promttendo

Promttendo is a term used in some online discussions of prompt engineering to describe the ongoing process of crafting prompts in order to guide a language model toward a desired output. The form appears to be a portmanteau of "prompt" and the Italian gerund suffix "-ando," reflecting a habit in some communities to coin terms with Italian-inflected morphology.

It is not widely standardized or formalized; exact origin is unclear, and usage is mostly found in informal forums, blogs, and social media among practitioners exploring iterative prompting workflows. Promttendo typically connotes an approach rather than a single prompt: practitioners may build a hierarchy of prompts, add meta-prompts, and refine instructions over multiple interactions to steer tone, structure, or factuality.

Common practices described as promttendo include layering instructions (a system prompt, followed by role prompts, followed by task-specific prompts), prompts that request self-critique or verification, and constraints to limit outcomes (length, style, or banned content). The goal is to create a cohesive prompting pipeline that yields more reliable results than a one-shot prompt.

Reception of the term is mixed. Some users find promttendo a handy shorthand for sustained prompting workflows; others view it as jargon-heavy or region-specific. It has to be distinguished from "prompt engineering" in general and from concepts like prompt injection or system prompts, which address different risks and mechanisms.

See also: prompt engineering; chain-of-thought prompting; meta-prompt; prompt injection.