The R/C/T/C Prompt Engineering Methodology
Most people write AI prompts like search queries - a few words, hope for the best. The result: generic, unusable output that requires round after round of revision.
The Role / Context / Task / Constraints methodology fixes this. It gives AI the four pieces of information it needs to produce expert-level output on the first try, regardless of which model you use.
The Four Components
Define who the AI should be. This sets the persona, expertise level, and professional lens through which the AI approaches the task. A well-defined role shapes tone, vocabulary, depth of analysis, and domain-specific thinking.
Example
“You are a senior marketing strategist with 15 years of experience in B2B SaaS, specializing in product-led growth. You have managed marketing teams at companies ranging from Series A startups to Fortune 500 enterprises.”
Why it matters
Without a role, AI defaults to a generic "helpful assistant" persona. With a role, it draws on domain-specific knowledge and communicates like an expert in that field.
Provide the situation, background information, and parameters the AI needs before it starts working. Context includes the target audience, industry, company size, current state, and any relevant constraints.
Example
“Our company is a 50-person B2B SaaS startup at $5M ARR, selling project management software to mid-market engineering teams (100-500 employees). We currently rely heavily on outbound sales and want to build an inbound content engine. Our blog gets 2,000 monthly visitors. Budget is $5K/month for content.”
Why it matters
Context prevents generic output. The more specific your context, the more tailored and actionable the output becomes. AI cannot infer your situation - you have to tell it.
Specify exactly what to produce. Define the deliverable type, format, structure, length, and any specific sections or elements that must be included.
Example
“Create a 90-day content marketing plan with the following structure: (1) Content pillars - 3-4 topic clusters with 5 article ideas each, (2) Publishing calendar - weekly cadence with specific dates, (3) Distribution strategy - channels and tactics for each piece, (4) Success metrics - KPIs to track with specific targets for month 1, 2, and 3.”
Why it matters
A vague task ("write a marketing plan") produces vague output. A specific task with clear deliverables, formats, and structure produces output you can actually use.
Set guardrails and quality requirements. Constraints tell the AI what to avoid, what tone to maintain, what compliance rules to follow, and what quality standards the output must meet.
Example
“Constraints: Do not suggest paid advertising tactics (out of scope). Keep language at a practitioner level - no jargon-free summaries for executives. Every article idea must target a keyword with at least 200 monthly searches. Format the plan as a table, not prose. Include estimated time-to-produce for each content piece.”
Why it matters
Constraints are what separate good output from great output. They prevent the AI from going off-track, ensure the output meets your specific quality bar, and eliminate the need for manual revision.
Putting It Together
When you combine all four components, you get a prompt that leaves nothing to chance. The AI knows who it is, what situation it is in, what it needs to produce, and what quality standards to meet. The output is specific, actionable, and professional - on the first try.
This is the methodology behind every prompt in PromptVault. We have applied R/C/T/C across prompts spanning marketing, sales, coding, legal, finance, HR, and more.
Skip the learning curve
Every PromptVault prompt already follows this methodology. Copy, paste, get results.