reports · graded insights
Most AI reports are black boxes: bullet points without a defense. Aegus Reports runs every insight through the Institute — multi-source weighting, per-claim confidence, citation coverage, and dissent when sources disagree. You see the reasoning, not just the conclusion.
How many independent sources support the claim, and how tightly they agree. Low agreement does not kill the claim — it gets flagged as contested terrain.
Every claim carries a 0–100% confidence score based on source quality, recency, and internal reasoning. Low-confidence claims are surfaced, not buried.
The fraction of claims in the final report that carry an explicit source citation. Uncited claims are auto-flagged for revision.
Narrow reports with 2 insights score lower than broad reports with 8. Breadth factors in against depth so we do not ship thin decks.
If the corpus contains evidence that directly contradicts the majority view, it is surfaced as dissent — not hidden. You get the fight, not just the winner.
Composite score rolls up to a single grade so you know at a glance whether the deck is ready to send or needs another source pass.
Same Institute framework that grades PATH trading signals. Every report routes through it before delivery. Expand the evidence chain on any card to see the per-source reasoning.
Report draws on 12 sources with 82% average confidence; sources align — 95% of claims carry citations.
Report draws on 6 sources with 65% average confidence; sources align — 75% of claims carry citations.
Report draws on 4 sources with 55% average confidence; sources diverge — 50% of claims carry citations.
Ships before 9 AM local every Monday. 6–9 insights, each graded. Connects Stripe, Shopify, QuickBooks, GA4, HubSpot — whatever you use. Branded PDF in your inbox.
Deeper synthesis across the quarter, including competitor moves, narrative shifts, and recommended experiments. Dissent section always included.
Reports reads from the same Institute that grades trades, calls, and content. The pattern library is shared — what the voice agent learns from client calls informs which insights the report surfaces next week. See the full diagram.
Tell us what your business actually needs visibility into. We scope the source stack (Stripe, Shopify, HubSpot, GA4, etc.), build the first weekly report, and grade every insight against the same Institute framework the rest of the stack uses.