ArexAi logoArexAi

Real product evidence

See the interface. Reproduce the workflow.

This gallery uses the live ArexAi Analyst interface—not a conceptual marketing render. Product claims below are tied to visible flows, current limits and documented privacy boundaries.

Synthetic example · no customer data

Reproduce a verified result

Before: two separate CSV files

CustomerId,Revenue
C001,120
C002,180
C001,100
CustomerId,Segment
C001,Retail
C002,Wholesale

Operation: match CustomerId, approve the relationship, group Revenue by Segment and calculate Sum.

After: Retail 220 · Wholesale 180 · Total 400

Source check: sales B2 + B4 = 220; B3 = 180. All 3 records match unique customer keys. The report includes cell evidence and a source fingerprint.

Other reproducible checks: Revenue 400; Cost 240; Actual 400 versus Target 390 gives +10; Stock totals 22. Values are unitless synthetic observations, not business performance claims.

Synthetic example · no customer data

Sales total

Before: three Revenue values. Operation: sum Revenue. After: 400; verify every contributing cell.

120 + 180 + 100 = 400

synthetic-sales.csv · B2 + B3 + B4

CSV ↓ synthetic-sales.csv

Synthetic example · no customer data

Inventory quantity

Before: three item quantities. Operation: sum Stock and review the smallest quantity. After: 22 in total; P002 has 3. No reorder is made.

8 + 3 + 11 = 22

synthetic-inventory.csv · B2 + B3 + B4

CSV ↓ synthetic-inventory.csv

Synthetic example · no customer data

Actual versus target

Before: Actual and Target for three periods. Operation: compare totals. After: absolute variance +10. This does not imply a forecast.

400 − 390 = +10

synthetic-budget.csv · SUM(B2:B4) − SUM(C2:C4)

CSV ↓ synthetic-budget.csv
Live ArexAi Analyst upload workspace showing supported spreadsheet inputs and example prompts
Captured from the live ArexAi Analyst application on 29 September 2026. Interface details can evolve as the Early Access Beta is updated.
1

Profile a workbook

Upload XLSX, XLSM or CSV and inspect sheet structure, field types, missing values, duplicates and ranges.

2

Ask a business question

Request KPIs, trends, anomalies or cross-sheet comparisons in plain language.

3

Verify the calculation

Use calculation-first analysis so numeric claims are checked against workbook-derived context before explanation.

4

Export a reviewed copy

Approve a workbook plan before ArexAi creates or redesigns an .xlsx output and verifies its integrity.

Current beta limits

8 MB maximum file size

Up to 20 sheets per workbook

Up to 1,000 sampled rows per sheet

Up to 100 detected fields

Privacy boundary

Raw workbook rows remain browser-local. AI requests use minimized structural or statistical context. Users should still remove secrets and unnecessary personal data before any analysis.

Read the complete privacy policy