Prove the control logic before commissioning,
with AI-enabled functional acceptance testing.

Check the PLC program against the control philosophy — with real PLC hardware or simulation — and verify field-device behaviour against a deterministic digital twin. See what the logic commanded, how the equipment responded, and why every check passed or failed, scan by scan.

We tailor the demo to your process — so you can see, on your own sequence, how much commissioning time and how many late control-logic surprises UnitFAT takes out.

UnitFAT system overview showing automated program tests, digital-twin field devices, PLC and simulation connectivity, and an AI-editable framework.
TEST INTENT PLC DECISION SIMULATED RESPONSE VERIFY

How it works

One Closed Test Loop:
Verify PLC Program, Control Logic, Plant Response.

The test program applies deliberate operator and fault stimuli. The PLC keeps ownership of permissives, interlocks, and sequence commands. The digital twin returns realistic device timing and process feedback — and every check records the whole chain.

EDGE64 / UNITFAT CONTROL-LOGIC FAT LOOP
Program Automatic Testing Running
01
TEST PROGRAM

Sequential program testing

Observe while running

START FAT PERMISSIVE CHECK DOSING SEQUENCE HOLD / RESUME ROUTE DISCHARGE PLC STATE IDLE · READY OUTPUT COMMAND FV100_CMD · ON SIM FEEDBACK ZSO100 · PROVEN ASSERTION PASS
T08 · SCAN 142 · RUNNING
02
CONNECTIVITY

PLC / simulation connectivity

Controlled I/O boundary

REFERENCE PLC · RUN BACKPLANE BUS PS CPU SCAN DI AI DO AO TEST I/O SIM I/O STIMULI FEEDBACK SCAN EXCHANGE
REFERENCE PLC · IN PROCESS
03
PLANT MODEL

Field-device simulation

P&ID + connected device nodes

TK-100 FV-100 P-100 FT FT-100 TO PROCESS ON/OFF SHUTOFF CENTRIFUGAL FV-100 · VALVE OPEN P-100 · PUMP RUNNING INPUT ← PLC Open command ON OUTPUT → PLC Open proof ON INPUT ← PLC Start command ON OUTPUT → PLC Running feedback ON PROCESS MODEL FROM DEVICES valve open / pump run TO FT-100 18.4 L/min
COMMAND → PHYSICS → FEEDBACK

Capabilities

DIGITAL TWIN LITE

Simulate instrument failure while the sequence runs.

Abnormal conditions are the hardest part of a manual FAT to stage. UnitFAT injects them at the exact step where the behaviour matters, then keeps the whole response on record.

  • Stuck valves, failed proofs, and dead transmitters on demand
  • Injected mid-phase, without editing the control program
  • Repeat the same failure at the same scan, every run
Digital Twin Lite simulating an instrument failure during a test sequence
AI drafting, editing, and verifying UnitFAT test sequences Cursor Claude OpenAI Codex
AI-ASSISTED TESTING

AI writes, edits, and verifies the test sequences.

The test model is plain, readable definitions — so an AI assistant can draft sequences from your control philosophy and re-verify them after every logic change, with the engineer reviewing the diff.

  • Draft test phases and checks straight from the control philosophy
  • Regression-verify the suite after a program change
  • Works with the tools your team already uses — Claude, OpenAI Codex

Outcomes

What your team gets out of UnitFAT.

Program testing today takes weeks to set up, and there is no dedicated platform built for it — so it gets done manually, under commissioning pressure, and rarely the same way twice. UnitFAT is that platform: AI-assisted, autonomous, and repeatable from the first run.

PLC program verified against the control philosophy
SYSTEM INTEGRATORS

Bring structure to FAT execution.

  • Turn a sequence narrative into reviewable test phases and checks.
  • Re-run abnormal conditions without rebuilding the test manually.
  • Use one evidence trail for controls, process, and client review.
BATCH-EQUIPMENT OEMs

Reuse a proven test model.

  • Keep standard equipment behaviour consistent across machine variants.
  • Exercise recipes, routing, device timing, and failure response before site.
  • Make control changes easier to regression-test on the next build.
COMMISSIONING TEAMS

Take fewer surprises to site.

  • Arrive with a sequence that has already been proven end to end.
  • Follow PLC commands and simulated feedback scan by scan when something diverges.
  • Turn a site issue into one exact scan you can replay back at the shop.

About us

Making generative AI genuinely useful in industrial automation.

JA
Jees Antony CEO & Principal Engineer · 12+ years in Industrial Automation LinkedIn

Edge64 is committed to improving generative-AI adoption in industrial automation — building engineering tools where the work is actually delivered: control logic, testing, commissioning, and the evidence that follows.

UnitFAT is our first product. Because the whole test model is AI readable, engineers can extend or inspect it with the tools they already use, from Claude to OpenAI Codex.

Our next milestone

A Production-grade SCADA platform that AI and Automation Engineers can co-develop.

Join the waitlist

UnitFAT proved that control logic can be tested and replayed with the help of AI, and that improved Engineers efficiency. The same discipline now moves up the stack — to the HMI screens, alarms, and supervisory logic that get changed most often and verified least.

  • 01
    Readable by engineers and AI alike Screens, tags, and supervisory logic live as plain text definitions — open a file and read the intent, instead of clicking through a proprietary editor to find it.
  • 02
    Every edit proven before it ships Each change is replayed against simulation before it reaches the plant — the same verification loop UnitFAT already runs on control logic.
  • 03
    Built for plant reality On-premise and offline-capable, fully versioned, with an audit trail that stands up in a FAT review or a post-incident investigation.

Contact us

Bring us one control sequence that matters.

Your time matters, so the first conversation stays specific: one process sequence, the abnormal conditions your engineers struggle to recreate, and the evidence your FAT reviewers expect.

What you'll walk away with

  • Where AI fits your control logic — and where it puts you ahead of the competition
  • Where your commissioning time leaks — the sequence gaps your team hits on site, during and after handover
  • A concrete read on how UnitFAT would improve the functional testing of your program

No prep needed. No sales deck. If it is not a fit, we will say so on the call.

BOOK YOUR REVIEW

Tell us where to send the invite.

45 minutes, engineer to engineer. No prep needed.

Your details are used only to respond to this technical enquiry.