What is human approved automation?
The plain definition of the category BlueArc builds in: AI that asks first, a person who signs off, and a receipt behind every action. Written for operators, not analysts.
The definition, in plain speech.
Human approved automation is software where AI does the tedious work and a person signs off every decision that matters. The AI reads the messy inputs, drafts the quote, the entry or the order, then stops and asks. Nothing ships, lodges or invoices itself; every action waits for your green light and leaves a receipt.
Unlike set and forget tools that act on their own rules, and unlike DIY ChatGPT experiments that leave you checking everything yourself, human approved automation keeps your judgement in charge of the outcome. BlueArc builds it into seven products for Australian operators, hosted in Australia in the Sydney region.
Definition current as at 24 July 2026.
Three ways to put AI to work.
| What matters | Human approved automation | Set and forget automation | DIY ChatGPT |
|---|---|---|---|
| Who decides | A person signs off every decision that matters | The tool acts whenever its trigger matches | You do, after checking every answer yourself |
| What the AI does | Reads the mess and proposes the work | Fires the mapped action, right or wrong | Drafts text when you remember to ask |
| Proof afterwards | A recorded confirmation behind every action | Logs, if the tool keeps them | A chat history, if you kept it |
| Knows your trade | Built around one workflow, localised for AU and NZ | Generic triggers and actions | Generic answers |
| When it is unsure | It asks first | It acts anyway | It answers anyway |
The Green Light Method, in one pass.
Propose
The AI turns what arrives, a message, a photo, a sheet, a feed, into the work it implies.
Check
The proposal sits beside its source so a person can compare the two at a glance.
Green light
A person approves it. The moment of consequence always belongs to a human.
Act
The software carries out exactly what was approved, and nothing else.
Receipt
The action is recorded, so the proof exists before anyone asks for it.
The full loop, step by step and product by product, lives on the Green Light Method page. The beliefs behind it are on the BlueArc manifesto, under the founder's name.
Honest answers.
What is human approved automation?
Software where AI reads the messy inputs and proposes the work, and a person signs off every decision that matters before anything happens. The proposal is never the action: it waits for a green light, and the confirmed action is recorded.
How is it different from set and forget automation?
A set and forget tool fires its action when its trigger matches, right or wrong, and you find out later. Human approved automation stops at the moment of consequence: the AI drafts the entry, the order or the quote, and a person confirms it before it goes anywhere.
Is that not just a person checking everything the AI does?
The AI does the reading, sorting and drafting, which is where the hours go. The person makes the calls that matter: approve the quote, release the order, confirm the position. Checking a prepared proposal takes seconds; preparing it used to take the evening.
What is the Green Light Method?
The five step loop BlueArc products run: Propose, Check, Green light, Act, Receipt. The AI proposes work from the inputs, a person checks it, gives the green light, the software acts, and the action is recorded so there is a receipt to point at later.
Does the AI ever act on its own?
No. Nothing ships, lodges, invoices or certifies itself in a BlueArc product. Every proposal waits in a review queue for a person, and that control never leaves a human.
your green light.
Book a short, tailored walkthrough on your own quotes, orders or sheets. No obligation.
Book a demo →