Skip to content

What we engineer

Our AI Systems.

These are systems we are engineering — askThat, hrThat, ledgerThat, postThat and hector. We engineer the one that fits your job. Nothing is a boxed product you license off a shelf.

01

askThat

Ask your files. Answers point back to the page.

We are engineering a search layer over your manuals, specs and binders. Every answer points back to the page it came from — and stays quiet when the source does not say.

Retrieval-augmented generationVector searchOCRSSO / role-based access
Talk this through

askThat

  • Answers that cite the document and page
  • PDFs, scanned manuals, SOPs and policy binders
  • Declines when the file does not say
  • Access limited to your own documentation
02

hrThat

Leave, attendance and reviews for your team.

We are engineering HR around how your team already works — clock-in, leave, and reviews — so those steps stop living in a spreadsheet.

ReactPostgreSQLPWARole-based access
Talk this through

hrThat

  • Clock-in with a selfie and a location check
  • Leave requests that go to the right person
  • Review cycles you can actually finish
  • Works on a phone in the field
03

ledgerThat

Till, stock and the books in one place.

We are engineering a till and ledger so your sales, stock and accounts stay in sync — without retyping the day into a spreadsheet at night.

Next.jsPostgreSQLPOSMulti-tenant
Talk this through

ledgerThat

  • A till that writes to your books
  • Stock that matches what you sold
  • Payments you can see at close of day
  • A record you can show at month-end
04

postThat

Plan and send your social posts from one queue.

We are engineering a queue for your social posts — write once, schedule, and send — so you are not hopping between five apps.

ReactSchedulingSocial APIs
Talk this through

postThat

  • One queue for the channels you already use
  • Schedule ahead, change before it goes out
  • A simple calendar of what is next
  • Handover on your own accounts
05

hector

A local assistant on your machine, for your work.

We are engineering a local assistant that runs on your machine — for the jobs you do not want leaving the building.

Local modelsOn-device agentsPython
Talk this through

hector

  • Runs on your device first
  • Helps with the files and tasks you already have
  • Calls the cloud only when you choose
  • Yours to keep, on your hardware
The AI layer

Where a model might earn its place

AI is an extension of the work above, not a separate product. It would run on the data those systems already capture — and we will say so plainly when ordinary automation is the cheaper answer.

Predict, don't just monitor

Anomaly detection and forecasting models over the telemetry you already collect — so a failing pump, a drifting cold room or a slow month shows up before it costs you.

  • Baselines learned per asset or branch
  • Failure and demand forecasting
  • Alerts wired into the channels you use

Read the paperwork for you

Document and OCR extraction feeding askThat and ledgerThat, so manuals stay searchable and invoices, delivery notes and orders land in your accounting without anyone retyping them.

  • Invoice and delivery-note extraction
  • Confidence scoring with human review
  • Written straight into your systems

Judgement inside workflows

AI-assisted checks, classification and routing inside hrThat steps — triaging requests, grading photo evidence and sending each job to the right person automatically.

  • Automatic triage and routing
  • Photo and free-text classification
  • Every decision logged for audit

Not sure which starting point fits?

Book a walkthrough and we will map what you run today against these patterns before recommending anything.

Book a Walkthrough