Prompts to build a Living Knowledge system
These prompts are based on the You don’t have an AI problem. You have a “Dead Knowledge” problem story featuring Bryan Cassady, published on the AI Lab by ActiveCampaign.
Get the prompt template
How to use these prompts
These are ready-to-use prompts pulled from the Living Knowledge workflow Bryan Cassady runs with marketing clients. Copy, paste, and swap in details where you see [BRACKETS]. Every AI tool and model behaves a little differently, so treat what comes back as a starting point—review the output and refine from there.
Prompt 1: Find alignment gaps in interview transcripts
Best for: Surfacing where leadership’s story and front-line reality diverge, before you build anything on top of it.
Use with: NotebookLM, after uploading your interview transcripts.
Use this
What do respondents agree on about what this organization does and who it serves?
Variables to fill in:
- No variables—run this exactly as written against your uploaded transcripts.
What to expect: A summary of shared themes across respondents. Good output names specific points of agreement, not generic statements like “everyone values quality.”
Follow-up prompt
Where do responses contradict each other? Identify the three sharpest disagreements.
What phrases do people use to describe our value that we are not using in our official messaging?
Prompt 2: Build a True N.O.R.T.H. strategic anchor
Best for: Setting a one-page objective before you brief AI on any campaign, proposal, or sprint.
Use with: ChatGPT (True N.O.R.T.H. GPT) or Gemini (True N.O.R.T.H. GEM).
Use this
I would like your help writing a True N.O.R.T.H. for this challenge: [describe your project or campaign objective]. Ask me questions to build this step by step.
Variables to fill in:
- [describe your project or campaign objective]—a plain-language description of what you’re trying to do, e.g., “a tagline campaign for our premium knife brand”
What to expect: The GPT walks you through each letter of the framework—Narrative, Objective, Restrictions, Tactical constraints, Here’s the starting place—and returns a one-page document you can save as the first source in your knowledge system. Takes about 20 minutes.
Prompt 3: Surface competitive gaps
Best for: Turning a folder of saved competitor campaigns into a competitive-intelligence read in an afternoon.
Use with: Gemini, connected to a NotebookLM notebook of uploaded competitor campaigns.
Use this
I’ve uploaded [NUMBER] competitor campaigns into this notebook. Based strictly on what’s there, identify the three angles nobody is using. For each one, explain why it’s absent and whether that absence is a gap or a deliberate choice.
Variables to fill in:
- [NUMBER]—how many competitor campaigns you uploaded, e.g., “12”
What to expect: Three specific angles, each with a stated reason for the gap (oversight vs. deliberate avoidance). If the output stays generic (“focus on quality”), the source library likely needs more campaigns.
Prompt 4: Stress-test a talk or presentation against expert sources
Best for: Getting outside-expert feedback on a keynote or presentation in an unfamiliar domain, without hiring consultants.
Use with: NotebookLM (loaded with your presentation plus several books or sources from domain experts) connected to Gemini.
Use this
From these [NUMBER] authors, give me a review of my speech. What have I missed? What assumptions am I making that an expert would challenge?
Variables to fill in:
- [NUMBER]—how many expert sources you uploaded alongside your presentation, e.g., “five” (Cassady’s own example uses five books by domain experts)
What to expect: Specific gaps or challenged assumptions tied back to the uploaded sources, not generic presentation feedback.
Prompt 5: Evaluate a deliverable before it goes to a client
Best for: Catching contradictions and weak points in a proposal before a senior reviewer—or the client—sees it.
Use with: Gemini, connected to your Living Knowledge notebook.
Use this
Before I share this with the client, evaluate it against everything you know about their preferences, their past feedback, and their stated objectives. What are the three weakest points? What contradicts something they’ve told us? What would make them push back immediately?
Variables to fill in:
- No variables—run this against the specific deliverable and client notebook in context.
What to expect: A short list of concrete weak points and contradictions, ideally referencing specific past feedback or rejected recommendations in the notebook—not generic proofreading notes.
Tips for better results
- Interview your people before you upload any existing documents. Uploading the “official story” first seeds the system with messaging that may not match how customers, sales, or front-line staff actually describe the work.
- Be selective about what goes into the standing library. Skip LinkedIn posts, press releases, and polished PR copy—assume they’re AI-generated or sanitized until proven otherwise. Interview transcripts are more valuable than published content.
- Upload entire folders, not individual files. NotebookLM can handle documents 1,000+ pages long without breaking, so there’s no need to pre-trim sources.
Ready for the full story?
Read You don’t have an AI problem. You have a “Dead Knowledge” problem featuring Bryan Cassady, published on the AI Lab by ActiveCampaign.
Related
More data from the AI Lab.