Setting up a project isn't just filling out a form. At almost every step, Gumshoe is actually asking AI models real questions about your brand and using their answers to build what you see. Each step is a small diagnostic in its own right, and everything it surfaces becomes part of what your project asks AI models when your Visibility Audit actually runs.
Here's what's really happening at each step:
- Confirm your brand, geography, and language: the AI-written summary of your brand isn't boilerplate, it's Gumshoe asking AI models to describe your brand in their own words. If something reads as off or outdated, that's a real clue about how AI currently understands you, and editing it here corrects the record before anything else is built on top of it.
- Select your product category: AI models are asked what they see as your business's product categories, and their answers become your options. Which category you pick shapes every persona and prompt that follows.
- Choose your plan: this step sets the depth of your ongoing tracking, more personas, more prompts, more models, and more frequent runs on higher plans, since deeper access to AI feedback means a clearer, more reliable picture.
- Review your personas and prompts: personas represent who AI models think your ideal customers are, built from the same diagnostic process. Their prompts are the actual questions AI models expect those customers to ask. Reviewing them shows you what AI already believes about your audience, and editing them refines what your project will go on to ask.
- Select models and start project: the models you choose here are the ones your project will actually query going forward, so this step is where you decide whose perspective you want to track.
Every step is really just Gumshoe asking AI models a piece of the same question: how do you see this brand? What you edit along the way sharpens that picture before your first Visibility Audit runs.