Omnatix

Decisions from the data a business already has.

Five dashboards across five industries, each with an AI you can question in plain language. Open any of them and it shows you where the money is and what to do next.

Leshaw is a real client, shown with permission, on sample data. The other four businesses are fictional, with findings planted on purpose, because a demo that only displays numbers proves nothing.

Boards and hardware

Leshaw Business Hub

One live view of the day: counter sales, trade quotes, the cutting floor and stock.

Operations hub with an AI assistant on the live figures
Open demoLive sample data
FMCG distribution

Meridian Provisions Co.

Revenue is up double digits. Gross profit is flat. Nobody can say why.

R12.6m
of gross profit that did not arrive
Open demoDemo period to 2026-08-31
Freight and fleet

Kestrel Logistics

Delivery is treated as fixed overhead, so nobody knows which routes lose money.

R19.7m
identified against R27.6m of contribution
Open demoDemo period to 2026-09-01
Microfinance

Sable & Finch Credit

Arrears are reported monthly and backward looking. By then the money is gone.

R10.3m
of credit loss above the clean book rate
Open demoDemo period to 2026-08-31
Clinic group

Lumen Health Network

Every clinic is full and the group still loses money. Capacity and cash are managed apart.

R17.4m
billed for care delivered and never collected
Open demoDemo period to 2026-08-31

An open AI model (Qwen, or Gemma or Nemotron when it is busy) answers from SQL that runs in your browser on each dashboard's own data. Every answer with figures shows its query and rows.

From raw files

CSV exports land as they are. Nothing is corrected by hand. Mixed date formats, duplicate rows, missing costs and orphan product codes are all resolved in the pipeline, visibly.

Tested every build

Bronze, silver and gold layers in dbt, with a test suite that runs on every refresh. A failing test stops the publish, so a number nobody checked never reaches a screen.

No server

Each dashboard runs its own query engine in your browser, with no database to host, which is why these cost almost nothing to keep online. Ask any of them a question and an open AI model answers from SQL you can see, run on that dashboard's own data.

Honest about age

On a client's live data, each dashboard counts the days since its newest transaction, against your clock, and turns red when a feed stops. The four fictional demos cover a fixed period, so their badge tracks the nightly rebuild instead.

See what this finds in your own data.

Send us a few of the exports your systems already produce, and we will show you what the same review finds in your own numbers, with every finding ranked by what it is worth.