/aiml
AIML Models
Site
1 / 8
Decision intelligence · runs on your warehouse

Subscribe to the forecast.
Run it on your own data.

Exl Athena turns pre-built AI/ML engines into something you subscribe to and run in minutes — sales forecasting, demand, inventory, production and budget planning, churn — all reading your governed warehouse tables and writing results straight back into BI. No data science team required.

7
planning & forecasting engines
one catalogue, one click to subscribe
3
connected plans
demand → production → budget
results written to your warehouse
and surfaced in Data Intelligence & BI
Subscribe · pick engines by industry Configure · map your columns, preview the SQL Decide · results land in BI
Navigate with or the dots below · full-screen for presenting
The problem

Planning lives in disconnected spreadsheets.

Forecasts sit in one workbook, the production plan in another, the budget in a third — none of them agree, and by the time they're reconciled the numbers are stale. Off-the-shelf ML is a black box nobody in operations trusts.

🧩

Siloed plans

Demand, production and budget are built separately, so a change in one never flows to the others.

📦

Black-box ML

Vendor models hide their inputs. Planners can't see the data or the query behind a forecast, so they don't trust it.

🐌

Slow to value

Standing up a model means a data-science project. Months pass before anyone sees a number.

Athena makes the engines a subscription — configured on your data, transparent end-to-end, connected into one plan.

01 · Subscribe

A catalogue of engines — one click to unlock.

Browse the models relevant to your industry, unlock the ones you need, and run them whenever you like. Each card remembers its last run.

Filter by industry — Retail, Manufacturing, or all.

🔓

Unlock a model to subscribe; then Configure & Run anytime.

Every card shows the last run — output table, row count, status.

exl-athena · AIML Models
IndustriesAll IndustriesRetailManufacturing
Subscribed AIML Models
Unlock and run pre-built AI/ML engines. Results are written to your warehouse and appear in Data Intelligence & BI.
📈
Sales Forecasting
Subscribed
Predict future sales volumes from historical data and market trends.
Last run · FACT_SALES_FORECAST · 1,240 rows · 5 of 5 forecast · 2h ago
📦
Inventory Optimization
Subscribed
Optimize stock to minimize holding cost while meeting demand.
📊
Demand Planning
Reconcile forecasts into an actionable, part-level demand plan.
🔁
Customer Churn
Flag customers at risk of leaving and the key attrition drivers.
02 · Configure & run

Transparent, not a black box.

Map your warehouse columns to the model, preview the exact SQL and the real rows it will read, then run. Nothing is hidden.

Map date, measure and dimension to your own columns — smart defaults pre-filled.

SQL

Preview the generated query and 100 sample rows before you commit.

Runs as a background job; results are written to your warehouse.

Run Sales Forecast
Run Sales Forecast
Retail Demo
FACT_SALES
m_datecompleted
m_qtyordered
m_productnum
Product
Monthly
6 months
Data previewSQL
transaction_dateentitymeasure
2026-05-01P-102311,204
2026-05-01P-10244860
2026-06-01P-102311,318
03 · The line-up

Seven engines, one operating picture.

Each is pre-built, configurable on your data, and writes back to the warehouse — pick the ones your business runs on.

📈

Sales Forecasting

Future sales volumes from history and trend, per entity.

forecasting
📊

Demand Planning

Part-level demand via BOM-exploded dependent demand.

demand
📦

Inventory Optimization

Reorder point, safety stock & EOQ from demand + supplier.

inventory
🏭

Production Planning

A feasible plan across work centres and materials.

production
💰

Budget Planning

Revenue, cost, margin, EBITDA & variance from the plan.

budget
🔁

Customer Churn

At-risk customers and drivers from RFM signals.

churn
🅰️

ABC Analysis

Rank inventory or customers by revenue importance.

abc

More landing

The catalogue grows — new engines slot straight in.

roadmap
04 · The differentiator

One connected plan — not three disconnected ones.

The planning engines chain together. Demand feeds production; production feeds the budget — so a change in the forecast flows all the way to EBITDA. This is the S&OP loop most tools can't close.

Step 1

📊 Demand plan

Sold quantities explode through the BOM into component-level dependent demand.

Step 2

🏭 Production plan

Demand becomes a feasible plan across work centres, with cost structure per unit.

Step 3

💰 Budget plan

The production plan consolidates into revenue, cost, margin & EBITDA by cost centre.

🔗 Forecast change → EBITDA, automatically 🧾 Real prices from your sales history 🏢 Roll-ups by plant / region / enterprise
05 · The results

A forecast you can see — and ship.

Every run produces a clear projection with confidence, the top movers, and accuracy you can hold it to — then writes the table back to your warehouse for BI.

History + forecast with a confidence band, per entity.

%

Accuracy (MAPE) reported, so the number is accountable.

Written to your warehouse → live in Data Intelligence & BI.

Sales Forecasting · results
Monthly sales — actual & 6-month forecastunits · entity P-10231
ActualForecast▨ confidence band
+17%
Projected next 6 mo
8.4%
Accuracy · MAPE
5
Entities forecast
1,240
Rows to warehouse
The takeaway

From raw warehouse to a decision — in minutes.

Subscribe to the engines you need, configure them on your own data, and get a connected plan you can act on — results flowing straight into BI.

🔓

Subscribe

Unlock pre-built engines by industry, no data-science project.

⚙️

Configure

Map your columns, preview the SQL and data, then run.

📊

Decide

Connected demand → production → budget, live in BI.

Let's subscribe your first engine.