What this is
You already know the 4Cs. You've done some version of Company, Category, Consumer, and Culture analysis on every engagement you've ever worked on. The goal is to get incisive questions across each area to focus your research and ultimately start collecting the building blocks for an insightful strategic foundation.
A well-prompted model can generate tailored exploration questions across all four Cs in seconds, tuned to the specific brand, category, and problem you're working on. The questions are inputs to give your team a running start at designing research and surface angles you might not have considered.
It is important to explain the context of a 4Cs process to whichever model you are using. The way I do it is by defining each part of the process and giving the model a few examples of questions we ask before prompting the instance to assist in the process.
The approach is two steps. First, you prime the model by explaining the 4Cs process and giving it example questions that show the quality bar. Then you give it your brief and let it generate. The priming is what makes the difference between generic textbook questions and ones that would make a junior strategist stop and think.
Step 1: Prime the model
Start a new conversation in Claude, ChatGPT, Gemini, or whatever model you use. Paste the following as your first message. This teaches the model what each C is, what kind of questions belong in each, and what good looks like.
Step 2: Give it the brief
Once the model has the framework in context, send your brief along with instructions for what you want back. Paste the following, replacing the placeholder with your actual brief.
That's it. Two messages, and you have a tailored set of 20 exploration questions to start working from.
Why two steps
The priming step is doing most of the work. Without it, the model defaults to generic strategy questions it's seen in business school case studies. By giving it your definitions and examples first, you're doing three things:
Defining what each lens is actually looking for — not textbook definitions, but the kind of inquiry your team runs. Setting a quality bar through the examples — the model sees the range and depth you expect. And separating the framework from the brief so the model doesn't just pattern-match your examples to the new client.
That separation is the key move. If you put everything in one prompt, the model tends to anchor on the examples and give you rephrased versions. When you prime first and prompt second, it internalizes the quality standard and generates from there.
Example
The model will generate 20 tailored questions (5 per C) specific to this brief, calibrated to the depth and range shown in the priming examples but original to this problem.
This is the lightest version
What you see above is a two-message conversation that works and produces useful output. But it's also the simplest possible implementation.
If you want to go further, these tools can be built to do significantly more: persistent memory across all of your projects so the model learns your team's conventions and past work, direct connections to your research databases and data tools like GWI or YouGov, structured workflows that move from 4Cs exploration through audience development through comms planning in a connected sequence, and team-wide deployment so every strategist has the same foundation.
That's a different conversation. If you want to explore what that looks like for your team, get in touch.