GeoGen Playbook: How to Improve Your Visibility in AI Responses

This playbook combines narrative guidance with clear, concise checklists for maximum effectiveness. You’ll set up the platform, choose buyer‑style prompts, interpret responses, competitors, and citations data and turn them into actions that move your visibility across AI assistants.

Adding Your First Entity

To see how your brand actually shows up in AI answers, start by adding your entity. The platform interacts with AI models through their publicly accessible interfaces, exactly how a user would, so the results mirror live experiences. Open the top‑left menu, select Select Entity, then Add New Tracking. Choose Website (Default) unless you’re tracking a person, then enter your brand name, domain, and a one‑sentence description.

You can generate that description by clicking the sparkles icon in the Description field. The platform will crawl your site and produce a draft. Treat it as a starting point. Tighten it so it states your offering plainly. A strong description is short, literal, and includes your core product or service; add a region only if it truly matters; skip slogans.

Prefer to write it yourself? Use this frame: “[Company] provides/offers [product/service] [in region if applicable].” For example: [Example Company] provides residential solar installation and maintenance. [Example Company] offers frozen vegan meals in Scandinavia. Once you save, the platform runs initial checks and creates your first report of how your brand appears across the AI landscape.

Note: Bare in mind that when adding your first entity, it will take a while (5-10 minutes) for the dashboard to get populated, you can check the loading wheel on the top-right to see the status.

Prompts Configuration

Everything begins with the real questions your buyers ask. AI search is conversational. People don’t type “project management software”; they ask “Which tool helps a remote team to stay organized”. The platform runs your prompts across multiple models every 24 hours so you see patterns, avoiding one‑off noise.

For your first set of prompts, write a compact set that reflects awareness, consideration, and key buying decisions. Keep branded and unbranded prompts separate so you can measure pull without nudging. Include context a human would share e.g. Best project management tool with a popular chat tool integration under €15 per user? team size, industry, constraints, typical use cases etc. so models can make meaningful trade‑offs.

Some examples (to use as tone guides, not a shopping list):

• What are the leading [category] solutions for [use case]?
• Best [category] for [industry] with [constraint: budget, region, compliance]?
• Compare [YourBrand] with [Competitor] for [use case].

Workflow: Go to Prompts, add items individually, or upload a CSV file, and save. If you have gone through the wizard, all prompts are now saved and fired off to the AI models to gather their responses for analysis. You can see the progress of this first run at the top right…add screenshot.. On a daily basis, the platform will rerun all your prompts against all the models in your plan. Daily runs will reveal which brands models prefer, how they justify picks, and whether you appear as a leader, a niche option, or not at all.

Prompt Tagging & Views

Once you are satisfied with the prompts you have selected/created, you will be able to add one or more tags to each one of them so screens and charts can be filtered by tags.. Tags turn a long prompt list into clear, comparable slices of insight. Instead of scanning dozens of lines, you can flip the view to see visibility by product, market, language, or buyer persona. Start simple: give each prompt a handful of labels that describe what it targets. For instance, you might tag a prompt with the product tier it relates to (such as “starter plan” or “enterprise plan”), note the intent stage (“awareness”, “consideration”, or “decision” stage), include the persona if it matters (“cmo”, “it-admin” etc..), and capture geography or language when relevant (“eu”, “english”, “german” or “USA” etc…).

Applying tags is straightforward. When you add a prompt, type the tags into the Tags field. You can also add or edit the tags later.

Once your prompts are tagged, screens and reports become far more actionable. You can filter visibility in every section of the platform to a single product line, compare performance between English and French queries, or contrast how you appear to CMOs versus IT admins. If you notice that prompts tagged as “starter” underperform in “EU”, you can jump straight into those prompts’ citations to see what sources shaped the answers and what content or listings you may be missing.

Competitors Selection

The platform surfaces the brands that appear alongside you and ranks them with the same visibility metrics you see for your own entity. A default leaderboard shows the competing entities most cited in your model responses. If you want a tailored view you can Manage Competitors. Select brands you care about; hide others by clicking the X on a card; restore any from the Rejected tab later.

Revisit your competitor section often, in the early days of your GEO effort after every daily run, as the competitors list will keep growing as we extract new ones from responses. New entrants in any market can rise quickly once they earn citations.

Citations Analysis

The Citations tab is where strategy becomes action. Citations are the sources (website/urls) AI models use to provide recommendations. In other words, citations are critical content sources AI models rely on. The platform ranks the citations referenced in your responses by frequency, and lets you expand each source to see the exact URLs used.

Do study the article formats that get cited, learn from the headlines that pull inclusion, and the angles that shape your category narrative so you can apply those to your advantage.

How to act on what you see:

• Join the communities cited: join the conversations that match your prompts. Contribute useful answers with neutral comparisons and verifiable facts. Build credibility before mentioning your product; most spaces are tightly moderated. Take the lead if some of the community platforms that are known to be trusted sources by AI models would not be cited yet e.g. popular community platforms or Q&A sites.
• Feature on cited blogs and media (listicles, comparisons, buyer guides): contact the author or editor of cited blogs. Offer specific, verifiable additions, missing features, updated pricing, screenshots, or customer evidence. Sponsored placements can be appropriate if clearly labeled and genuinely useful.
• Add / improve your site with citation‑friendly content: You will find that vendor web pages also feature as citations. In other words your own web pages - as those of your competitors - can be used by AI models to inform their recommendations. Reverse‑outline the top cited articles. Publish a cleaner, faster page that answers the buyer question directly, documents claims with dates and sources, and includes objective comparison tables. Target long‑tail topics you can realistically rank for.
• Authority Piggybacking (guest/sponsored): Pitch a fresh article on reputable sites that usually outrank current citations in your niche. Keep structure and headline similar, add new data, and include decision criteria. The goal is to force your entity into the AI model’s citations list.

Weekly rhythm (keep it light, keep it moving):

• Review new citations and pick three moves: one on‑site page to draft/refresh, one author to pitch, one guest/community opportunity to pursue.
• Track which actions lead to new or improved citations over the next few runs.

Monthly tune‑ups:

• Expand your prompt set with 3–5 new buyer questions.
• Refresh outdated claims or screenshots.
• Adjust competitors if new brands are rising.

Metrics and What They Mean

The platform has a dedicated dashboard screen that provides summarised insights for your entity. The insight can be tailored by applying tags and model filters. Furthermore most pages include insightful metrics and filters that tell you whether you’re winning without turning this into spreadsheet theater.

Key metrics to track are:

• AI Visibility Score: Trend over 30/90 days beats day‑to‑day noise.
• Brand mentions per model: Where you’re gaining traction vs. lagging.
• Citation coverage: Count of top sources that now include your brand.
• Share of recommendation: Percentage of answers naming you.
• Win/loss vs. key competitors: Where you replace or are replaced.
• Content velocity: On‑site and off‑site pieces shipped each month.

Strategic Prompt Creation Ideas

Classic SEO leaned on search volumes. AI search doesn’t. You win by anticipating buyer intent and crafting prompts that mirror real decisions, then letting the platform test them daily for reliable patterns.

Start with your customer persona. Map their goals, constraints, and triggers. Reverse‑engineer the questions they’d ask to avoid wasting time. In AI chats, people get specific fast: they include budget limits, compliance needs, team size, regions, and “non‑negotiables” like a genuine free trial. If a free trial is rare in your niche, that becomes a high‑leverage angle both for third‑party listicles (pitch it as a USP with proof) and for your own content that targets that intent directly.

---

Core principles

• Write prompts in the buyer’s voice, not marketing prose.
• Add the constraint that actually decides the deal (price cap, data residency, SSO, SLA, language, integrations).
• Separate branded from unbranded prompts so you can see organic pull.
• Prefer concrete, testable phrasing over vague superlatives.

Persona → prompt mapping

• Role & job‑to‑be‑done: “As a RevOps manager, what’s the best [category] to consolidate [systems] without breaking [process]?”
• Company profile: “Best [category] for a 15‑person remote team in the EU with a specific compliance standard.”
• Constraints & must‑haves: “Which [category] offers a real 14‑day free trial and SSO under €50/seat?”
• Switching triggers: “If I’m moving off [Competitor], what’s a simpler alternative with [missing feature]?”

High‑leverage prompt patterns (adapt, don’t copy verbatim)

• Constraint‑led: Best [category] that supports [integration] and [compliance] under [budget].
• Feature bundle: Which tools combine [feature A] + [feature B] for [use case]?
• Operational reality: What works for a non‑technical team with [system] already in place?
• Region/legal: EU‑friendly [category] with data residency in [country] and a data agreement.
• Trial/contract terms: [Category] with true free trial and monthly billing, no annual lock‑in.
• Timeboxed outcomes: Fastest way to launch [use case] in a week with minimal setup.
• Replacement: Alternatives to [Competitor] that fix [pain] without losing [feature].

Turn USPs into prompts

If you offer something rivals lack (e.g., real free trial, transparent pricing, local support, offline mode), encode it into prompts. Then:

1. Pitch third‑party listicles with the USP highlighted and lightly evidenced (screenshots, policy pages, or user quotes).
2. Publish a matching explainer on your blog (e.g., “Best [category] tools with a real free trial”), using neutral comparisons and dated claims. This doubles as citation‑friendly content.

“50 prompts in 30 minutes” workflow

1. List 5 buyer personas × 3 stages (awareness, consideration, decision).
2. For each, note 2 must‑haves and 1 constraint (budget, region, stack, policy).
3. Combine into one natural sentence per case. Keep branded and unbranded versions.
4. Paste into the platform, tag by intent/use case, and run.

Prompt quality checklist

• Natural, first‑person or conversational phrasing.
• One clear decision criterion (two max).
• Includes relevant context (team size, region, stack).
• No fluff (avoid “best in class”, “innovative”).
• Branded and unbranded variants created.

Examples across the funnel

• Awareness: What are the top data-compliant [category] tools for a 20‑person startup in Berlin?
• Consideration: Best [category] with CRM and chat tool integrations and a real 14‑day free trial?
• Decision: Compare [YourBrand] vs [Competitor] for a non‑technical team with €1k/month budget.

What to avoid

• Stacking too many requirements into one prompt, models default to generic answers.
• Brand‑stuffing unbranded prompts, it skews visibility.
• Vague asks like “best tool” with no context, little decision signal.

With this approach, every daily run spends its limited checks on prompts that reflect real buying moments. The output is sharper: clearer brand mentions, more actionable citations, and faster insight into which angles actually move recommendations.

Once you have compiled a list of prompts that you wish to begin tracking you can manually add them to the platform, or bulk upload them via CSV file.

Closing: Ship, learn, repeat

AI search rewards brands that teach, not shout. You do not need a thousand prompts or a content factory. You need tight questions that reflect real buying moments, content that answers them with clarity, and a weekly rhythm that compounds.

Your next 7 days

• Day 1: Add your entity. Write a one‑sentence description that says exactly what you offer.
• Day 2: Draft 15 buyer‑style prompts. Tag by intent. Save branded and unbranded versions.
• Day 3: Run them. Note which brands appear and why.
• Day 4: Pull the top citations. Reverse‑outline one article and plan your answer.
• Day 5: Publish a citation‑friendly page on your site.
• Day 6: Pitch one existing author for inclusion with a factual, helpful update.
• Day 7: Join one active community thread and contribute real expertise.

Keep the loop small: measure, learn, adjust. Replace guesses with patterns. Let the platform show you where to push next.

Ready to see where you stand? Add your entity now and run your first 10 prompts. In a week, you will know more about your AI visibility than most teams learn in a quarter.