The Aelo Loop · Five welded steps
Scan → Diagnose → Prescribe → Act → Prove.
Every other AEO tool stops at Scan. Aelo carries the loop all the way to the receipt, the number that says whether what you shipped moved the answer.
Scan
Every AI, every prompt that matters, on a schedule.
The scanner runs your target prompts across ChatGPT (GPT-4o), Gemini 2.5, Claude 3.5 Sonnet, Perplexity's Llama 3.1 Sonar, and Google's AI Overview. Every run is analyzed by an in-loop model that extracts the structured signals, mentioned, position, sentiment, competitors, citations. Nothing is fabricated. When a provider fails, we surface an honest empty state.
- Multi-LLM parallel scans (5 providers today; more on request)
- Analyzer-in-loop: sentiment score, reasons, brand aliases
- Weekly (Radar) or daily (Command) cadence, Vercel Cron
- Schedules survive limit hits; overage is prevented, not billed
Diagnose
The exact URLs the AI leans on when it answers.
For every mention (and every miss), Aelo attributes the citation graph, the specific Reddit threads, Wikipedia articles, G2 pages, and technical documentation the LLM pulled from. We compare your citation footprint to competitors and highlight Citation Gaps, the pages that produce answers but never mention you.
- Per-scan citation extraction with own-domain flagging
- Citation Gap analysis vs. every tracked competitor
- Source classification (Reddit, docs, Tier-1 media, own domain)
- Own-domain-cited alerts when an LLM starts using your site
Prescribe
Actions ranked by expected visibility movement.
Aelo turns diagnosis into an ordered work queue: draft this forum reply, publish this comparison page, add this schema block, close this citation gap. Every item is scoped, actionable, and tied to the specific prompt it will affect. Content briefs and Reddit-safe replies are pre-drafted by an in-loop model.
- Forum reply drafts (non-spammy, community-aware)
- Content briefs targeting the exact intent queries
- Schema.org JSON-LD snippets for machine readability
- Prioritized by baseline gap × prompt intent volume
Act
You ship. Aelo snapshots the baseline the moment you do.
Mark a forum reply as posted, or hit publish from Content Studio, Aelo captures the frozen visibility snapshot at that moment. That baseline is denormalized into the intervention record, so even if raw scans are pruned later, the 'before' number survives. This is what makes the receipt trustworthy.
- One-click 'mark as posted' from Forum Hub
- Content Studio publish → auto-intervention
- Baseline snapshot per (prompt × platform), frozen in JSONB
- Every action is a row you can audit later
Prove
Re-scan. Compare. Deliver the receipt.
Hit Measure and Aelo re-scans every target prompt on every configured provider, computes the delta vs. baseline, and writes the receipt, visibility change in points, position change, a Verdict (improved / no_change / regressed), and the timestamp. Nothing is smoothed. If the follow-up regressed, it says regressed.
- Per-intervention before/after receipts (Verdict + delta)
- Weekly digest email rolls receipts into one summary
- Public monthly India AI Visibility Index built from anonymized deltas
- Every number is a query away from the raw scan
Under the hood
Serverless, multi-tenant, real-time. Trust every row.
See the loop, running.
Your first scan finishes in under a minute. Interventions unlock on Command.