How AI is blurring the lines between product managers and engineers

The relationship between software engineers and product managers (PMs) has historically been defined by clear role boundaries, with engineers writing code and PMs defining product requirements. Artificial intelligence, particularly large language models (LLMs), is fundamentally changing this dynamic by making product development more accessible to non-technical team members.

The changing landscape: AI applications are increasingly driven by prompt engineering rather than traditional software development, with many companies now entrusting prompt creation to product managers and domain experts rather than engineers.

• Major tech companies are allowing PMs and subject matter experts to directly shape AI applications through custom interfaces
• At companies like Duolingo, language specialists handle prompts, while at Gusto, customer service teams and PMs take the lead
• This shift represents a significant departure from traditional software development practices where engineers controlled most technical aspects

Core components of AI applications: The fundamental building blocks of AI applications are proving to be simpler than traditional software architecture, with an emphasis on prompts and tools rather than complex code.

• Basic LLM applications typically consist of a base model (like GPT-4 or Claude) and prompt templates
• More sophisticated applications use techniques like retrieval augmented generation (RAG) or agent-based systems
• The same codebase can create vastly different applications – from tutoring to legal assistance – simply by changing prompts and tools

The persistence of prompt engineering: Despite predictions to the contrary, prompt engineering is becoming a permanent fixture in AI development.

• Early “prompt hacks” like telling models to “think step by step” are being replaced by more sophisticated prompt design
• The need to clearly communicate goals to AI systems remains essential, regardless of model sophistication
• Future interactions may include multiple media types, but clear instruction specification will remain crucial

Impact on engineering roles: Traditional software engineering is evolving as AI takes on more coding tasks.

• Tools like Github Copilot and Cursor are automating routine coding tasks
• Engineers are spending less time writing code and more time understanding user needs
• The social and judgment aspects of engineering work remain difficult to automate

New tooling requirements: The shift toward prompt-driven development is creating demand for specialized tools and platforms.

• Teams need evaluation frameworks to measure the impact of prompt changes
• User-friendly interfaces are required for non-technical team members
• Observability tools must track data flow and help debug AI systems
• Platforms like Humanloop are emerging to fill these specific needs

Looking ahead: The distinction between product management and engineering roles appears to be diminishing as AI continues to reshape software development practices and processes. This convergence may lead to new hybrid roles that combine technical expertise with strong product understanding and communication skills. Success in AI development will increasingly depend on teams that can effectively bridge the technical-product divide.