Property36 staffMelbournecadenceproperty.com.au
How Cadence cut its client presentation time by 50%
Cadence Property Group develops and invests in Australian property. It has 36 staff and about $1 billion in its pipeline.

At a glance
- 50%
- Less time to put together a presentation, "at least" in Brock's estimate
- 7 weeks
- From 5 people using AI to most of the office
- 1 email
- To set up a new deal in DealCloud and its SharePoint folder
“For our team, it's a game changer. ... It used to take forever to put together these docs, so lets us look at more deals.”
Before
Five people were using AI at kickoff, mostly as a chatbot. Putting together a presentation was, in Brock's words, "massively time-consuming", and the executive assistants sorted their inboxes by hand.
What we built
We helped the team build the foundations, then taught them to build on top of them.
- Step 1
Email sorting in the assistants' inboxes. It labels and drafts, a person sends.
- Step 2
A company AI brain that knows Cadence's projects and history.
- Step 3
A shared library of skills anyone in the office can use.
- Step 4
Training for the team, and a working session every fortnight.
After
Seven weeks after kickoff, most of the office was using it, and staff with no coding background were building their own tools. Brock puts presentations at half the time, at least. One assistant called the email sorting "revolutionary". In the week to 23 October the email sorting handled 3,686 emails, and 24 people used the brain.
The full story
Too much reading and reporting
Cadence develops and invests in industrial property. Brock Ellett, its operations manager, describes the side of the business that takes the time: "there's a lot of compliance, there's a lot of administration, there's a lot of reporting, and there's a lot of information processing."
The team already used AI, but mostly to ask it questions. In Brock's words, "it took a long time for people to stop using AI as like a glorified Google search engine." Presentations were the other big cost. Cadence is particular about how they look.
“Everything must look, you know, perfect before it ever sees an external party.”
Not software engineers
Cadence first tried pushing AI further on its own, building small dashboards. Brock is plain about the limit: "the truth is we're in the property sector. You know, we're not software engineers."
They spoke to a few firms and chose us. What mattered most to him was "the speed at which they are able to... come along the journey with us in AI and the flexibility."
Start with the inbox
One of the first things we helped them build was email sorting for the executive assistants' inboxes. It started with labels only, until the labels were right. One assistant remembers the first version: "When we first implemented it, things were just going all over the place and wrong and I was like, it was all too much."
Once it labelled her email the way she would, she was asked if it was saving time. "Oh, so much time."
Sam King, in the transactions team, learned the same thing. "I went a bit too hard at the start," he said, with 15 categories. He cut them back to a few. Now it flags the emails that need his reply, which matters "if I'm out on the road for two hours and I come back to the inbox."
From one build to many
Next came the company AI brain and a shared library of skills. Charlie Buxton, the CEO, explains why the brain beats searching the server every time: "the brain has done that in advance, and it's distilled it into information that's just easy to work through."
With what Brock calls "some nurturing from the Hourglass team", staff started building their own tools. Liam in finance has no coding experience. He built interactive websites that show the finance numbers, and Brock says that reporting now takes about half the time: "Timeframe itself has probably been cut in half as well." The rest of the office noticed: "here's someone with no software engineering experience that has pretty much built an entire platform."
New deals changed too. One of the team describes setting one up now:
“I drag an email and I tell it to set up the deal with DealCloud and set up the property folder in SharePoint and it just does it.”
What runs today
The brain reads Cadence's SharePoint and email, and staff sign in with their Microsoft login. It is backed up to Cadence's own GitHub. The email sorting labels the executives' mail every five minutes and never sends or deletes anything. In the week to 23 October it handled 3,686 emails, and 24 people used the brain.
On top of that sit the team's own tools: a finance dashboard that syncs both ways with Excel, a valuation tracker, a business-plan scenario modeller, and skills for sign-off sheets and equity drawdown notices.
What changed
“AI has really, really cut in half at least the time it takes to put together a suitable presentation.”
Before, five people at Cadence used AI, mostly as a search box, and putting a presentation together was "massively time-consuming". Seven weeks after kickoff, most of the office was using it. Brock is careful with the numbers ("It's hard to put a number on time saving"), but he puts presentations at half the time at least, and the finance reporting at about half.
The team now builds for itself. Brock says most people are making their own data views and web pages in place of spreadsheets, and they named their own AI leads "Claude Lords".
Brock's advice to a friend thinking about working with us:
“This space moves so fast, you need to have someone in your corner.”










