SAS-AM Applied IntelligenceBook a conversation (opens in a new tab)

Practical AI, kept in Australia.

We help Australian businesses choose one AI project worth doing, build it, and hand it to their own team. Where your data calls for it, the system runs privately on Australian infrastructure.

Take the readiness check

How the work runs

How an engagement runs, in four steps.

We own a torque wrench to get the right torque, not to tighten bolts. We treat AI the same way: a tool, bought for the outcome it delivers.

  1. 01

    We ask before we advise

    We start with your workflows. You tell us where the time goes, and we keep asking until the real problem is on the table.

  2. 02

    We pick one thing worth doing

    Every candidate is scored for value, risk and how ready the data is. Most ideas do not survive this. One or two do, and those are the ones we scope.

  3. 03

    We build it beside your team

    A working system on real work, with a person reviewing at each checkpoint. Where the data demands it, it runs on Australian infrastructure and stays there.

  4. 04

    We hand it over

    Your people learn to run it, check it and switch it off. We measure the result against the number we agreed at the start.

Proof

Results from delivered work.

Read the case studies

A major Australian transport operator

3,200 documents sorted in two days. The estimate by hand was three to four weeks.

Before a major review, someone has to find the documents that matter. The system did the sorting. Our people made every judgement.

The same pattern applies to supplier contracts, tender responses and policy libraries.

to triage 3,200 documents, against an estimate of 3 to 4 weeks
2 days
accuracy on checked samples
94%
fewer consultant hours on document review
~60%

Source: SAS-AM project record, client anonymised.

An asset intensive operator

Thousands of free text job records, turned into data a manager can act on.

Each record was matched to a known cause of failure, with a confidence level. Anything uncertain went to a person. The system also surfaced causes nobody had written down.

The same pattern applies to service tickets, customer complaints and warranty claims.

job records processed
3,126
classified by the system, the remainder went to a person
99.2%
causes of failure the existing analysis had missed
337

Source: SAS-AM project record, client anonymised.

These results come from large, asset intensive organisations. We do not yet have a published result from a smaller business. That is why our first engagement has a fixed scope you can read before you commit.

Client names are withheld by agreement. References are available in conversation.

What we believe

AI should do the dishes.Your people should do the artwork.

Many people use AI to do their artwork while they still do their dishes. We think that is backwards. Our work starts by finding the busy work, so we can take it away.

From the masterclass Less Hype, More Real AI Please, presented at MAINSTREAM 2026.

Read the thinking behind it
A watercolour in two halves. On the left, a tired person washes a pile of dishes while a robot arm paints at an easel. On the right, a smiling person paints at an easel while a small robot washes the dishes.

People decide

A person at every checkpoint.

Nothing an AI system drafts leaves the building until someone accountable has looked at it. How closely they look depends on what it costs to be wrong.

Work itemReview levelStatus
Supplier invoices matched to purchase ordersSpot checkApproved
Reply to a customer complaintFull reviewAwaiting review
Weekly operations summarySpot checkApproved
Change to a safety procedureDirector sign offHeld for sign off
Illustrative example of a review queue
Shane Scriven, smiling with arms folded, standing outdoors beside a row of trees.

Who you will deal with

I never want my work to produce more work for somebody. That is a hard rule.

Shane Scriven

Managing Director, SAS Asset Management

  • Writes the practice's published articles on AI, from work he has delivered.
  • Presented the masterclass Less Hype, More Real AI Please at MAINSTREAM 2026.
  • Runs this practice day to day on the same governed AI systems we build for clients.
About the practice

Where to start

Start wherever suits you.

Three ways in. The first asks for three minutes and nothing else. You only hear a price when you ask for one.

  1. 1

    Take the readiness check

    Eight questions and an honest result, shown straight away.

    Three minutes, no details

    Start the check

  2. 2

    Request the starter scope

    A fixed scope first engagement. See exactly what you get before you ask the price.

    A short form

    See the scope

  3. 3

    Book a conversation

    Thirty minutes with Shane Scriven, who will ask more than he tells.

    Half an hour

    Book a conversation (opens in a new tab)