We scan new podcasts and send you the top 5 insights daily.
Hardcoding prompts into code repositories creates an unnecessary bottleneck where product managers author text for engineers to copy and paste. Moving prompt tuning and evaluations to dedicated internal tooling increases operational velocity by allowing in-house domain specialists to directly configure agent behavior.
To maximize leverage, reframe every SDLC component—docs, tests, review agents—as a way to 'prompt inject' non-functional requirements into the agent. This approach teases out expert knowledge from engineers' heads and makes it part of the automated system, guided by the agent's mistakes.
Prompts are written in English and encapsulate the AI's core logic and personality. It is a mistake to treat them as code firewalled within the engineering team. Product managers, as domain experts, should have direct access to edit and experiment with prompts through user-friendly admin interfaces.
The transformative power of AI agents is unlocked by professionals with deep domain knowledge who can craft highly specific, iterative prompts and integrate the agent into a valid workflow. The technology itself does not compensate for a lack of expertise or flawed underlying processes.
By defining the entire software factory (agents, prompts, configs) as code, teams can version it like any other software. This allows for replaying past tasks with new configurations to test for improvements and enables AI agents to directly modify the factory's code to self-improve.
Complex prompting is a transitional phase for AI interaction, not the end state. Truly useful AI tools will abstract this complexity away, using agents to translate user intent into optimal prompts. The focus should be on creating intuitive, directorial controls rather than teaching users to be prompt engineers.
An AI coding agent's performance is driven more by its "harness"—the system for prompting, tool access, and context management—than the underlying foundation model. This orchestration layer is where products create their unique value and where the most critical engineering work lies.
Instead of manually crafting complex "mega prompts" or training rules for AI assistants, ask the AI to generate them for you. You can have a dialogue with the AI to refine its suggestions, dramatically speeding up the process of creating sophisticated workflows.
Expert prompting is less about the final text command and more about architecting the entire context available to the AI. This includes building the right tooling ("harness"), defining custom "skills," and providing relevant data. Small prompts can seem magical only because of this extensive prep work.
Instead of locking prompts in code repositories managed by engineers, empower PMs to own and iterate on them. This treats prompts as a core product component, ensuring AI behavior directly serves user needs and business strategy, as practiced at Watermark.
Early on, a central AI team managed a single, complex few-shot prompt, creating a bottleneck. The key shift was to a tool-calling architecture where individual product teams own their agent's tools and definitions. This distributed ownership, enabled by strong evaluation frameworks, dramatically increased development velocity.