Reports

Overview

Reports generate professional, consulting-grade AI Brand Intelligence reports from your AEO Optima data. Each report includes up to 25 data sections with SVG charts, KPI cards, competitive analysis, intelligence scores, and prioritized recommendations — designed to the visual standard of McKinsey and BCG deliverables.

Report Types

AEO Optima offers seven report types, each tailored for a different audience and purpose.

Report generation is not available on the Free plan (0 reports/month). It starts at Starter; the richer report types unlock on higher tiers.

Every report type can be generated on demand; all types except Custom — including Per-Prompt Detail and AEO Program — can also be delivered on a recurring schedule (see Scheduled Reports).

Output Formats

Each report type can be exported in multiple formats:

Report Sections

Reports contain up to 25 sections organized into categories:

Core Metrics

AI Brand Score vs. Visibility Score: A report's headline AI Brand Score is a broader composite (Visibility 35% + Sentiment 15% + BNCI 15% + CMCS 15% + ETAS 20%) than the Visibility Score shown on your dashboard, so it's usually a higher number. A dashboard Visibility of 33 and a report AI Brand Score of 45 are consistent, not a discrepancy — see Core Concepts for the full definition.

• Executive Summary — AI Brand Score, letter grade, key takeaways, competitive position
• Scorecard — Overview of all metrics in KPI cards with RAG (Red/Amber/Green) coloring
• Overview — The Overview page's own table: the last 6 months by month for the whole project, whatever dates or questions the rest of the report covers, with where each card's figures come from and what each row is compared with. In every report type except Per-Prompt. In PDF and Excel.
• Quarterly Scorecard — Each KPI from Performance → Scorecard against the target frozen when the baseline was locked, its state in words and the counts behind every rate (standard, comprehensive and AEO Program reports). Before a baseline is locked it says so in one line. The value calculator is never in a report: it uses your organisation's confidential sales figures, and a report can be read by anyone on the project or through a share link. See Scorecard.
• Key Insights — Auto-generated insight highlights for the period

Visibility & Sentiment

• Visibility Overview — Trend line, platform breakdown, confidence intervals
• Visibility Forecast — 30-day predicted trend with confidence bands
• Sentiment Analysis — Score, distribution by platform, trend over time
• Sentiment Drift — SDI tracking how sentiment narrative is evolving

Competitive Intelligence

• Mention share — your slice of all brand mentions, vs. competitors
• Competitor Ranking — Ranked table with visibility scores
• Market Entropy (MEI) — How fragmented or concentrated the market is

Citations & Content

• Citation Summary — Citation rate, source types, unique sources
• Top Citation Hubs — Most-cited domains and their influence
• Citation Gaps — Sources citing competitors but not you

Intelligence Scores

• BNCI (Brand Narrative Coherence) — How consistently AI describes your brand
• CMCS (Cross-Model Consistency) — Agreement between platforms
• MEI (Market Entropy Index) — Market concentration
• SDI (Sentiment Drift Index) — Sentiment stability
• ETAS (Entity Topical Authority) — Authority recognition

BNCI, and ETAS on a 30-day report, are scored by AI models that check your brand facts against the answers. The full report (the Reports page, a scheduled report, or an AI assistant) scores them as it builds. The quick views (the live share link, a test send of a scheduled report, the voice preview) read the scores our nightly run already made for the last 7, 30 or 90 days. If those are not ready yet, the view says the score is being retried and asks for them; a custom date range is scored only in the full report.

Goals & Intelligence

• Goal Progress — Active visibility goals with progress bars, pace status (ahead/on track/behind/at risk), milestone tracking, due dates
• Intelligence Insights — Top findings from computation engines, severity-coded, with confidence scores
• Action Efficacy — Success rates per action type, average lift, time to impact — shows which strategies work best

Advanced

• Anomaly Detection — Unusual spikes or drops in metrics
• Shopping Visibility — Product mentions, pricing accuracy
• GEO Audit — AI readiness score by category

Integrations

• GA4 Traffic — Google Analytics traffic correlation
• GSC Search — Google Search Console data
• AI Visibility ↔ Real Traffic (Fusion) — joins AI citations, AI-referral traffic, and Google rank on your own page URLs: a per-page fusion table + insights (cited-page→traffic, AI-vs-SEO divergence, AI-Overview share, first-party Google AI visibility, …). Columns adapt to your connected sources; the section is omitted when there's no fusion signal.
• Revenue Attribution — AI-driven revenue correlation

Actions

• Recommendations — Prioritized, actionable optimization tasks with expected impact

Reference

• Glossary — Definitions of all metrics used in the report
• Methodology — How scores are calculated, sample sizes, statistical methods

Date Ranges

Reports can cover:

Voice Register

Every report is written in one of three voice registers, controlling tone and reading style. Choose at generation time, or set a project default in Settings > Project.

Voice applies across all narrative sections (Period in Review, executive summary commentary, recommendations) but does not change the data — same metrics, same charts, different framing. When a report is shared via a live share link, the recipient sees the same voice the original was generated with.

How to Generate a Report

1. Navigate to Reports from the sidebar.
2. Select your report type (Executive, Standard, Comprehensive, Competitive, or Custom).
3. Choose output formats — select one or more (PDF, Slides, Excel, CSV).
4. Set the date range (7, 30, 90 days, or custom).
5. Pick a voice register (or use the project default).
6. For Custom reports, select which sections to include.
7. Click Generate Report.
8. Once generated, download from the Recent Reports list or the Report History page.

Question Visibility Sheet

A direct download, not a generated report: every active question, cycle by cycle and engine by engine, as Excel or CSV. Take it from the Question visibility sheet card under the Reports library, or from the Download menu on the Answers page (which uses the window, or the cycle, the page is showing). Any role that can read the answers can take it; nothing is stored in the library.

• Rows are your tracked questions, with their type (Branded, Non-Branded or Competitor).
• Columns run cycle by cycle (C1, C2, … counted from the project's first capture, at its schedule's rhythm), then engine by engine — every engine switched on for the project, plus any other that answered in the window, so an engine that gave no answer still has its column — then three cells:
  • Visible? — whether the answer named the brand. When an engine gave more than one answer to the question in that cycle, "(k of n)" says how many named it. "No answer" means nothing was captured, which is not the same as "No".
  • With Link — whether an answer that named the brand linked to a web page. A source the engine mentioned without a link does not count. "N.A" when the brand was not named.
  • Note — whose pages those links were: Self (your own site), 3rd party, Self and 3rd party, and Social media with the platforms named (for example, "Social media (LinkedIn, YouTube)"). Review sites count as 3rd party.
• Whole cycles. The sheet covers each whole cycle that overlaps the dates you pick, headed with its real first and last day. A project without enough captures to have a rhythm gets one column group for the dates you picked, with no cycle number.
• Limits. A sheet holds up to 31 cycles (a month of a project that captures daily); pick a shorter window for more. An answer that came back blank is left out.

The sheet's first row repeats this key in one line, so it reads without this page.

Report Sharing

Share reports with stakeholders via secure links — no account required for recipients.

Creating a Share Link

1. Find the report in your Recent Reports list or Report History.
2. Click the Share icon.
3. Configure access controls:
   • Password protection — Require a password to view
   • Link expiry — Auto-expire after 7, 30, or 90 days (or never)
   • Allow comments — Let viewers leave section-level feedback
   • Allow download — Control whether viewers can download copies
4. Click Generate Share Link.
5. Copy and send the link to your recipients.

Live Report Viewer

Recipients open the share link in their browser and see an interactive report with:

• Sticky sidebar navigation — Jump between sections
• Interactive charts — Hover for data points
• Searchable tables — Filter large datasets
• Metric definitions — Expandable "What does this mean?" panels
• Comment threads — Section-level discussion (if enabled)
• Download options — PDF and Excel downloads (if enabled)

Share Limits by Plan

Scheduled Reports

Automate report generation and delivery on a recurring basis.

1. Go to Reports → Scheduled Reports.
2. Click Create Schedule.
3. Choose frequency (weekly, monthly or quarterly), report type, and recipients.
4. Pick the time of day and timezone the report should arrive. The schedule fires at that local hour, adjusting automatically for daylight saving time.
5. Optionally add delivery channels (see below) to post the report beyond email.
6. The system generates and delivers the report on schedule.

Scheduled reports are available on Pro-Individual and above. Every plan with the feature can choose a weekly, monthly or quarterly cadence; the number of active schedules is limited per plan (see Plans).

A weekly report covers the last 7 days and a monthly one the last 30. A quarterly report is sent on the day you pick in January, April, July and October and covers the previous calendar quarter: the January report covers 1 October to 31 December.

Send time & timezone

Each schedule has its own timezone and hour-of-day. The next-send time is computed as that local wall-clock time and stored as a precise UTC instant, so the report always arrives at — for example — 9:00 AM in Asia/Kolkata regardless of where the servers run or whether daylight saving is in effect. Existing schedules created before this feature continue to deliver at 09:00 UTC until you change them.

Delivery channels

Beyond email, a schedule can fan the same report summary out to:

• Slack — paste an incoming-webhook URL from your Slack app. We post a Block Kit message with the AI Brand Score, snapshot count, and a button linking to the full report.
• Webhook — any HTTPS endpoint. We POST a signed JSON summary (report.delivered) with an X-AEO-Signature HMAC header when you provide a signing secret. Point it at Zapier, Make, or n8n to forward into Notion, Google Drive, a database, or anywhere else.

Channel secrets (the Slack URL, the signing secret) are encrypted at rest and never shown again after you save them — the UI displays only a masked hint. Channel delivery is best-effort: if Slack or your endpoint is down, email still goes out and the failure is recorded for that run.

Quota behavior

Each scheduled delivery counts toward your plan's monthly report-generation quota — the same pool used by manually generated reports and the MCP generate_report tool.

• Quota exhausted — if a schedule fires while your organization has used its monthly report generations, that occurrence is skipped and the schedule's creator receives an email notice. The schedule stays active and resumes automatically when the monthly quota resets.
• Plan no longer includes the feature — after a downgrade, the schedule is paused and the creator is notified. Re-enable it from the Reports page once your plan includes scheduled report delivery again.

Notices go only to the schedule's creator, never to the recipient list.

When a scheduled report cannot be generated

If a report cannot be generated when its schedule fires (see The report itself, below), nothing is sent — no recipient receives a copy without its lead. The attempt is recorded as failed in the report library and the schedule tries again at the next hourly run: at most six more times, and never past the schedule's next delivery. If the failed attempt could not be recorded, or its attempts could not be counted, the period is given up at once instead of retried. If it still cannot be generated, that period's report is given up — the schedule's creator receives one email saying so, and the next delivery runs on its usual date.

When part of a report cannot be loaded

A report never presents a missing figure as an empty one, and each state gives its own reason:

• The report itself. The four question types lead every report. The project's settings (including its timezone, which sets the report's dates) and the answers nearly every section is built on come with them. If any of these cannot be read, no report is produced. Running it yourself shows "We couldn't generate this report just now. Try again in a few minutes", and the attempt appears in the library as failed. Nothing is wrong with your data — trying again usually works.
• The AI Brand Score. It is never shown when any of the measures it blends — visibility or one of the intelligence engines — could not be loaded: a score put together from only part of its inputs would be a different number under the same name. It reads "Not available" on the PDF, in the email, in the exports and in the Slack summary, and the Visibility section is printed as not available when its own read failed. When the cause is an intelligence engine whose figures failed when they were prepared, the note reads "one engine's data couldn't be loaded; retrying": those figures are prepared again in the background within about an hour. The API response, the report.generated and report.delivered webhooks and the MCP generate_report tool return the score as empty with the reason alongside: scoreNotLoaded (API and webhooks) or score_not_loaded (MCP).
• A section. Any other section that cannot be read is still printed, under its own title: "Not available — this part could not be loaded when the report was generated". That includes Prompt Coverage, which is never shown as "0 prompts", and the Competitive Landscape: when competitor mentions cannot be read, the competitive position also reads "Not available" (never a 0% share), the takeaways leave out their competitor line, and the rest of the report is still sent. A section is never left out and never shown as zero. The CSV and Excel exports and the shared live view list those parts by name, and the MCP tool names them in sections_not_loaded.
• The capture coverage and the period in cycles. If the capture coverage cannot be read, the cover says so rather than showing a window with no gaps. If the report's period in capture cycles cannot be read, the report library says "Couldn't load the cycle period" — different from "Period not recorded in cycles", which is shown only for reports made before periods were recorded.
• Changes against the previous period. If the previous period cannot be read, every change in the report reads "Change not available" — never "no change", and never a gain from zero. The current figures, the AI Brand Score included, are still printed.

Custom Branding

Enterprise-Brand, Enterprise-Agency, and Custom plans (any tier with white_label: true) can customize report appearance:

• Company logo — Displayed on the cover page
• Brand color — Applied to headers, charts, and accent elements
• Company name — Replaces "AEO Optima" branding
• White-label mode — Removes all AEO Optima references

Report Quality Assurance

Every generated report passes through 20 automated QA checks. A failure at error severity blocks delivery (the report is marked failed and a platform alert fires) rather than shipping a misleading report:

Generation Limits

API & MCP Access

Reports can be generated programmatically via the REST API or MCP tools:

• POST /api/reports/generate — Generate a report
• GET /api/reports/history — List past reports
• POST /api/reports/share — Create a share link
• MCP tools (one-off reports + shares): generate_report, get_report_history, get_report_download, create_report_share, list_report_shares, revoke_report_share
• MCP tools (recurring schedules): list_scheduled_reports, create_scheduled_report, update_scheduled_report, delete_scheduled_report — manage timezone-aware recurring email reports. Slack/webhook delivery channels are configured in the dashboard, not via MCP.

Best Practices

• Use Executive reports for weekly updates — Quick, focused, and the lightest report to generate.
• Use Comprehensive for quarterly reviews — Complete picture with all intelligence scores.
• Use Competitive for strategy meetings — Focus on market position and competitor moves.
• Use Custom for specific teams — SEO team gets citations, strategy team gets competitive, leadership gets executive.
• Schedule monthly reports — Automate delivery so stakeholders always have fresh data.
• Share with password protection — Add a layer of security for sensitive competitive intelligence.

Tip: Generate a Comprehensive report at the end of each month and share it with your team. Use the interactive live viewer for discussion and the PDF for archival.