TL;DR — Key Takeaways

  • Small teams. Much bigger impact: Agentic AI enables forward deployed engineers to solve increasingly complex enterprise problems without expanding engineering teams at the same rate.

  • Customization without chaos: AI accelerates development and integration, but uncontrolled customization risks creating fragmented, difficult-to-maintain enterprise applications.

  • Guardrails are the new competitive advantage: Secure architectures, supported extension points and human oversight allow FDE teams to move faster without compromising product integrity.

After more than two decades working on technology modernization across enterprises, I’ve learned that successful modernization is rarely about technology alone. It happens where product capability, engineering execution, business outcomes and customer adoption come together.

I’ve seen that from different sides — making architecture and engineering decisions, working through business and commercial realities, and, most importantly, seeing what it takes to turn a transformation strategy into something that actually works for the customer.

And through all of it, one question: Did we create the outcome the enterprise was looking for?

One challenge has remained remarkably consistent: The gap between what a product provides and what a large enterprise actually needs.

Forward Deployed Engineering has historically been one way to bridge that gap. But Agentic Application Development is changing both the economics of that model and the boundaries around it.

FDE, Traditionally

The Forward Deployed Engineer model emerged to solve a specific problem: The largest, most strategic enterprise customers rarely adopt a product entirely out of the box – their scale, complexity or unique environments make standard product adoption difficult.

The value of the model is proximity. FDEs understand both the product and the customer’s environment, allowing them to identify gaps, solve integration challenges and adapt the solution to real-world requirements quickly.

The role sits at the intersection of customer, engineering and product. FDEs work with the customer to understand what needs to happen, leverage the product’s capabilities and supported extension points to make it work, and bring broader or recurring requirements back to the Product team.

Product owns the platform; FDEs extend it to solve customer-specific needs, without creating a separate product.

Traditionally, this model has allowed enterprises to get more value from the product while keeping the core platform scalable.

Agentic Development changes that equation.

FDE in the Agentic Era: the Customization Surface Has Changed

The biggest change isn’t simply that FDEs can use AI to write code faster. The bigger shift is that AI expands the solutioning surface itself.

From Custom Code to AI-Assisted Solutioning

Not every enterprise will approach AI in the same way. An FDE working with a customer early in its AI journey may rely primarily on AI-assisted development, code generation and automation. A more mature organization may be ready to introduce agents, AI-driven workflows and greater behavioral configuration.

The FDE’s role is to meet the customer where they are, choosing the right combination of product capabilities, configuration, custom development and AI to solve the problem.

What changes is the breadth and speed of the solutioning toolkit. FDEs can move from requirement to prototype, integration and working solution much faster, while still keeping the implementation aligned with the product’s architecture and the customer’s AI maturity.

Agentic Architecture

Agentic applications introduce a different set of considerations than traditional application architectures. But there isn’t a single starting point for every enterprise. Some organizations will already have an AI or agentic framework and established patterns; others will still be deciding what that foundation should look like.

For the FDE, understanding that starting point becomes part of the solutioning process. If the customer already has the right framework, the focus is on integrating with it rather than replacing it and extending the product through supported interfaces. If they don’t, the development platform itself may need to provide that foundation — bringing together the agentic framework, integrations, security, evaluation and operational controls needed to build and run the solution.

A few areas become particularly important.

Architecture & Integration

Traditional applications typically follow predictable data and service flows. Agentic applications introduce dynamic context management and runtime orchestration. Enterprise APIs, services and data sources need to be exposed as well-defined tools that agents can safely use.

Guardrails & Security

With more autonomy comes a different set of controls — from prompt injection and data leakage through context to permissions, execution boundaries and sandboxing. These controls need to be part of the architecture, not added after the solution is built.

Development & Testing

AI-assisted development can accelerate design, code generation, integration and troubleshooting, while agents can take on increasingly complex development tasks. But the more dynamic the application becomes, the more important it is to validate behavior against the customer’s real data, business rules and environment.

Testing also expands beyond traditional functional correctness. FDEs may need customer-specific evaluation datasets, real-world scenarios and edge cases to validate accuracy, behavior and regression in the actual operating environment.

Operations & Human Oversight

Production agentic systems also require visibility into context and token usage, agent loops, semantic drift, unexpected behavior and when human intervention is required. Critical workflows should have explicit human-in-the-loop intervention points.

For the FDE, this changes the solutioning conversation. The question is no longer simply how do we integrate this customer’s environment with the product? What foundation does this customer need, and how do we enable the right level of intelligence and autonomy while keeping the solution secure, observable and within the product’s boundaries?

That creates a natural boundary for the FDE: extend through the platform’s supported interfaces rather than around them.

This is also where the underlying development platform matters. A platform with an integrated agentic framework and deterministic, architecture-first code generation can provide the FDE with a consistent foundation when the customer doesn’t already have one — while allowing customers with an existing framework to integrate rather than replace what they have.

And the more powerful the customization mechanism becomes, the more important the boundaries around it. The goal isn’t infinite customization. It is controlled extensibility — enough flexibility to solve real enterprise problems without creating a different product for every customer.

The FDE / Professional Services Boundary is Changing

As AI accelerates software development, it also changes the economics of how enterprise technology gets delivered. A lean FDE team can take on significantly more technical complexity without scaling linearly with the amount of engineering work.

That makes the distinction between FDE and Professional Services less about who builds and more about where each organization creates leverage.

FDE: Product and Engineering Leverage

FDEs operate close to the product and the customer’s technical environment. They solve complex integration and implementation challenges, extend the product within its architectural boundaries and iterate quickly against real customer requirements.

An FDE team may often be one to three engineers supporting a strategic enterprise. With AI-assisted development and automation, that small team can take on a broader solution surface and move faster from requirement to working software.

The FDE success measure is therefore closely tied to the product: How quickly can we get the software working in the customer’s environment — and does it stick?

Professional Services: Engagement and Business Leverage

Professional Services operates at a broader engagement level. Its role can span implementation, process redesign, adoption, change management, enterprise architecture and other consulting capabilities required to realize the value of the technology.

PS organizations can scale through multiple delivery teams or specialized capabilities depending on the engagement. Their measures may include delivery milestones, engagement outcomes, scope, utilization and ultimately the business outcomes defined for the engagement.

The two models are complementary.

FDE brings deep product and engineering leverage close to the customer. PS brings broader delivery and organizational leverage around the customer.

Agentic Development doesn’t eliminate the need for either. It changes where each can create the most value — allowing FDEs to take on more technical complexity without scaling the team proportionally, while creating an opportunity for Professional Services to focus more of its capacity on the broader organizational and business challenges that sit beyond the product itself.

And the value doesn’t have to stay with one customer. When FDEs identify repeatable patterns or capabilities across implementations, those learnings can feed back into the product, making the platform stronger and future customer engagements faster.

Conclusion

Two decades in, I’ve learned the hardest part is rarely proving something can be built. It’s building it in a way that works for the customer today without making the product harder to evolve tomorrow.

What’s different now is the breadth of what an FDE can influence. With AI and agents, FDEs can shape not just the solution around the product, but how the application behaves from prompts and workflows to agent-driven decisions.

That’s a meaningful shift in the role. FDEs have more leverage to solve complex customer problems, but that leverage also makes product boundaries more important.

Architecture, extension points and guardrails aren’t constraints on FDE speed. They’re what make that speed sustainable.

Get that foundation right, and FDEs can take on more complex enterprise requirements while keeping those solutions aligned with the product rather than creating a growing collection of customer-specific implementations that become difficult to manage, scale and evolve.

Frequently Asked Questions

How is agentic AI changing forward deployed engineering?
Agentic AI expands the capabilities of forward deployed engineers beyond traditional integration and customization. FDEs can now use AI-assisted development, intelligent agents and automated workflows to deliver enterprise solutions faster while adapting to each customer's technical environment and AI maturity.
What's the difference between FDE and professional services in the agentic era?
FDE teams focus on deep technical integration, product extensions and engineering solutions close to the customer. Professional services teams address broader business transformation, implementation, adoption and organizational change. AI increases the technical reach of FDEs while allowing professional services to concentrate on wider business outcomes.
Why are architecture and guardrails essential for AI-powered FDE teams?
Agentic applications introduce new risks involving autonomous decisions, security permissions, data leakage and unpredictable behavior. Clear architectural boundaries, supported integrations, evaluation frameworks and human oversight help organizations accelerate development without creating unmanageable customer-specific solutions.