One Analytics Engine for Your Entire Platform.
Stop pulling fragmented reports from six different dashboards. Aeion BI is a visual report builder that queries across Commerce, CRM, Finance, and 30+ other modules natively, turning siloed data into unified business intelligence without the need for expensive ETL pipelines.
Break Down Data Silos
Visual Report Builder
Query any collection without writing SQL. Select fields, apply groupings, aggregate data (sum/avg/count), and visualize results in 14 distinct chart types natively.
43 Pre-Built Metric Packs
Get instant value out-of-the-box with curated fields, suggested aggregations, and recommended charts for major modules like Commerce (AOV), CRM (Pipeline), and Finance (P&L).
Cross-Module Queries
View the full picture. See revenue by customer segment, support cost per deal, or fulfillment time by region by combining data seamlessly through the service layer.
Scheduled PDF Delivery
Automate your board reports. Schedule daily, weekly, or monthly reports delivered directly to your inbox or Slack channel as formatted PDFs or Excel workbooks.
Threshold Alerting
Monitor crucial metrics and set alerts for percentage drops or hard thresholds. Know when revenue dips or ticket volume spikes before it becomes a crisis.
Dashboard Composer
Drag and drop saved reports onto a 12-column grid. Build custom, auto-refreshing command centers and share them via the self-service portal.
Supercharged by Aeion AI & WebGPU
AI Natural Language Reports
Just ask. "Show me revenue by product category for the last 30 days." The AI instantly selects the optimal data source, applies aggregations, and renders the perfect chart for your review.
WebGPU Compute Cluster
We compile aggregation logic into WebGPU Shading Language (WGSL), pushing heavy data processing directly to the browser's GPU. What takes minutes on a traditional server runs locally in milliseconds at teraflop speeds.
What's in the Box — BI Capabilities by Area
43 Metric Packs
Pre-built curated fields + aggregations + charts per module
Report Builder
Visual report compiler — pick a collection, choose fields, aggregate, done
21 Report Templates
Pre-mapped to common business questions
Alerting Engine
Threshold + percentage-drop monitoring with notification fan-out
Export
PDF + Excel + CSV + Parquet export with proper formatting
Scheduled Delivery
Daily / weekly / monthly cron with timezone awareness
Formula Evaluator
Safe calculated-field engine for custom metrics — no code required
Goal Tracking
Progress + variance + forecasting + ETA
Data Source Service
Multi-source query coordination (cross-module joins)
AI Explain
"Why is revenue down 12%?" → AI explains anomalies from the underlying data
AI Report Builder
"Show me revenue by category" → AI builds the report from natural language
The 43 Pre-Built Metric Packs
Commerce
AOV, conversion rate, cart-abandonment, repeat-customer rate, top-products, refund rate, GMV
CRM
Pipeline value by stage, win rate, deal velocity, lead-source attribution, rep performance
Finance
P&L, cash flow, AR/AP aging, expense by category, MRR/ARR, gross margin, runway
Marketing
Campaign ROAS, attribution by touchpoint, email open + click rates, list growth, churn cohort
Helpdesk
First-response time, resolution time, CSAT, ticket volume by category, SLA breach rate
Fleet
Utilization rate, fuel cost per mile, route efficiency, driver scorecard, maintenance backlog
Production
Schedule variance, budget variance, scene-day completion, crew utilization
Hospitality
ADR, RevPAR, occupancy, channel mix, no-show rate, upsell rate, NPS
Health
Patient volume, avg encounter time, no-show rate, prescription patterns, payor mix
HR / Recruit
Time-to-hire, offer-acceptance rate, source quality, attrition by manager, comp-band fairness
LMS
Completion rate, time-on-content, quiz performance, cohort retention
Events
Registration trajectory, check-in rate, session attendance, sponsor ROI
AI-Native Report Building
Two AI surfaces — natural-language report builder + explain-this-number anomaly analysis.
"Build me a report" → user types in natural language → AI selects the optimal collection, aggregation, time range, and chart type. Generates a working report in seconds. Examples:
- "Show me revenue by product category for the last 30 days" → bar chart, grouped by
category.name, summed bylineItems.price, filtered tocreatedAt > now - 30d. - "Which sales reps closed the most deals last quarter?" → table, ranked by
count(deals where stage=closed_won and closedAt in last_quarter), grouped byassignedTo.name. - "Compare ticket volume across our 5 support categories this week vs last week" → 100% stacked bar, x=category, y=count, series=this_week|last_week.
"Explain this number" → user clicks a metric → AI explains the underlying drivers + anomalies. Example: "Revenue down 12% week-over-week" → AI reports "Top driver: Product X (-$45K, -18%) due to a stockout starting Tuesday; new-customer count flat; AOV up 3%." It reads the exact data behind the chart it's explaining, not a summary of it.
The catch: AI doesn't make up data. The model receives the report's actual underlying query result + a structured prompt — no hallucination room. If the data doesn't support an explanation, the AI says so. Pro logs every AI query for audit and lets you route through your own BAA-compliant LLM provider.
Scheduled Delivery + Threshold Alerts
Cron-driven report delivery + real-time threshold alerts.
Schedule a report:
Report: Weekly Revenue P&L
Cadence: Every Monday at 8:00 AM (America/New_York)
Format: PDF + raw CSV attachment
Recipients: cfo@company.com, board-readonly@company.com
Slack: #exec-team
The scheduler renders the report with the data current at delivery time, exports to PDF / Excel / CSV / Parquet, and ships it via email, Slack, or webhook. Timezone-aware (8 AM in EDT, not 8 AM UTC).
Threshold alert:
Metric: Daily ticket-volume (Helpdesk → open_tickets count)
Trigger: > 25% above 7-day rolling average
Cooldown: 4 hours (no re-firing within window)
Action: Page on-call manager + post to #ops-alerts Slack
Alerts run on a 1-minute cron. Anomaly detection options include: absolute threshold, percentage deviation from rolling baseline, z-score from historical mean. Combine conditions with AND/OR. Cooldown prevents alert fatigue.
Goal tracking: set a target ("$2M MRR by Q4"), the system computes daily progress + variance + linear forecast + ETA. Surfaces on dashboards as progress bars. Alerts when off-track by configurable thresholds.
Calculated Fields — Safe Formula Evaluation
Build custom metrics without writing code — and without the security risk of a general-purpose scripting engine.
Standard reports cover the 80%. The other 20% needs calculated fields — "gross margin %" = (revenue - cogs) / revenue. BI's formula engine:
- Sandboxed by design. Formulas can only do arithmetic and lookups against your report's own fields — there's no path to run arbitrary code, read files, or reach outside the report data.
- Operators:
+ - * /arithmetic,< <= > >= == !=comparison,&& ||logical,?:ternary,if(cond, then, else). - Functions:
sum,avg,min,max,count,count_if,round,floor,ceil,abs,pow,sqrt,log,exp,concat,substring,upper,lower,date_diff,date_part,coalesce,null_if. - References:
[Revenue],[Cost.amount],[customer.tier]— fields from the report's data scope. - Type checked. Numeric operators on string fields fail at validation time, not at evaluation. Schema-aware error messages.
- Hardened against malicious input, including pathological patterns designed to hang a naive evaluator.
Calculated fields appear in the field picker like any other; charts, exports, and AI reports all consume them uniformly.
Report Embedding — Share a Report Outside Aeion
Public report toggle
flip a single report to public and get an embeddable URL — `GET /embed/:reportId`.
iframe embed
drop the URL into a standard iframe; no JS SDK or web component required.
Export controls
the embedded view supports the same CSV/PDF/Excel export as the report itself.
Audit log
every export from the tenant's own reports is logged for compliance.
Frequently Asked Questions
No. The visual report builder generates SQL behind the scenes. Pick a collection, choose fields, apply aggregations (sum/avg/count/min/max), drop a chart on it. For complex queries, click "Explain" to see the generated SQL — useful for learning or for handing off to a developer to optimize. Power users can write raw SQL via the Pro custom-query mode.
BI queries through the same service layer as the rest of the platform, not raw SQL joins bolted on after the fact. It coordinates a single logical query across multiple collections — "revenue by sales rep" joins orders to contacts to users automatically. Joins use the same tenant scoping + RBAC as your application; cross-module reports respect every access control rule.
BI is the visual + scheduled + alerting + AI surface every user touches. Underneath it, queries run through the same tenant-scoped database as the rest of the platform, with the same indexing and query-planning Aeion already relies on elsewhere — there's no separate analytics-only infrastructure (like a dedicated read replica) to manage.
"Show me revenue" — the AI asks one clarifying question if ambiguous ("by category, by region, by date range?") then proceeds. Once it has enough context, it builds the report and shows you both the chart AND the underlying query as plain English ("Revenue is summed from paid orders, grouped by..."). You can edit either side and the AI tracks the changes.
Live by default. Dashboard tiles refresh every 60 seconds when a user is viewing the dashboard; idle dashboards don't refresh (saves DB load). Scheduled reports always run on fresh data at delivery time. For high-cardinality dashboards under heavy load, Pro tier enables materialized-view caching with configurable refresh intervals.
Yes — every report exports as CSV / Excel / Parquet (Pro) with the underlying row data, not just the aggregated visualization. Useful for "this looks interesting, let me pull it into Tableau" workflows or for data-warehouse pipelines.
A report is scoped to your own tenant's data the same way it is anywhere else in Aeion — marking it public only changes whether its embed URL requires a login, not which tenant's data it shows. There's no per-viewer JWT or RBAC scoping on the embed itself today, so a public report's URL shows the same data to anyone who has it — treat "public" as visible to anyone with the link, not as a customer-specific view.
Yes on Pro. BI accepts external Postgres / BigQuery / Snowflake connections. Queries are pushed down to the external source where possible (predicate pushdown for joins + filters) and merged with Aeion-resident data in the result set. Custom adapters available for niche systems (e.g., your data lake's REST API).