Every feature, honestly labeled

One scan asks Perplexity, ChatGPT, Gemini, Claude and Google AI Overviews the questions your buyers ask.

Every prompt runs three times per engine; results are rates with 95% confidence intervals, not one-off anecdotes.

Being named in an answer and being linked as a source are different problems — we never blend them into one score.

Who gets named instead of you, per engine, with domains deduped so www and non-www are one rival, not two.

Track the prompts that matter, each recording where it came from — onboarding, manual, or discovered.

Tracked prompts re-run automatically on their cadence, so trend lines keep moving without anyone clicking scan.

Every run is kept — open any past scan and see the full answers, scores and citations from that moment.

A fixed prompt panel sampled repeatedly to measure how volatile each engine is — the error bars behind our error bars.

One trending 0–100 headline per client, computed from the mentions, citations, share-of-voice and sentiment we already measure.

The same prompt asked in different languages, locales and buyer personas, tracked separately.

The AI answer block at the top of Google — captured with its references and the organic results beside it.

Modeled and wired, but off by default: the upstream SERP provider cannot yet return AI Mode content reliably.

Not just whether engines mention you — how they talk about you, classified per answer.

The Reddit threads, YouTube videos and listicles engines lean on for your prompts — and whether you appear in them.

A ranked to-do list of the specific pages engines cite for your prompts, and what to do about each.

A point-in-time check of which AI crawlers can actually reach a site — a fact, not an estimate.

Whether assistants recommend your products when buyers ask what to buy.

Engines six and seven, keeping the sampled methodology rather than racing the coverage number.

An ongoing feed of which AI bots visit which client pages — the recurring proof-of-work every competitor now ships.

Store a client's ground-truth facts and flag sampled answers that contradict them.

A streaming agent over your own scan data — ask why visibility moved and what to do next.

Adding a site seeds buyer prompts, scans them 3× across five engines, and runs a technical audit — while you watch.

A daily sweep across every client that flags visibility drops and competitor surges before the client notices.

Monthly client reports drafted as a grounded narrative from real scan data.

Scan data and missing-schema findings turned into a prioritized what-to-do list per site.

The alerts and report drafts we already generate, delivered where the team actually lives.

Each client is a first-class workspace: its own sites, prompts, reports, portal and billing records.

Owner, manager, analyst, client-viewer and partner roles — with members scoping down to a single client.

Agency partners get their own earnings dashboard with a share percentage set at creation.

Every billable action lands in an append-only ledger at the rate in force that day.

Platform admins create agencies atomically — org, owner, entitlement and default client — with rollback if any step fails.

Password resets, role changes, suspensions and limit edits — recorded with who and when, uneditable.

Admins can see what an agency sees to answer support questions — deliberately read-only, never act-as.

Public signup is closed; accounts come from provisioning or a token-gated internal page.

Password and email changes verified against the current password, plus an email reset flow.

Live, always-current portals your client opens from a link — your brand, never ours.

Portals served from agency.aeranking.com — or from the agency's own domain with automated DNS verification.

Colors, font presets and footer text — from a safe preset system that never loads external resources.

Periodic human-readable reports from real scan data: headline, movement, what was done.

Run a shareable AEO audit on a prospect before they sign — the pitch deck writes itself from live data.

Issue branded invoices per client from share links; payment itself moves outside the platform for now.

Every action priced from a versioned rate card; the version is stamped onto each ledger row.

Atomic credit reservation with per-agency ceilings set at onboarding.

Record each client's retainer and cost; see margin per client at a glance.

Revenue, cost and margin per agency, with a cost-reconciliation check — platform admin only.

Partners see their share rate, basis and accrued earnings — and nothing else.

Fully built, deliberately hidden: the pilot is invite-only and every payment surface is gated off.

Hashed, prefix-identified tokens scoped to an agency or a single client, with rate limits.

Pull visibility metrics, prompts and run data into your own stack.

Query your visibility data from Claude, Cursor, n8n or any MCP-capable client.

Blog posts publish through a CMS webhook and are pushed to search engines the moment they land.

The reporting stack agencies actually assemble in, fed directly.

Connect visibility to outcomes: AI-referred sessions and conversions beside the share-of-voice they came from.

Push Optimization Studio articles straight into the client's CMS instead of copy-paste.

Every scan identifies the prompts where competitors are cited and you are not.

Research, draft and polish agents turn a gap into a publishable, source-cited article.

Schema, crawler access and answerability checks with concrete fixes and generated markup.

New prompt ideas mined from what Google users actually ask, one click to track.

Tie each published article to the visibility change that followed, per prompt and engine.

Visibility checker, crawler access, answerability scorer, schema generator and five more — no account needed.

Open data on which domains AI engines cite most, from our own scan pipeline.

Direct, sourced answers to the field's questions — and comparisons honest enough to name who should not buy from us.

A public directory of agencies on the platform, with leads routed straight to the agency.

Invite-only access requests carry the qualification the operator needs to size a workspace.

This page — every feature with its honest status, findable location, and what is planned next.

The platform practices what it sells: llms.txt, IndexNow, entity graph, FAQ schema, canonical discipline.

The page and its standards exist; it publishes nothing until a real measured engagement exists.