AI can help you think through an audience, sharpen an offer, prepare content, and learn from a test. It becomes much more useful when you give it a specific decision and reliable context.
The Founder Marketing Moves resource contains 40 original prompts for four kinds of work: Discover, Decide, Share, and Improve. It is designed for people building a service, a product, a workshop, or another small business. You can use it with your existing assistant and planning tools.
It does not require an automated agent system. It also does not imply that every research or marketing action described is already built into Waymaker.
Start with the uncertainty, not the prompt collection
Ask what is stopping the next useful action. Do you need a clearer problem, a more understandable offer, a way to reach people, or an explanation of what happened?
| Your uncertainty | Start here | What to keep |
|---|---|---|
| I do not know whom to help | Discover prompts | A focused question and a research plan |
| My offer is hard to explain | Decide prompts | An offer statement and evidence gaps |
| I do not know where to share it | Share prompts | One relevant distribution experiment |
| I have activity but no clear learning | Improve prompts | A grounded decision and next checkpoint |
One good prompt session can be enough for the day. Producing forty unfinished plans rarely helps you move.
Give your AI partner a small context pack
Explain what you are building, whom you think it helps, what you can deliver, and the time or budget you can invest. Add actual customer notes, examples, and past attempts when available. Mark your beliefs as assumptions.
Include this instruction: “Use only the evidence I provide or sources you actually retrieve. Separate facts, interpretations, and hypotheses. Ask for missing information when it changes the decision. Do not invent customers, results, or completed actions.”
Avoid pasting identifiable customer records unless the tool and task are appropriate for that data. Often an anonymized summary is enough to reason about the problem.
Follow one example from idea to learning
Suppose you want to help independent tutors prepare clearer parent updates. You have heard a few tutors describe the task as repetitive, but you have not established whether they would pay for help.
Discover: use WM-D01 to review the actual conversations and WM-D02 to prepare neutral follow-up questions. Find out when the task occurs, how they solve it today, and what makes it difficult. Do not replace those conversations with simulated customers.
Decide: use WM-C01 to write a narrow pilot offer and WM-C02 to identify who is ready to try it. A useful segment might be tutors who said they send recurring parent updates and volunteered to discuss a pilot. It should not be “all tutors who probably dislike admin.”
Share: use WM-S01 to create a demonstration with fictional sample data. Use WM-S03 to plan an appropriate community contribution. Show a useful before-and-after example, and follow the community's rules before including a promotional link.
Improve: use WM-I01 to define evidence before the test, then WM-I02 to inspect the response. If people save the example but nobody asks about the service, you may have a helpful resource without a validated paid offer. Ask what they would pay to avoid doing and what they prefer to handle themselves.
This is an illustrative workflow. The test's outcome is unknown until you run it.
Build a style guide from real choices
AI often produces generic copy because the instructions are generic. Give it approved examples, a clear audience, the promise you can support, and wording that does not fit you.
Use WM-D07 to create a short voice guide and WM-D08 to test whether your main story is understandable. Keep your own judgment in the process: which phrase sounds natural, which detail matters, and which claim you can defend.
A preference profile can describe what you have learned about an audience. It is not a machine that reliably predicts purchase decisions before you test them.
Help different people differently
An explorer, a potential buyer, and an existing customer should not necessarily receive the same next step. Use declared needs and reliable behavior to offer appropriate help.
Read the founder's segmentation guide before building complicated journeys. Start with two or three understandable groups, handle missing information, and stop purchase reminders when someone buys.
Let a prompt produce a working artifact
Keep outputs in a form you can use: an interview script, an offer card, a segment rule, a demonstration outline, a distribution plan, or an experiment note. Give the artifact a date and note what evidence informed it.
Connect it to your next task and checkpoint. If you already use Waymaker for planning, bring the reviewed output into that workflow using the capabilities available in your account. If not, a document works. This article does not depend on a new integration being installed.
Know which work an assistant has actually done
“Here are interview questions” is not “customers were interviewed.” “Here is a campaign plan” is not “the campaign launched.” “Here is a suggested audience” is not “the audience exists and can be reached.”
Ask the assistant to show its sources and list missing inputs. When it recommends a tool, confirm that the capability, access, cost, and integration fit your situation before buying or connecting anything.
For research about competitors, separate a visible claim from your interpretation. For AI search, observe actual answers rather than assuming an optimized paragraph will be cited. For performance analysis, supply actual results with time windows and definitions.
Review the result with four questions
- Is it grounded in information we actually have?
- Can I explain the recommendation without repeating the AI's confident phrasing?
- Is the next action small enough to carry out and useful enough to learn from?
- What evidence would make me change the decision?
Do not demand certainty from a small experiment. A few responses can reveal confusing language or a recurring obstacle, but they do not establish a reliable market forecast.
Make the next session easier
Save the goal, action, observation, uncertainty, and next move. Reuse the approved context instead of starting from scratch. Update it when the offer, audience, or evidence changes.
Download Founder Marketing Moves and choose the prompt that resolves your nearest uncertainty. The aim is a useful decision followed by real work.
Reference note: HubSpot's public Loop Marketing prompt-library resource informed the request for a staged resource. The 40 Waymaker prompts and this founder workflow are original. They do not reproduce HubSpot's collection or claim its reported performance. Checked September 27, 2026.
Ready to put this into action? Waymaker helps you go from idea to your first paying customer, with AI doing the heavy lifting alongside you.
Start building free