How to Work with ARPAI GEO

On the Dashboard, click New Project. Enter your company website URL and click Create Project.

The system scrapes your website and automatically fills your Company Profile: brand name, industry, products, services, target audience, and USP. Review and edit if needed — click EDIT in the Company Profile section.

Now add your first monitoring location. Click Add Location, select a country, region, or city, and choose the query language. Each location runs independently — add as many as you need.

Click your location tab, then click Location Monitoring (or use Run All on the All Locations view to start all locations at once). The system generates 15 search queries — 10 non-branded + 5 branded — across 8 buying-intent categories and sends them to all AI engines simultaneously.

When finished, you see the SOV Scorecard — three key metrics:

Below the scorecard, each AI engine card shows per-engine visibility, citation rate, sentiment (positive / neutral / negative), and position distribution.

Scroll down to Queries & Mentions. Every query is color-coded:

Click any query row to open the Brand Ranking modal — see the exact position of your brand and all competitors for that query in each engine. If you have history, a position dynamics chart shows trends over time.

Further down the page: Top 10 Competitors (who appears most often across all queries), Top 20 Resources (which websites AI cites as sources), and Competitor Analysis (brand vs. query relevance matrix).

Go to the Promotion tab, then the Analysis & Topics sub-tab. Click Analyze.

The system analyzes all sources that AI engines cited in your monitoring results and produces:

• Content Publishing Platforms — websites and blogs where you should publish articles. Each card shows priority (High / Medium / Low), the platform name, and recommended content types.
• Review Platforms — directories and review sites where AI looks for ratings and opinions about your industry. Colored by priority.
• Marketplace Platforms — if applicable, product marketplaces that AI cites in its answers.
• Missing Platforms (amber cards, marked NEEDED) — platforms where competitors are present but you are not. These are your biggest growth opportunities.
• Blind Spots (amber tags in the Source Analysis results) — specific queries where none of the AI engines mention you at all. These are real customer questions that AI answers without ever bringing up your brand.
• Competitor Content Map — what types of content your competitors publish, and on which platforms.

Switch to the Strategy sub-tab (inside Promotion). Click Update Strategies.

The AI agent reads your monitoring results, source analysis, competitors, and real-time market intelligence. It generates a 3-phase roadmap:

• Quick Wins — actions you can execute this week with immediate impact
• Medium-Term (1–3 months) — content campaigns and platform presence building
• Long-Term (3–6 months) — authority and entity optimization

Each phase opens with an executive summary. Inside, every recommendation follows this format:

• Fact — the data point from your monitoring that triggered this recommendation
• Action — concrete steps you can execute
• Expected Result — which metric should improve and by how much
• Verification timeline — when to run the next monitoring to check

The strategy also includes an engine-specific playbook — tailored tactics for each AI engine based on how it ranks brands.

In your location monitoring results, scroll to the Recommendations section. Each recommendation uses the Fact → Action → Expected Result format:

• Fact — the data point from your monitoring (e.g., “Brand not mentioned in 4 of 10 non-branded queries”)
• Action — a concrete step with clear instructions
• Expected Result — which metric should improve and by how much

Each recommendation has an effort badge — Quick Win, Medium Effort, or Major Initiative — so you can prioritize. The system also tracks progress across monitoring cycles: recurring issues are flagged, and resolved issues are marked as such.

Click Add to Plan to convert any recommendation into a testable hypothesis. The hypothesis captures baseline metrics, expected change, implementation checklist, and a verification date (default 21 days). After your next monitoring run, the system auto-checks every hypothesis and gives a statistical verdict: Confirmed, Refuted, or Too Early.

Run monitoring regularly — once or twice a week — to track changes as they happen. Click Location Monitoring on your location page.

The system runs the same queries across all AI engines and compares results with your previous scan. At the top of results, the Key Changes banner highlights what moved:

The system automatically checks your hypotheses against new data. Each hypothesis receives one of four verdicts:

Use the Dynamics chart to see trends over 7 days, 30 days, 90 days, or 1 year. Toggle between Visibility and Citations. Each AI engine gets its own color-coded line so you can spot which engines respond to your changes first.

Product monitoring: In the Products / Services section, add individual products. Each gets its own sub-monitor with dedicated SOV metrics and competitor analysis. Useful for e-commerce, SaaS with multiple pricing tiers, or agencies managing a product portfolio.

Multiple locations: Add locations for every market you operate in. Each location runs its own queries in the local language with region-specific competitors. A brand that scores 60% SOV in New York might score 10% in London — each market has its own competitive landscape.

Scheduled monitoring: Set up automatic monitoring schedules so the system runs on its own. Focus on strategy execution while data collection happens in the background.

Custom queries: In the Custom Queries section, add your own search queries. They run alongside auto-generated ones on every monitoring scan. Use them to track specific product names, competitor comparisons, or niche topics.

In the Monitoring tab, find the AI Referral Analytics Code section. Copy the JavaScript snippet and paste it before </body> on your website.

The pixel automatically detects visitors coming from AI engines (ChatGPT, Perplexity, Gemini, and others) using first-party cookies and referrer/UTM signals. It is not a cross-site tracker. Once installed, you see:

• AI Traffic dashboard — total visits, daily chart, breakdown by engine, device type, countries, languages, and top landing pages
• Conversions & ROI — if you add window.arpaiConversion('purchase', 99.99) to your conversion events, the system tracks: conversion rate, average days to convert, revenue by AI engine, and ROI calculation
• Pixel status indicator — green "Pixel active", yellow "Waiting for AI traffic", red "No recent data"

The pixel uses first-party cookies (_arpai_vid, _arpai_src) and referrer/UTM detection. It is not a cross-site advertising tracker and does not share data with third parties. As the website owner deploying the pixel, you are responsible for any cookie consent required in your region. Learn more about the AI Traffic Pixel →