Structuring STAR Work Accomplishments with Two-Pass AI Prompts
Ads
When you ask an AI model to write resume bullets in a single command, it routinely invents vague phrasing like spearheaded visionary synergy while fabricating metrics. Recruiters skip these generic statements immediately.
By separating the writing workflow into a STAR accomplishments AI prompt run across two passes, you maintain control over factual details while systematically enforcing the Situation, Task, Action, Result framework.
Why Single-Pass Prompting Degrades Resume Quality
Drafting an effective bullet requires two distinct tasks: extracting verifiable facts from raw experience, and formatting those facts into concise, impact-focused syntax.
When an LLM attempts both tasks at the same time, formatting constraints overpower factual precision. The model fills information gaps with corporate clichés. Dividing the process into two steps avoids this:
Pass 1: The Fact Extraction Prompt
Paste your project notes, sprint summaries, or quarterly reviews into your chat window alongside this extraction prompt.
Read my raw project notes below. Do not format them into resume bullets yet. Output a clean list identifying: 1. The business context or technical constraint (Situation/Task) 2. The exact operational tools, code, or workflows I used (Action) 3. The verifiable outcome, metric, or deliverables produced (Result) If my notes lack a verifiable metric or clear action, explicitly list it as [Missing Data]. Raw Notes: [PASTE MESSY PROJECT NOTES HERE]Check the [Missing Data] tags in the response. If the model flags an unmeasured outcome, supply an estimated range or operational change directly in the chat before moving on.
Pass 2: The STAR Formatting Prompt
Once the extraction pass establishes the factual parameters of your accomplishment, run this second prompt within the same conversation window.
Using only the verified data points from Pass 1, generate three distinct bullet options formatted for an executive resume. Adhere to these constraints: 1. Start each bullet with a precise past-tense action verb (e.g., Engineered, Migrated, Negotiated; avoid Spearheaded or Championed). 2. Follow the structure: [Action taken] + [Technical or procedural mechanism] + [Measured business impact]. 3. Do not introduce tools, adjectives, or numbers not present in the extraction data. 4. Keep each bullet under 28 words.Reviewing and Selecting Your Bullets
Select the variation that most accurately reflects your day-to-day work. If the prompt generated an outcome phrasing that feels exaggerated, edit the sentence to match your actual experience.
Save the structured outputs into an offline document alongside the raw extraction notes so you have the verified facts ready for interviews.
Frequently Asked Questions
What if my past project had no quantitative metrics?
Focus on operational changes rather than financial metrics. Documenting that you reduced manual deployment steps, decommissioned legacy servers, or unified cross-departmental documentation provides verifiable proof of impact without inventing numbers.
Why should I ban words like 'spearheaded' or 'championed'?
Recruiters often treat vague verbs as indicators of passive participation. Concrete verbs like configured, authored, refactored, or audited convey clear ownership.
Can this prompt handle non-technical work history?
Yes. The framework works for sales, operations, healthcare, or logistics by focusing on volume, compliance adherence, customer retention, or turnaround times.