MCP Server Use Cases & Workflows
Real-world workflow examples for the Aurora MCP server — from weekly reports to competitive analysis.
These examples show complete workflows you can run by talking to your AI assistant with the Aurora MCP server connected. Each one describes the scenario, what to ask, which tools run behind the scenes, and what you get back.
1. Monitor Brand Visibility Trends
Scenario: You want a quick week-over-week comparison to see if your brand is gaining or losing ground in AI responses.
Prompt:
Compare my brand's visibility this week vs last week across all AI providers.
Behind the scenes: The assistant calls list_campaigns to find your campaign, then get_visibility_trend with a 14-day window — the daily series includes answers-per-day, so it can compare the two weeks in one call.
Example response:
Here's your week-over-week comparison for My Brand US:
Provider Last Week This Week Change ChatGPT 65 72 +7 Gemini 54 58 +4 Perplexity 80 85 +5 Your visibility is up across all providers. The biggest jump is on ChatGPT (+7 points), likely driven by increased citations from your recent blog posts.
2. Find Most-Cited Competitor Domains
Scenario: You want to know which competitor websites AI models are recommending to users who search for queries you track.
Prompt:
Which competitor domains get cited most often for my tracked queries? Exclude my own domain.
Behind the scenes: The assistant calls get_citation_share (or get_sources at domain level) — the summary already ranks domains by citation share, so it just filters out your own domain.
Example response:
Top competitor domains cited in AI responses (last 30 days):
- competitor-a.com — 67 citations across 12 queries
- competitor-b.com — 38 citations across 8 queries
- industry-blog.com — 29 citations across 15 queries
- reddit.com/r/yourspace — 22 citations across 6 queries
competitor-a.com dominates on ChatGPT (45 of their 67 citations), while industry-blog.com is spread more evenly across providers. Reddit citations appear mostly on Perplexity.
3. Compare ChatGPT vs Gemini for Your Brand
Scenario: Your marketing team wants to understand how different AI platforms treat your brand — do they mention you equally, or is one platform much better for you?
Prompt:
Compare how ChatGPT and Gemini handle my brand — mentions, citations, and visibility scores for the past month.
Behind the scenes: The assistant calls get_brand_visibility with a 30-day range, then filters for chatgpt and gemini providers and compares the metrics.
Example response:
ChatGPT vs Gemini for My Brand US (last 30 days):
Metric ChatGPT Gemini Winner Total Mentions 142 98 ChatGPT Total Citations 56 71 Gemini Avg Visibility Score 72 64 ChatGPT ChatGPT mentions you more often and gives you a higher visibility score overall. But Gemini actually cites your website more frequently — it links to your pages even when it mentions you less in the text. If citation-driven traffic matters most to you, Gemini is your stronger channel.
4. Generate a Weekly Summary for Marketing
Scenario: Every Monday you send a quick brand-in-AI update to your team. You want the AI assistant to draft it for you.
Prompt:
Write a short weekly GEO summary for my marketing team — brand visibility, top competitors, and any notable changes from last week.
Behind the scenes: The assistant calls get_brand_visibility (this week + last week), get_citations, and get_keyword_counts to compile a full picture, then formats it as a human-readable summary.
Example response:
Weekly GEO Report — Mar 3–9, 2026
Visibility: Up 6% overall. Perplexity remains our strongest platform (85 avg score), with ChatGPT climbing to 72 (+7 from last week).
Citations: Our domain was cited 89 times this week (up from 76). The new pricing guide drove 23 new citations on its own.
Competitors: competitor-a.com gained 12 citations this week, mostly on ChatGPT. They published a comparison article that's getting picked up.
Keywords: "affordable" appeared 34 times in AI responses about our space — consider whether we want to lean into or away from this positioning.
Action items: Monitor competitor-a.com's comparison content. Consider publishing a response piece.
5. Analyze Which URLs Cite Your Brand Most
Scenario: You want to know exactly which pages on your site (or competitor sites) are being picked up by AI models, so you can optimize or replicate high-performing content.
Prompt:
Show me the individual URLs that get cited most by AI providers. Group by my own URLs vs competitor URLs.
Behind the scenes: The assistant calls get_sources with level: "url" — results come back already grouped per URL with citation counts, providers, and best position — and splits them into own-domain and competitor categories.
Example response:
Your most-cited pages:
URL Citations Top Provider example.com/pricing 23 ChatGPT example.com/blog/guide 18 Perplexity example.com/features 14 Gemini Top competitor pages:
URL Citations Top Provider competitor-a.com/comparison 19 ChatGPT competitor-b.com/reviews 12 Perplexity reddit.com/r/yourspace/best-tools 9 Perplexity Your pricing page is your strongest asset — it's the most-cited page across all domains. competitor-a.com's comparison article is worth reading; ChatGPT picks it up for nearly every comparison query you track.
6. Check Your Brand on Yandex (Russian Market)
Scenario: You're expanding into the Russian market and want to know if Yandex's AI search (Neuro) mentions your brand when users ask relevant questions.
Prompt:
How does my brand appear on Yandex compared to other AI providers? Show me Yandex-specific visibility and citations.
Behind the scenes: The assistant calls get_brand_visibility and get_citations, then filters for the yandex provider to give you a focused view of your Russian-market AI presence.
Example response:
Yandex Neuro results for My Brand (last 30 days):
Metric Yandex All Providers Avg Mentions 8 42 Citations 3 24 Visibility Score 35 68 Your Yandex visibility is lower than average. Note that Yandex Neuro only triggers its AI answer for about 60% of queries — for the other 40%, users see standard search results instead. Of the queries that did trigger a Neuro answer, your brand appeared in 3 of them with direct citations to your site.
Top cited domains on Yandex:
- competitor-a.ru — 12 citations
- habr.com — 8 citations
- example.com (you) — 3 citations
7. Full Provider Comparison
Scenario: You want a bird's-eye view of how your brand performs across all 7 tracked AI platforms to decide where to focus your GEO optimization efforts.
Prompt:
Give me a complete comparison of my brand's visibility across all AI providers for the past month. Rank them from best to worst.
Behind the scenes: The assistant calls get_brand_visibility with a 30-day range, groups by provider, and ranks them.
Example response:
Brand visibility across all providers (last 30 days):
Rank Provider Avg Score Mentions Citations 1 Perplexity 85 52 34 2 ChatGPT 72 45 18 3 Google AI Mode 68 38 22 4 Copilot 65 35 15 5 Gemini 58 31 12 6 AI Overview 45 22 8 7 Yandex 35 8 3 Perplexity is your strongest platform by far — it cites your pages more than any other provider. Google AI Mode and Copilot are solid middle-ground performers. Yandex is the weakest but that's expected if Russian-market content isn't a priority yet.
8. Share of Voice & Competitor Deep-Dive
Scenario: A quarterly review is coming up and you need to know exactly where you stand against competitors — not just overall, but per AI engine and per prompt.
Prompt:
What's our share of voice vs competitors, per AI engine? Which prompts are we losing, and to whom?
Behind the scenes: The assistant calls get_share_of_voice (overall + per-engine breakdown) and get_competitor_report (per-competitor profiles and the prompt × engine matrix).
Example response:
Share of voice: You hold 21% of brand mentions across your tracked prompts; competitor-a leads with 34%.
Engine Your SoV Leader Google AI Mode 32% you Perplexity 24% competitor-a (38%) ChatGPT 14% competitor-a (41%) You're losing "best X for teams" and "X pricing comparison" prompts to competitor-a on ChatGPT — their comparison article is cited in 8 of 10 answers. Your strongest untapped lever is ChatGPT, where your citation rate is half your mention rate.
9. Read What the AI Actually Said
Scenario: Your visibility dropped on one prompt and you want to see the actual answers to understand how you're being described.
Prompt:
Show me the last 5 Perplexity answers for my "best CRM for startups" prompt and summarize how my brand is positioned.
Behind the scenes: The assistant calls list_prompts to find the prompt ID, then get_answers filtered to provider: "perplexity", and reads the raw answer texts.
Example response:
Across the 5 most recent Perplexity answers, your brand appears in 3. You're consistently described as the "budget-friendly option" and recommended for teams under 10. Competitor-a is framed as the "full-featured" choice. Two answers cite your pricing page; none cite your feature pages — which matches your positioning problem: Perplexity doesn't know about your enterprise tier.
10. Work Through Your Insights Report
Scenario: Aurora generated an insights report and you want to triage its recommendations conversationally.
Prompt:
Get my latest insights report. Walk me through the top recommended actions and accept the ones about content briefs.
Behind the scenes: The assistant calls list_insights_reports, then get_insights_report for the executive summary and actions. With write access, it can call update_insights_action to accept or dismiss each one as you decide.
Example response:
Your report from Aug 2 recommends 7 actions. The top 3 by priority:
- Publish a comparison page vs competitor-a (high impact, medium effort) — they win 60% of comparison prompts
- Refresh the pricing guide (high impact, low effort) — your most-cited page, but 8 months old
- Pitch industry-blog.com (medium impact, low effort) — cited in 15% of answers, never mentions you
I've accepted the two content-brief actions as requested. ✅
11. Set Up a New Campaign
Scenario: You're onboarding a new brand and would rather describe it than fill in a form.
Prompt:
Set up tracking for Ofri, a Swiss handyman marketplace at ofri.ch. Competitors are renovero and hausinfo. Track the German-language questions people ask when looking for an electrician or plumber in Zurich, Lucerne and St. Gallen.
Behind the scenes: With the campaign-creation grant, the assistant calls create_campaign — first without confirm, which returns the exact brand, competitor, keyword, prompt and engine list it would create along with your remaining plan headroom, and writes nothing. You review it, then it calls again with confirm: true.
Example response:
Here's what I'd create — nothing is saved yet:
- Campaign: Ofri (Switzerland, German)
- Own brand: Ofri — ofri.ch
- Competitors: renovero, hausinfo
- Prompts: 3 — "Wie finde ich einen lokalen Elektriker in Zürich?", "Wie kann ich einen zuverlässigen Klempner in Luzern finden?", "Wo finde ich Bewertungen für Handwerker in der Region St. Gallen?"
- Engines: ChatGPT, Perplexity, Gemini, Google AI Mode, Google AI Overview
- Plan: 1 of 10 campaigns used
Note I kept the brand name out of the prompts — including it guarantees a mention and inflates share of voice. Say the word and I'll create it.
After confirmation: ✅ Created. Collection starts on the next scheduled run (09:00 UTC daily) — there won't be data or a report until then.
Built-in Prompts
Instead of writing your own workflow, run one of the server's built-in prompts (your MCP client lists them alongside the tools):
weekly_visibility_pulse— a week-over-week narrative: own-brand trend, citation-share movers, and one recommended focus for next week.engine_scorecard— a per-engine table (ChatGPT, Perplexity, Gemini, …) of share of voice and where improvement is most valuable.competitor_movers— which competitors gained or lost visibility in the last 30 days, which prompts they win, and concrete counter-moves.
Each prompt takes a campaign argument (name or ID) and orchestrates the right tools automatically.