Topical Research

Topical Research is where you turn a core idea, brand context and personas into a structured outline of topics and keywords. The output is a four level hierarchy that later feeds SERP Clustering and the Topical Map.

Use this page as a guide for setting up a research run, generating topics, editing the outline, finding AI search gaps, and exporting your work.

Topical Research helps you:

• Expand a core topic into Main Topics and subtopics in up to four levels
• Ground topic generation in real search behavior using live Google and YouTube autocomplete
• Generate scoped keyword sets tied to those topics
• Shape and refine the outline before any clustering or map work starts
• Detect and resolve duplicate topics with semantic analysis
• Find content gaps by comparing your outline against what AI search engines recommend for your buyer personas

The outline you build here becomes the starting point for:

• SERP Clustering
• Topical Map

1. Open your project in Floyi.
2. In the left navigation, select Topical Research.
3. You will see the research workspace for that project, including context inputs and the outline views.

You will get better results if you set up Brand Foundation and at least one persona first.

• Brand profile: Brand Foundation
• Personas: Audience Insights

You can technically run Topical Research without them, but the topics will be less aligned with your positioning and buyers. AI Search Gaps requires at least one buyer persona.

At the top of the Topical Research view you will see several input sections. Use them to frame the run.

Common fields include:

1. Brand Foundation

   • Select the brand profile that applies to this project.
   • Floyi will use this to align topics with your product, positioning and language.

2. Buyer personas

   • Choose one or more personas from Audience Insights.
   • This tells Floyi which buyer journeys to focus on.

3. Core topic

   • Enter the main theme you want to explore, for example
     • “Email deliverability for SaaS”
     • “Revenue recognition for B2B subscription companies”

4. Optional user inputs

   • Additional keywords, topics or entities you want to make sure are considered.
   • Notes about audience stage, product tier or constraints you care about.

After you fill or adjust these sections, make sure you save. Use the edit button in each block to update content, then click save inside that block.

Once your context looks right, move to the Generate Bar at the top of the outline section.

The Generate Bar is a four column panel. Each column represents one step in the generation process.

1. Generate MTs

   • Creates Main Topics from your core topic, brand context and personas.
   • Start here. Review and clean your pillars before expanding further.

2. Generate ST2 + ST3

   • Generates second and third level subtopics together for each Main Topic.
   • Each ST2 gets at least three ST3 children to prove depth.
   • This is the main expansion step after your MTs look right.
   • During this step, Floyi automatically queries Google Web and YouTube autocomplete to ground the AI in real search behavior. See Live Autocomplete below.

3. Generate ST4s

   • Adds a fourth level of subtopics under each ST3.
   • Scoped per ST2, so you can regenerate specific branches without touching others.
   • Use this for granular planning where you need more detail.
   • Google Web autocomplete is used automatically during this step.

4. Generate Keywords

   • Generates keywords attached to the most granular topics.
   • A scope selector lets you choose the target: all leaf nodes, a specific Main Topic, or a specific ST2 branch.
   • If ST4 exists, keywords attach there. If not, they attach to ST3.
   • Google Web autocomplete is used automatically during this step.

Each column in the Generate Bar shows:

• The count of items that will be generated
• The estimated credit cost based on your selected AI model
• A disabled state if the previous step has not been completed yet

You can run these steps in stages. For example, generate MTs, edit them, then generate ST2 + ST3.

Generation runs as a background task. While it runs:

• A progress banner appears at the top showing current progress, such as “Processing 3 of 8 Main Topics”
• You can cancel the task from the banner
• If you leave and come back, the banner picks up where it left off
• The banner dismisses automatically when generation completes

During subtopic and keyword generation, Floyi automatically queries Google Web autocomplete (and YouTube autocomplete for the ST2 + ST3 step) behind the scenes. This grounds your topic map in real search behavior rather than relying solely on what the AI model learned during training.

How it works

1. The AI reviews your topic and decides which autocomplete queries would be most useful.
2. Floyi fetches results from Google Web autocomplete and, during ST2 + ST3 generation, from YouTube autocomplete as well.
3. The AI uses those results to improve its output, surfacing topics that reflect what people are actually searching for.

What each source adds

• Google Web autocomplete surfaces commercial, tool, and product search patterns that reveal how users explore your topic in Google.
• YouTube autocomplete adds a second intent layer with tutorial demand, audience segments (e.g., “for musicians”, “for ecommerce”), and comparison queries that often do not appear in web search.

No action required

This is fully automatic. There are no settings to configure and no extra steps. The AI decides when and what to query, and the results are used to improve the topics it generates. You do not see the autocomplete results directly — they are consumed by the AI and reflected in the quality of the generated topics.

Topical Research offers three ways to look at your outline. Switch between them using the view toggle at the top of the outline section.

The default view. Each Main Topic gets its own vertical column, similar to a Kanban board.

• Columns are independently scrollable and resizable by dragging the right edge.
• Topics appear as a nested tree: ST2 nodes expand to show ST3, which expand to show ST4 and keywords.
• Use Expand All and Collapse All controls at the top left to open or close the entire tree.
• Color coded borders help distinguish Main Topics at a glance.
• Topics added from AI Search Gaps show a teal sparkle badge so you can see which topics came from AI search discovery.
• Best for seeing the shape and balance of your topic map across pillars.

A hierarchical list with collapsible groups and search.

• Main Topics act as sticky section headers.
• Click any group header to expand or collapse its children.
• A search bar at the top filters across all levels and keywords in real time.
• Keywords display inline under their parent topic.
• AI Search Gaps topics are marked with a teal sparkle badge in this view as well.
• Best for scanning through a large outline quickly or finding specific topics.

The traditional flat table with one row per topic or keyword entry. Columns for MT, ST2, ST3, ST4 and keywords.

• This is the same table used for edit mode and bulk row actions.
• Best for detailed row by row editing and manual data entry.

On smaller screens (under 1024px wide), the view automatically switches to Spreadsheet since the Silo layout needs more horizontal space.

Once topics and keywords appear, you can shape them before clustering.

You can edit topics directly without entering a formal edit mode:

• Double click any topic name to rename it inline. Press Enter to confirm or Escape to cancel.
• Right click any topic to open a context menu with Rename and Delete options.
• Deleting a topic removes it and all its children. A confirmation toast appears.
• Changes save automatically.

For bulk editing, switch to Spreadsheet view and click Edit.

Inside edit mode you can:

• Select all rows using the header checkbox.
• Select specific rows using individual checkboxes.
• Add a new row using the plus icon. Fill in the relevant columns, such as MT and ST2, then the keyword.
• Edit text inline by clicking any cell to rename a topic, adjust a label or correct a keyword.
• Delete selected rows to remove topics or keywords that are off scope or duplicates.

When you finish editing:

• Click Save to apply changes.
• Click Cancel to exit edit mode without saving your latest adjustments.

You can repeat this edit cycle as many times as you need. Manual edits do not use credits.

After generating several levels of topics, you may end up with near duplicate names across different branches.

1. Click Check Duplicates in the toolbar above the outline.
2. Floyi runs a semantic analysis that compares all topics and flags pairs that are too similar.
3. Results appear as warning icons next to flagged topics in the Silo and Outline views:
   • Red triangle for strong duplicates (very high similarity). These should be merged or removed.
   • Amber triangle for likely or possible duplicates. Review these and decide.
4. Hover over any warning icon to see the matched topic and similarity percentage.
5. The Summary Stats Bar at the bottom shows the total duplicate count. Click the count to rerun the analysis.

Use this to clean up your outline before moving into clustering, where duplicates would waste credits and create noise.

After you have built your topical outline, AI Search Gaps helps you discover topics that AI search engines recommend for your audience but that your outline does not yet cover. This is a reverse gap analysis: instead of starting from keywords, you start from the questions your buyers actually ask AI assistants like ChatGPT, Google Gemini, and Google AI Mode.

The feature runs a five step pipeline in the background:

1. Generate persona prompts — Floyi uses AI to create 15 conversational questions that your selected buyer persona would realistically type into an AI assistant. These questions reflect the persona’s pain points, goals, journey stage, and vocabulary.

2. Select best prompts — The AI narrows the 15 prompts down to the 5 most diverse and representative queries, balancing journey stages and intent types.

3. Query AI search engines — Floyi sends those 5 prompts to the AI search engines you selected (ChatGPT, Google Gemini, and/or Google AI Mode) and collects the full responses.

4. Extract topics — The AI analyzes all responses and extracts every distinct topic mentioned, formatted as clean topic names suitable for your hierarchy.

5. Match against your research — The AI compares every extracted topic against your existing outline and classifies each as:

   • Gap — a topic your outline does not cover yet, placed at the right level with a suggested parent.
   • Already covered — a topic that matches something already in your research.
   • Not relevant — a topic that does not fit your brand or positioning.

1. Click AI Search Gaps in the toolbar above the outline. This button requires at least one buyer persona and is only available while your research is still editable (before clustering or map work begins).
2. Select a buyer persona from the dropdown. This determines the voice and perspective of the queries.
3. Choose one or more AI search engines to query: ChatGPT, Google Gemini, and/or Google AI Mode. Each engine adds approximately 4 credits per run.
4. Click Find AI Search Gaps to start the run.
5. A progress indicator shows each step as it completes. The run typically takes 2 to 3 minutes. You can close the modal and come back — the results will be waiting.

When the run completes, you see three sections:

• Gap topics — Topics your research does not cover. Each gap shows:

  • The topic name and a short description
  • The suggested hierarchy level (ST2, ST3, or ST4) and parent topic
  • A confidence indicator (high confidence gaps are most likely to be genuine misses)
  • The number of AI search responses that mentioned this topic (frequency)
  • Subtopics that will be added along with the gap (ST3s for ST2 gaps, ST4s for ST3 gaps)

• Already covered — Topics that matched an existing topic in your outline. Expand to see each match.

• Not relevant — Topics that AI search mentioned but that do not fit your brand or positioning. Expand to see the reason.

Gap topics use an opt-out model. All gaps are pre-selected by default.

1. Deselect any topics you do not want to add. Use the checkbox on each row, or click Deselect All to start fresh.
2. Click Add Selected to Research to insert the selected gaps into your outline at their suggested positions.
3. A success banner confirms how many topics were added. If any topic could not be placed (because its suggested parent was not found in the hierarchy), a warning banner lists those topics.
4. Added topics appear in your outline with a teal sparkle badge and are tagged as ai_discovery source so you can distinguish them from your generated topics.

You can also remove a previously added gap topic using the Remove link next to it in the results.

Click New Discovery Run at the bottom of the results to start a fresh analysis. You can run with a different persona or different engines to surface gaps from multiple perspectives.

Previous results are saved. When you reopen the AI Search Gaps modal, Floyi loads the most recent run’s results so you can continue reviewing and adding topics.

The AI Search Gaps button is disabled when:

• No buyer personas exist for the brand (create them in Audience Insights first)
• Topical Research is fixed (you have already moved to clustering, topical map, or topical authority stages)

If you started with topics only, you can add keywords later.

1. Clean your topic hierarchy first, at least down to ST3.
2. Click Generate Keywords in the Generate Bar.
3. A scope selector appears. Choose where to generate keywords:
   • All leaf nodes generates keywords across your entire outline.
   • Specific Main Topic limits generation to one pillar.
   • Specific ST2 limits generation to one branch.
4. Review the new keywords:
   • Remove queries that do not fit your strategy.
   • Add missing variants that matter to your buyers.
   • Fix obvious naming issues.

Scoped generation is useful when you only want to add keywords to a new branch you just created, without regenerating keywords for the entire map.

You can rerun keyword generation after adjusting topics, but remember that each run uses credits.

Click the Export menu (download icon) in the toolbar to see your options:

• Export to XLSX downloads an Excel file that mirrors the outline structure with columns for MT, ST2, ST3, ST4 and keywords.
• Export to Google Sheets creates a live Google Sheet with the same structure. Floyi will prompt you to connect your Google account on first use.

Exports are helpful for:

• Sharing your plan with stakeholders
• Doing additional tagging or notes in a spreadsheet
• Keeping a static snapshot before you move on to clustering and maps

If you update the outline later, export a fresh copy instead of editing an old file.

You can also import an existing outline:

• Click Import from XLSX in the same export menu.
• Upload an Excel file with columns matching the MT, ST2, ST3, ST4 and keyword structure.
• Floyi will load the data into your outline, replacing the current contents.

This is useful when you have keyword research from another tool that you want to organize and expand inside Floyi.

At the bottom of the outline, a stats bar shows a quick inventory of your outline:

• Count of Main Topics, ST2s, ST3s, ST4s, keywords and flagged duplicates.
• If AI Search Gaps topics have been added, a teal counter shows how many gap topics are in the outline.
• The duplicate count is clickable and reruns the analysis.

Use this to check the balance and completeness of your outline at a glance.

Each plan includes a monthly allocation of topical research runs:

Research is tracked per brand. The first time you research a new brand in a billing cycle, it uses one allocation slot. All regenerations on the same brand within the same cycle are free. After your included research is used, additional runs use credits from your balance at the normal rate.

The allocation counter appears in the Summary Stats Bar at the bottom of the page and in Settings > Credits Overview.

Free users can generate Main Topics, ST2, and ST3. ST4 generation and keyword generation are disabled on the free plan. Upgrade to Creator or higher for full research depth.

After your included research allocation is used, or for actions outside the allocation:

• Generate topics at any level
• Generate keywords
• Run duplicate detection
• Run AI Search Gaps (credits are used per AI search engine queried)

Key points:

• When you are within your allocation, the Generate Bar shows Included instead of a credit cost.
• When your allocation is exhausted, the Generate Bar shows the normal credit estimate.
• Each step in the Generate Bar shows a credit estimate before you confirm the run.
• AI Search Gaps shows the approximate credit cost per engine before you start the run (approximately 4 credits per engine selected).
• Editing the outline by hand is always free.
• Larger outlines cost more to generate, especially when using heavier models.

The Generate Bar shows model recommendations for each step:

• Basic models (GPT-5 Mini, Gemini 3 Flash): Lower cost, good for ST4 expansions and keyword generation where volume matters more than nuance.
• Advanced models (Claude Sonnet 4.5, GPT-5, Gemini 3 Pro): Higher cost, better for initial MT and ST2 + ST3 generation where strategic alignment matters most.

Typical approach:

• Use an advanced model for your initial MT and ST2 + ST3 generation.
• Switch to a basic model for ST4 or keyword expansions if you want to conserve credits.

If you are out of credits, Floyi will stop you before starting the job and prompt you to add or manage credits.

Topical Research outputs drive later parts of your Floyi workflow.

• PAA Explorer

  • Dive deeper into any topic with PAA Explorer, which recursively expands People Also Ask questions into a visual tree up to five levels deep. Use it to discover question-based content opportunities, FAQ structures, and subtopic relationships that keyword research alone misses.

• SERP Clustering

  • Send the keyword column or selected rows into SERP Clustering to group by search intent and real SERPs.

• Topical Map

  • Use the combination of outline and clustered keywords to build a clean Topical Map for your site.

• Planning and briefs

  • The themes you validate here influence what appears later in your Planner and briefs.

If you are new to Floyi, a simple pattern is:

1. Set your Brand Foundation and one main persona.
2. Run Topical Research for one core theme.
3. Generate MTs, then ST2 + ST3. Review and clean.
4. Generate keywords with a scoped target.
5. Check duplicates and resolve any issues.
6. Run AI Search Gaps to find topics AI engines recommend that your outline misses.
7. Review gap topics and add the relevant ones.
8. Move those keywords into SERP Clustering for deeper work.

That gives you a focused, strategy aligned topic set instead of a random keyword list.