wanted to handle can be reorganized.

What technology changes goes beyond just the toolbar.

It will also alter the distribution of power within the organization.

At this point, a new title becomes important. It provides legitimacy to these individuals and gives the organization an identification mechanism.

For example, a legal professional who was initially just interested in AI tools and began exploring contract revision, risk control, and legal workflow automation—once the company defines this role as a legal engineer, this person is no longer merely “someone who enjoys experimenting with new tools,” but rather occupies a recognized, empowered, and promotable new position.

The hardest part of AI transformation is often not that employees don’t know how to use the tools, but that the organization lacks the language to recognize those who are already creating new value.

If you're working on AI for business, the insights from this article are very direct.

Don't just name the product—also consider: What new roles will your product create within your customers' organizations?

If you serve the legal industry, the early adopters may no longer just be lawyers or traditional legal operations staff, but legal engineers. If you serve sales and growth teams, you might see GTM engineers emerge. If you serve financial research or consulting, intelligence engineers may appear in the future.

These names are more than just slogans.

They will help clients internally organize and mobilize: who should be authorized, who should be heard, and who represents this new capability.

This is also where title arbitrage adds value for the company.

The product is sold externally, while the job title spreads internally within customer organizations. If a new job title truly gains traction, it will, in turn, build mental association for the product. In the future, when the market thinks of such roles, they will think of who first introduced it, who understands it best, and who can best help these professionals grow.

Palantir benefited from this type of upside on FDE.

Why is FDE worth revisiting today?

Because the boundaries of products and services offered by AI-native companies are becoming increasingly blurred.

It’s not always easy to distinguish whether an AI enterprise software is a pure product, a product with services, or a service that has been productized. On-site customer processes can, in turn, define the product roadmap; model failure cases become capabilities for the next version; and implementation teams are no longer just endpoints for delivery but integral parts of the product learning system.

In this case, the old title may underestimate the new capabilities.

Calling it "after-sales" may discourage engineers from joining; calling it "implementation" may worry investors about gross margins; calling it "customer success" may not make the product team view it as a product signal. But if its essence is transforming complex customer needs into replicable capabilities on-site, then FDE is a more accurate term than the old ones.

Of course, naming is not a universal solution.

Renaming customer success to FDE will not automatically complete the organizational upgrade. What truly needs to change are the reporting lines, incentive structures, hiring criteria, product feedback mechanisms, and how founders view the concept of “service.”

The name is just the first step.

The key question is whether the organization truly places these individuals at the core of product education and customer delivery.

The emergence of a new job title often indicates that old organizational language is no longer sufficient. Many of the challenges AI companies face today are precisely those that old terminology cannot accurately describe: products resemble services, and services resemble products; engineers need to be on-site with customers, while those customer sites are defining the product roadmap; after-sales is no longer just a cost center but part of a learning system.

This could be a key turning point for the next generation of AI enterprise software companies.

It’s not necessarily the one who completely eliminates the service who will win. More likely, it’s the one who can rename, reorganize, and reproductize the parts of the service that most closely address customers’ real problems and generate the deepest product insights—who will build stronger barriers.

The first person to clearly explain this will plant their flag in the customer’s mind.