When you onboard to Rankry, we automatically generate 50 prompts tailored to your category and audience. These are buyer-intent queries built around your niche, language, and competitive landscape. They work right out of the box, and your first reports are built on them.
But after a few weeks, once you’ve studied the data, spotted gaps in coverage, and developed a clearer picture of which queries actually matter for your market, the moment for customization arrives. You notice your customers are asking “CRM for a remote-first startup with 10 people,” not the generic “best CRM for startups.” You see an entire audience segment that isn’t covered.
Markets shift too. New categories emerge, buyer personas evolve. Your Semantic Core should evolve with your market.
What You Can Now Do with Semantic Core
The 50 auto-generated prompts you get at onboarding are no longer a fixed set. They’re a starting point you fully control. The total limit on the Base plan is 100 prompts: 50 of ours + up to 50 of yours.
Edit and Expand
Any of the 50 auto-generated prompts can be edited, deleted, or rephrased. If your buyers speak a different language, you can rewrite prompts in Ukrainian, Spanish, German, or any other language.
On top of the auto-generated ones, you can add your own prompts up to your plan’s limit. Organize them into custom categories, set priorities: core prompts for key queries, experimental for hypotheses.
Prompt Lab to Core (coming soon)
The next step we’re working on: the ability to move findings from Prompt Lab directly into Semantic Core. Test a hypothesis, discover an interesting pattern, add the prompt to weekly tracking in one action. For now, you can do this manually through Semantic Core settings.
When Customization Becomes Necessary
Auto-generated prompts cover the broad picture of your category. But after a few weeks of working with reports, you start seeing zones that need deeper exploration.
Different Audience Segments Ask Different Questions
A marketing team at a project management SaaS serves two audiences: remote startups of 10-50 people and enterprise teams of 500+. These segments ask AI differently. The startup searches for “simple tool for a small team.” Enterprise asks about compliance, SSO, and scalability.
If you track both segments with one mixed prompt set, visibility scores average out and insights lose specificity. With customization, you can create separate categories: 30 prompts for remote startups, 30 for enterprise, 20 general. Each segment gets its own analytics.
Multilingual Markets Require Multilingual Tracking
An e-commerce brand sells in the US and Ukraine. Without customization, they only see the English picture. With customization: 50 prompts in English for the US audience, 50 in Ukrainian for the Ukrainian market. Two parallel data slices, separate insights for each.
AI models respond differently to the same questions depending on language. Competitors rank differently too. Without multilingual tracking, you’re only seeing half the picture.
Prompt Lab Hypotheses Deserve Permanent Tracking
A Head of Marketing notices that buyers in ChatGPT are increasingly asking about integrations. They test in Prompt Lab: “Best CRM that integrates with Slack,” “CRM for teams using Notion and Figma,” “Sales tools with native Zapier support.” They discover that one specific integration angle is an invisible zone for their brand.
Next step: add these prompts to Semantic Core through settings so they appear in every weekly report. Marketing builds a content strategy around this zone and tracks progress week over week.
What’s Available
Base plan: up to 100 prompts in Semantic Core (50 auto-generated + up to 50 custom), full editing, custom categories.
Your Semantic Core should reflect the reality of your market. Now it does.
FAQ
Can I restore auto-generated prompts after deleting them? Yes. Auto-generated prompts are saved as a template. If you deleted a prompt and want it back, it’s available in the default prompts library for your category.
What’s the difference between Prompt Lab and Semantic Core? Prompt Lab is a tool for real-time experimentation: quickly test a hypothesis, see how the model responds to a specific query. Semantic Core is systematic weekly tracking: every prompt runs through all 5 models as part of each weekly report. Soon you’ll be able to move findings from Prompt Lab directly into Core, but for now this is done manually through settings.
Do I need to reconfigure Semantic Core if I change my market or product? If your positioning changes significantly, it makes sense to rebuild Core from scratch. You can request a fresh auto-generation of 50 base prompts for the new category and then customize them to fit your needs.
