Waymaker
Your AI cofounder can write five launch plans. Run the one you can learn from by Friday
Give competing ideas a deadline, an observable buyer action, and a reason to stop.
Imagine a founder with one weekend available and three plausible offers for independent photographers. An agent proposes a referral newsletter, a client onboarding portal, and an inquiry follow-up service. Each pitch sounds viable. The founder’s constraint is access: she personally knows four photographers and has no subscriber list.
A useful launch plan starts with the people who can tell you you’re wrong. For each idea, write the action someone could take this week. A compliment is easy to give all three. Sharing an anonymized inquiry that went cold requires the photographer to recognize the follow-up problem in their own work.
Choose an experiment that could disappoint you
For the follow-up service, prepare one before-and-after email using invented details. Invite a photographer to choose a past inquiry and tell you where the buyer stopped. Offer to draft a reply they can edit and send. Set a Friday decision in advance. If nobody will even discuss a real example, interview them about inquiry handling before writing more templates.
Ask your AI partner to keep a record with four columns: invitation, buyer action, exact objection, next decision. Anthropic’s guidance on agent evaluations makes a useful distinction between the path an agent takes and the result left behind. Apply that principle here: a finished launch plan is an artifact, while a buyer volunteering a case is a change in the world you can observe. Anthropic • Evaluating agent outcomes
After Friday, read the objections together. “We have plenty of replies but no time to quote” points to a different service. “I have another inquiry you can help me answer” earns a second delivery. Either answer is worth more than a sixth plan.