Last week I showed you the 3-prompt stack — the difference between prompting once and getting generic output versus building toward something you can actually use.

This week: the structure that makes every prompt in that stack work better.

It's called the RCTF Method.

Tell the AI who it is before it writes a word.

Not "act like an expert." Specific: "You are a senior B2B marketing strategist with experience in healthcare SaaS. You write for VP-level buyers who are skeptical of vendor claims and short on time."

The role sets the lens. It determines vocabulary, tone, assumed knowledge, and what the AI thinks counts as a good answer.

Give it the situation it's operating in.

Industry. Audience. Constraints. What's been tried. What's at stake. The context is what separates a generic answer from one that fits your actual problem.

Most professionals skip this. They jump straight to the task and wonder why the output reads like a template.

State the exact deliverable — not the general topic.

Not "help me with my pipeline review." Instead: "Write a 150-word executive summary of my Q2 pipeline for a CRO who needs to understand where we're at risk and what I'm doing about it."

Specific in. Specific out. This is the law.

Tell it exactly how you need to receive the output.

Bullet points or paragraphs? Under 200 words or a full briefing? Table format for a slide deck? Structured sections with headers?

Format controls how useful the output actually is. The best insight in the wrong format doesn't land in the meeting.

Each component eliminates a different type of vagueness.

Role eliminates perspective drift — the AI won't speak to the wrong audience. Context eliminates generic defaults — it can't fall back on template answers when it knows your specifics. Task eliminates scope creep — it won't pad the response when you've named exactly what you need. Format eliminates translation work — you get output you can use, not output you have to reformat.

Skip one and the others pick up some of the slack. Include all four and the output rarely needs a second draft.

Take one prompt you've already sent this week that didn't give you quite what you needed.

Run it through RCTF. Add the Role it was missing. Add the Context it didn't have. Tighten the Task. Specify the Format.

Compare the two outputs.

That gap — between what you got and what RCTF gets you — is what Nova AI is built to close permanently.

The compound workflow. How operators chain RCTF prompts together to produce outputs that would take hours to do manually — in under 15 minutes.

Nova | novaai.media