Applying the four Cs framework and building your data puddle
This guide is based on the Beyond Prompts: How to Avoid “Faster Mediocrity” with AI webinar, published on the AI Lab by ActiveCampaign.
Get the quick-start guide
What will I accomplish with this guide? By the end, you’ll have a strategic filter (the four Cs—CRM, content, channels, customers) you can run any AI-assisted marketing decision through, plus a “data puddle” of unique inputs that pulls your AI output above the average of the open internet. It’s based on the framework Bob Pearson walked Sean Blanda through in this webinar.
Before you start, you’ll need:
- Access to your CRM, your content backlog, your channel performance reports, and any customer interview or research notes you have
- A workspace inside ChatGPT, Claude, or ActiveCampaign Active Intelligence where you can keep files and context across multiple conversations
- A list of the campaigns or marketing decisions you’ll need to make in the next 30 days
Quick reference
- Total time: 60–90 minutes to set up the four Cs filter + initial puddle; ongoing as you add inputs
- Tools needed: Your CRM (e.g., ActiveCampaign), a workspace-capable AI assistant, somewhere to store reference docs (Drive, Notion, or the workspace itself)
- Key output: A four Cs decision filter, a folder of first-party inputs, and a habit of feeding new inputs back into the workspace
Watch this section
For full context on the following topics, watch these sections of the webinar:
- The four Cs framework explained — [15:55]–[18:30]
- CRM and content as the first two Cs — [18:14]–[19:30]
- Channels and customers as the next two Cs — [19:30]–[20:50]
- Return on information vs. return on investment — [20:30]–[22:00]
- Data lakes, ponds, and puddles — [23:20]–[25:30]
The workflow
Phase 1: Set up the four Cs as your decision filter
After this phase, you’ll have: a one-page reference for routing any AI-assisted marketing question through CRM, content, channels, and customers.
- Write down the four Cs as headings: CRM, content, channels, customers. Bob uses these as a strategic filter so AI work stays connected to outcomes instead of producing volume in a vacuum.
- Under CRM, list the inputs you have: pipeline, customer base, deal stages, lifecycle data. This is what AI uses to suggest who to add, who to retain, and what to offer.
- Under content, list what you publish and where it sits: website, newsletter, social, sales enablement. AI helps you produce variants for each segment, not generic copy.
- Under channels, list every channel you’re on plus its performance signal: which channels actually drive customers? Which are noise? Bob’s prompt: get rid of the channels that aren’t producing and double up where they are.
- Under customers, list what you know about them and what you wish you knew: brand voice fit, segment behavior, friction points. The gap between the two columns is your research backlog.
The framework keeps us honest. So we’re not just doing content, we’re not just doing channels. An individual can see the whole picture and have a much better view of how to be strategic.
Phase 2: Build your data puddle
After this phase, you’ll have: a working “puddle” of unique inputs an AI workspace can draw on — proprietary plus open-source — that competitors don’t have.
- Inventory your first-party data: customer interviews, sales call notes, churn surveys, support tickets, win/loss reasons, product roadmap, brand voice docs.
- Inventory the public data that complements it: category reports, competitor messaging, channel benchmarks, trend research. Bob’s framing — proprietary plus open source equals decisions your competitors can’t replicate.
- Pick a workspace that accepts files and keeps memory: a ChatGPT or Claude project, an Active Intelligence workspace inside ActiveCampaign. Pick one and stick with it so the puddle deepens over time.
- Upload the first batch of inputs: start with the five files that matter most—customer interview notes, brand voice doc, last quarter’s campaign brief, top-performing email examples, and a list of segments.
- Add to it weekly: every customer call, every campaign post-mortem, every win/loss note goes into the puddle. The goal is a brain you keep feeding, not a one-time dump.
Phase 3: Run AI work through the filter and the puddle
After this phase, you’ll have: an AI workflow where every decision is grounded in your unique inputs and routed through the four Cs.
- Pick a real upcoming decision: a campaign brief, a social repurposing job, a re-engagement workflow. Don’t practice on a fake task.
- Open the workspace that holds your puddle: the AI sees your inputs without you re-pasting them.
- State the task and the relevant Cs: “I’m briefing a re-engagement campaign for Segment X. Walk me through it across all four Cs—CRM signals, content angles, channel mix, customer-level personalization.”
- Ask for one improvement per C: “Based on this brief, what’s the one CRM signal I should add? What’s the one content angle I’m missing? Which channel should I drop? What customer detail would change the offer?” Each answer is a sharper version of what you started with.
- Score the work on return on information, not just volume: Bob’s reframe of ROI as “return on information” — did the precision of your decision improve, and can you prove the outcome did too? If not, the chain isn’t tight enough yet.
Phase 4: Make it a weekly habit
After this phase, you’ll have: a working rhythm where the puddle gets deeper every week and your AI output gets less generic with it.
- Pick a recurring marketing task: weekly newsletter, weekly social cutdown, monthly campaign brief. The same one every week.
- Run it through the four Cs filter every time: CRM signals → content angle → channel placement → customer-level variant.
- Drop the new inputs back into the puddle: the brief you wrote, the version that performed, the version that didn’t. The workspace gets smarter on your data.
And subscribe to The AI Lab, engineered by ActiveCampaign for new examples: every issue features another marketer’s workflow you can lift into your own puddle.
Related
More data from the AI Lab.