---
id: "action-establish-pov"
type: "action-item"
source_timestamps: ["\\\"§ Step 1. Establish an initial point of view", "so you have a basis for evaluating AI's output.\\\""]
tags: ["workflow", "preparation"]
related: ["framework-four-step-ai-development", "contrarian-friction-is-good"]
speakers: ["David S. Duncan", "Tyler Anderson"]
action: "Form a preliminary hypothesis and scope the task before opening any generative AI tool."
outcome: "Creates a baseline for evaluating and critiquing the AI's eventual output."
sources: ["reskilling"]
sourceVaultSlug: "hbr-seg-reskilling"
originDay: 10
articleStem: "hbr-edu-32-help-employees-get-better-with-ai"
sourceUrl: "https://hbr.org/2026/06/help-employees-get-better-not-just-faster-with-ai"
sourceTitle: "Help Employees Get Better—Not Just Faster—with AI"
---
# Establish an Initial POV Before Prompting

**Action:** Form a preliminary hypothesis and scope the task *before* opening any generative AI tool.

Before touching AI, explicitly scope the task — audience, utility criteria — and form a preliminary hypothesis of what a good answer should look like. If the task is entirely unfamiliar, use AI *strictly* to ask what a strong deliverable looks like and what judgment calls are involved, then form your own view based on limited context.

**Outcome:** Creates a baseline for evaluating and critiquing the AI's eventual output. This is [[framework-four-step-ai-development|Step 1]] of the model; it deliberately introduces [[contrarian-friction-is-good|friction]] (see [[quote-friction-is-necessary|the friction quote]]). Its hardest edge case is captured in [[question-junior-employee-baseline|how novices form a valid initial POV]].