The platform

A computation layer between your data and the answer.

ORE doesn't ask a language model to do your maths. A deterministic engine computes every metric, trend and forecast first — then the AI explains the result in plain English. Exact numbers. Real statistics. No hallucinated trends.

The principle

The AI is the narrator, not the calculator.

Generic AI tools hand raw rows to a language model and hope it spots the trend. It can't compute a percentage, has no statistical basis for "revenue is declining," and will confidently invent a number. ORE flips this: every figure is pre-computed, the model only translates.

  • Deterministic maths. Pandas, NumPy and SciPy compute aggregations, correlations and significance tests.
  • Grounded narration. The model receives a structured result and is told: these numbers are exact — do not recalculate.
  • Lead with what matters. Every answer opens with the single most important finding, then the action.
ore — query pipeline
user question
↳ intent extraction
↳ query execution
↳ analytics engine // computes
↳ structured result
↳ AI narrative // interprets
↳ reply + chart
Analytics engine

Ask it the questions you actually have.

ORE detects what you're really asking — a ranking, a trend, a comparison, a root cause — and routes it to the right computation. Then it picks the right chart for you, automatically.

Rankings & share

"Top customers by revenue." Groups, sums, sorts and computes share-of-total — rendered as a clean bar chart.

Trends over time

Linear regression with R-squared and direction. "Is revenue actually declining?" gets a statistically honest answer.

Period comparisons

"Compare Q1 to Q2." Sums each period, computes the change, and shows them side by side.

Composition

"Revenue by channel." Share-of-total across categories as a pie or horizontal bar depending on how many there are.

Correlation

Pearson correlation with p-values and strength classification — relationships you can trust, not coincidences.

Root cause

Comparison plus per-entity contribution, rendered as a waterfall so you can see exactly what moved the number.

Forecasting & prediction

See what's coming, not just what happened.

ORE fits time-series models with weekly and yearly seasonality to project where you're heading — and falls back to a robust linear projection when there isn't enough history. Either way you get confidence bands, not false precision.

  • Revenue projection. "Revenue projected to grow +12% over the next 90 days," with upper and lower bounds.
  • Demand by SKU. Per-product forecasts ranked by recent demand, so you order ahead of the curve.
  • Days-to-stockout. Velocity vs. on-hand vs. lead time → a ranked list of what to reorder, by urgency, with the supplier named.
ore — forecast
Forecast revenue for the next quarter.
ORE · seasonal model · 28ms Revenue is trending up — projected +12% over 90 days (from $1.84M to $2.06M), seasonal model.
Always watching

Catch the problem before it catches you.

Anomaly detection

Statistical outlier scoring flags the freight cost, order or supplier that's off-pattern, with a severity badge so you know what to look at first — critical, high, medium or info.

CriticalHighInfo

Autonomous workers

Assign a worker to a domain — inventory, sales, suppliers — and it monitors on a schedule, runs the same rigorous analytics, and surfaces findings or generates its report without being asked.

Inventory · activeSales · active
Briefs & reports

From one question to the whole picture.

Ask a big question and ORE composes a multi-section answer: executive summary, revenue, customer health, inventory status, delivery performance — plus recommendations. Need it for the board? Get it as a formatted PDF.

  • Analytical briefs. "How is my business performing?" → a structured, sectioned answer inline in chat.
  • Board-ready PDFs. Monthly operations, supplier review, inventory health — charts rendered, branded, downloadable.
  • On a schedule. A worker can auto-generate its report on the first of the month and deliver it to your desk.
monthly_operations.pdf
Executive summary Revenue up 12% MoM. Two suppliers slipped on lead time. Three SKUs need reorder this week.
Sections Revenue · Customer health · Inventory · Delivery · Recommendations
Generate a sample report
Connects to what you run

One structured picture of your operation.

Orders & order lines

Revenue, tax, discounts, channel, city, gross profit by product — the transactional heart of the business.

Inventory

Quantity on hand, reorder points, inventory value and velocity — the basis for every stockout projection.

Shipments & freight

Carrier performance, delivery delay, freight-as-percentage — where margin quietly leaks.

Suppliers

Lead times, contact details and performance — so a slipping supplier is visible, not a surprise.

Customers

Order counts, value and health — to find who's growing and who's quietly churning.

Your sources

ERP, WMS, CRM and Microsoft 365 — ORE connects the dots without forcing you to rip and replace.

Get started

Point ORE at your data. Watch it compute.

A 30-minute demo on real supply chain data — bring the question your current tools won't answer.