Karbon Excellence Awards Podcast

From a team perspective, I would
say that rather than AI helping us

live up to our values, I would say
our values have helped us remain on

course and how we harness AI and tech.

The Karbon Excellence Awards recognize
the accounting firms that are

raising the bar for the industry.

On this podcast, we ask leaders
of the winning firms a series

of questions about how they're
taking their firm's culture, client

experience, technology, and leadership

to the next level.

I'm your host, Twyla Verhelst, and this is
the 2025 Karbon Excellence Awards podcast.

Founded in 2009 +MORE group is a
network of New Zealand owned firms

delivering seasoned expertise,
practical solutions, and positive

outcomes for ambitious business owners.

As AI continues to disrupt the industry,
+MORE group is leading the change.

So much so, they've been recognized
as one of the 2025 Karbon Excellence

Award winners for AI Innovation.

Today, I'm joined by Megan
Plumridge from the +MORE Group.

She's a partner of the MGI+MORE
team based in Auckland.

We're here to explore the innovative ways
they're embracing AI in their practice.

Welcome Megan.

In your nomination form, Megan, you've
mentioned how you use AI in your

robotic process automation program,
VBA scripts and DAX functions.

These are pretty technical terms
that maybe not every accounting

professional will be familiar with.

What do these terms mean and
how do you develop this type of

technical knowledge inside your firm?

That's a really good question.

And to be honest, they're not terms
that I would've known five years ago.

I'll start with what they are.

So basically, they're all different
types of programming language that you

use within different types of software.

So VBA, for example, is used to automate
functions within Microsoft Office tools.

DAX is a language that you use to
create formulas within Power BI.

We also use other languages such as
which we use to sort of process data

that we've taken out of our Karbon
API and put it into other formats.

Or HTML, which we use to
format automated emails.

So there's all sorts of bits
of, language in there that we've

had to become familiar with.

But to be honest, with, the increasing
AI capability, we didn't necessarily

need to be totally across what all
these tools and these languages did

when we first started using them, and
that we just posed the right prompt

to our favorite GPT, that sort of led
us in the direction of what we needed

to do in order to build our knowledge.

From memory, we started with Power
BI, and we had to build DAX knowledge

by asking, our prompt would be, how
do I make this graph look like this?

Or how do I develop this formula,
so that can join these bits of

data together and you can use,
you know, very basic language.

Because obviously at that point you don't
really know what you're trying to do.

You just know what output you want.

And that's the key bit is knowing
what you're trying to achieve.

So once you've got a DAX query
developed in your GPT, you can then

start asking it questions as to what
each bit of that query does, so you

can build up your understanding.

in terms of how we've managed to
do this in a busy accounting firm.

I mean our, our accountants and
our partners are all client facing

and they've got plenty of client
delivery work to keep them busy.

So the fact that we have this firm
group model is really where that

capability has come from because we
share a head office capability, and

that's including an operations and a
business performance improvement function.

And so we've collaborated between
the professionals who know what

they want to see and what would
be cool to automate, for example.

And the operations and business
improvement team have developed their

skills and ability to deliver it.

About five years ago, when we first
started doing automation work, we

didn't really know how to do it either.

So they've had to learn.

It's not something that they arrived
fully formed with the skills to do it.

They've kind of built up their knowledge
and capability over the years just

by being curious and asking a lot
of questions and watching YouTube

videos and all the usual things we do
these days to, to improve our skills.

In terms of how to start out,
I think you just have to know

what you want to achieve.

So, be curious and have, have something
you want to do differently or better,

and start asking your ChatGPT or your
Copilot or your Gemini, whatever tool

you use, how, how to go about it.

And they will help you develop a
data flow or a process and you keep

asking questions and drill down into
something that is useful for you.

Excellent.

Thanks for sharing that.

I mean, there's so many ways that people
are diving into AI and getting that

comfort level built up and it's always
so interesting to hear how, how the

really innovative firms are doing that.

And so maybe let's talk a little
bit about the time savings.

I understand that inside of
your organization that

you're saving approximately

two weeks on building your reporting
dashboards, and a week or so on

structuring complex client data.

And you've been able to
launch some automated daily

reporting in the past 12 months.

So with data and business intelligence
systems being the focus of your firm,

which is an area where, you know, many
firms are kind of hesitant with AI.

From your experience, what are
some of the practical advantages

of you using AI in this way?

And then on the opposite side of that,
what precautions do you suggest that

other firms take in order to ensure
their data is secure and it's accurate?

We are very familiar in how our
clients, are able to grow and build

their businesses through understanding
their own business performance.

Through monthly reporting, through
KPIs, making sure that their own

data is up to date and accurate.

And so, to do the same within our firm
with our processes and our workflows

just seemed like a bit of a no brainer.

Lacking the data to understand
where our roadblocks were

was really frustrating to me.

I'm, I'm obviously a bit of a data
junkie, but I found it really frustrating

just not understanding where, where
our, our troublesome points were.

At the time that we started, particularly
our dashboarding, it was just post COVID

and we had a real struggle to ensure we
maintained our tax filing in line with

what was required by our online revenue.

And our clients had a lot of
other things to worry about.

They needed to support another areas
and tax returns was really their

last, the last of their concerns.

So, we needed to make sure we kept
on, on top of our filing and we had

real visibility as to what was going
on to make sure that they were kept

in, in line with legislation basically.

So we developed a tax filing
dashboard using Power BI.

But it was a real clunky beast
in that you had to download

four or five reports manually.

And then we developed that
into a Power BI dashboard.

We had a memorable couple of weeks
where it was just before the final

deadline and our business performance
manager, Matt, who's the only person

who could work it and who had the
license, went on paternity leave, and

I ended up drawing the dashboards on a
big spreadsheet for a couple of weeks.

And we realized that we needed, we
definitely needed a more robust tool.

We needed something that just took all
the drama out of making sure we had that

really vital information at our fingertips
where we were lagging, and that we were

gonna make our IRD filing deadline.

So, Matt sort of got into, he had
got into Power Automate, which is an

automation tool within Microsoft, and
developed ways of pulling the data

together and automated those dashboards.

And that was really our first
sort of little automation that

we developed within the firm.

In terms of precautions, I think
providing ongoing training to the

team is the most important thing.

That they understand where the
risks lie, they understand where

the power is, and and they can make
their own intelligent decisions as

to what they're doing with the data.

Giving them access to all of
ChatGPT or Copilot or whatever that

is licensed, not the free versions,
so that they can confidently

import reports or data or whatever,

and not have to worry is also a
key, I think a key requirement.

I know there is a cost to that.

They are well-trained.

They can sort of make up the cost of that
license quite quickly, through the, the

efficiencies they gain using the product.

So we've taken that quite
seriously in terms of our training.

We're actually doing ongoing,
training course in October where

every day the team have some sort of
video they need to watch in terms of

building up their Copilot knowledge.

We're using Copilot currently
as a, as a Microsoft firm.

So a whole month, every day they need to
watch a video of a couple of minutes in

terms of something that we are going to

designate to them, to
improve their skills.

So even though we've been using these
tools for a while, the training is

ongoing because they keep changing.

What's available and what they
can do keeps changing, so the

training needs to continue.

We also need to be really
cognizant of accuracy.

So you can't necessarily rely on any agent
or AI generated output to be accurate.

For one thing, AI really wants to make you
happy, so it'll tell you things that it,

thinks you want to hear really lovely, but

not

helpful when it comes to data.

So, we make the team know that
you need to check any AR generated

output, and any automation that we
build has to be rigorously checked

before we can actually roll it out.

And actually I re, I recheck
it quite regularly, especially

some of our dashboards.

At

no

point do we just assume it's right
and going to continue being right.

Because ultimately it
is just, it is a bot.

It's doing what we've told it to do.

And sometimes it can, it
can go a little bit awry.

The final piece of the puzzle,
I think is our AI policy.

So ensuring that we review this regularly
and we have an AI tech committee now,

which reviews the policy quarterly.

You make sure that it's a
living, breathing document,

that people are aware of it.

You haven't just stuck it on the internet
or stuck it in a folder somewhere.

Because the tech keeps evolving,
you need to make sure it's relevant

and it's covering the risks and
the particular risks to business.

I mean, the wonderful thing about AI
policy is you can get Copilot or ChatGPT

to help you write it and develop it.

And then just talk about it to
decide which bits of it makes sense.

I think you're , you're investing
in all the right places, right?

In terms of your training, you know,
giving, giving your team that, empowerment

to make decisions as to what is secure.

One of the terms that I like to use
is ' the rules of the road don't

change', in terms of you're using AI,
but you know your integrity and your,

you know, what you're, you're doing for
a client and the work that you produce

is all still held to that same standard.

You're just using a
different tool to get there.

But the, you know, the same sort
of, rules that you would have around

doing the work by hand manually
versus AI still stay the same.

Exactly that.

And making sure that the time
we save doesn't impact our

client relationships at all.

It has to improve our
client relationships, not

remove the need for

client relationships.

Obviously, relationships remain the
most important part of our business.

Yeah,

absolutely.

Keeping the human to human front
and center and leveraging AI.

On that note, I understand that you
have a proprietary bot, and I've heard

through the grapevine that her name is
Rosie, so you could Is her name Rosie?

Rosie which

was

Okay.

name.

We

had

Oh

we held a vote and that was
the name that was chosen.

I love that.

I love that.

So, tell us about that undertaking.

I mean, I can only assume that that
was a big undertaking, but tell us

about Rosie and what was involved
with building her and would you

suggest that other firms build their
own bot or Rosie to use internally?

I actually think when we started
this project, we had no idea what

we were getting ourselves into.

So I mean, that, that's almost the best
way to go into these, into these sort of

large, significant undertakings because
we probably wouldn't have done it.

But, we, we'd basically investigated
robotic process automation, RPA,

a couple of years previously.

I think pre COVID, and decided
against it as an option at the

time because it required a really
high initial capital outlay.

the consultants that could do the
work for us they'd been working

on really big corporates, you
know, within the health system.

And it was just much bigger than what we
needed at the moment for a firm our size.

Post COVID, we had a consultant reach
out to us and they had a new model to

deliver robotic process automation,
which had become available where we pay

a monthly subscription and we help them,
or we work with them to develop the bot.

And during that time,
there's no capital outlay

during

the build period.

We are just working
really closely with them.

And then once the bot goes
live, can work with them on any

enhancements they need and the, the
pricing basically stays the same.

So it was a really accessible way for
us to get into more serious robotic

pro process automation without having
to throw a whole bunch of cash at

something we weren't sure was gonna work.

Basically, with the whole process
was developing a product design.

It was like a tech, any
tech project really.

We had a product design document,
and then did a bunch of testing and

it slowly rolled it out and we had
the whole minimal viable product and

then ongoing enhancements to get it
to a point where it was doing sort of

enough to make it commercially viable.

And basically what Rosie does is
at the start of our annual accounts

process, she goes in and downloads all
the reports and documentation we need

from a variety of different sources.

So from our inland revenue,
from our ACC, which is our,

our, government health provider.

From Xero or other accounting software
and stores it all against the work

papers and completes the work papers
for specific areas of the accounts

that we know are relatively low risk
but still need to be taken care of.

Then that's left for the
accountant to go in and do the

more complex areas of the work.

Paper and review and finalized.

So it, it's not at all that
Rosie comes in and completes

an end-to-end process for us.

It's, it really is taking away that pain
point area of the bit where we didn't

feel like humans would necessarily
adding any value, but it had to get done.

Generally, I would say these days I
wouldn't advise the average firm to go

in and build their own proprietary block.

'cause I think that was a time when
there wasn't a lot of option out there.

and subsequently there's amazing products
out there that can be utilized that

doesn't require the amount of resource.

We, we threw it at over a
number of years in order to

sort of get it to a good state.

At the same time, you know, we love
Rosie and we talk about her all the time.

After we had Rosie working on our annual
accounts process, we then adapted her

to also do our management reporting.

So the same process for our management
accounts that we could get the first

hour, say the first hour of a management
reporting process taken care of by the

bot, then an accountant would come in
and do the value add part of the work.

So working across those two work types, we
then rolled that out to our other offices.

'Cause initially Rosie was only working
in Auckland it got to a point where

Rosie's now at full capacity and, we have
to decide what we're going to do next.

But yeah, as I said, I don't, I don't
think it's necessary at this point

feel you need to build a bot in order
to take advantage of automation.

I'm sure it's the same in the
States, but in New Zealand and across

Australasia, there's some amazing
products out there now that take care

of large sections of your accounting
compliance, which is awesome.

And it's really exciting to see where
that space is developing and how quickly.

In terms of some of the learning, I think
the first thing we learned, and it's

something we learned through implementing
Karbon as well, is that in order to

automate a process, you need to have only
one process, and that's something that

we had to work hard on across the firm
because we discovered that everyone kind

of liked to do things their own way and
so we needed to build in a discipline,

find the right workflow and optimize
process and get everyone on board.

It wasn't, an easy process, especially
as our first big automation project.

I think we've got a lot better at it now.

The other thing is that the software
is constantly evolving and the one thing

we really appreciated by going with
the subscription model and partnering

with, a really capable, robotic
process automation provider was that.

whenever a button moved or a login
pa, page changed, we needed to

update Rosie and how that worked,
and we needed to do it quite quickly

because everything would just stop.

So having a partner that we could
work with, meant that the bot

was kept moving quite quickly.

That happened a lot more
frequently than I appreciated.

'Cause as a human being, you
don't realize that buttons changed

slightly or the this works is

slightly different, but it's ongoing
as I think we all realize now.

The other thing I think we learned is
that using Copilot or ChatGPT develop

your data flows, is really useful.

So using, almost using
a bot to create a bot.

So using that sort of AI
technology to simplify a process.

So we've had process documents uploaded
into Copilot, and we say, how can we

do this in a more streamlined manner?

Or can you suggest other ways
we could approach this task?

That can be a really good way of
simplifying a process that has

a lot of human interaction going
on, because we think that the only

way to do it is the way we do it.

But that's not the case at all.

Exactly.

Abs- absolutely, and I, it's what I keep
saying to the team as well is if you're

trying to learn something, don't think you
need to go and find a way of learning it.

Just ask it how to learn it.

If you're trying to do

something

in

Copilot, ask it how you would
do that in Copilot and yeah.

It's, it's amazing how far
you can get with that approach.

Yeah.

It's so unnatural for us humans to the
system that we're trying to work with, how

to solve something and not look elsewhere.

So for those who maybe don't know, the
+MORE Group is a network of firms and

is also part of MGI Worldwide, which
is a top 20 international network

of accounting and advisory firms.

So Megan, I'm curious to know, what
have you learned about AI from other

firms, and why do you think that
it's important to have this exchange

of ideas amongst other firms?

We are currently running AI and tech
committees within our New Zealand

group of firms, and we have a, a
separate AI and tech committee within

the MGI Australasia network of firms.

And I do feel like having these
groups where we can freely share

ideas is a bit of a superpower for us.

But given the pace of change and the
variety of skills and the interest

levels across different firms
and across both networks, it means

we've, we've, we've been able to
investigate and trial different term,

different tools in different ways.

Some of our smaller firms have
volunteered to be guinea pigs, trialing

certain things that if we trialed
it in the larger firms potentially

could be a little bit disastrous.

And they've been able to give us
feedback so that we can improve it

before we roll it out to the wider group.

But also just, I, I think the
committees have been great.

And committee sounds
like a very grand term.

'Cause really all we're doing,
we're doing a lot of sharing

of ideas and chitchat, and.

of playing around with
various agents and AI tools.

So it's actually a lot
more fun than it sounds.

But I think what's been great
is where people lack confidence.

We can help bring them confidence or
where you're lacking an idea, someone

in the group will have a better idea.

It's just that sharing of skill
sets and sharing of, ideas that

has been so valuable for us.

We've also got, it's been quite
interesting in that across the group

we don't have any sort of, requirement
that people follow a specific tech stack.

So that said, we do encourage people
not to go too, sort of off piece.

But ultimately, we're all independently
owned firms, so if they choose to

do something that's kind of, that
is their, prerogative to do that.

And we can get some really great
feedback from some of those examples.

What has been quite interesting is
across the two networks, we have a

group of firms that are, have gone
with Copilot, they have full Copilot

licenses for the whole team, a number
of other offices have gone with ChatGPT,

and we've had some really good debates
over, which is better, which, which one

is the, the one to back for the future.

And, I think it's really powerful
to get those experiences.

It doesn't mean that our either
is right or either is wrong.

It just means that we are, we have a
lot more information available in order

to sort of back the choices and even
if perhaps ChatGPT does something that

Copilot doesn't do, it doesn't mean we
just suddenly all change, like kneejerk

reactions is not the way forward,
in the current environment, I feel.

But having the knowledge is really
important, so that you do know if there

is a reason to change, especially, there's
quite a lot of, I wanna say politics,

but there's quite a lot happening
behind the scenes in the open AI space.

So sort of understanding what
that means for you as a firm and

as a group is quite important.

So we've had some really good chats,
some quite robust conversations and

little competitions to see which tool
works better in different circumstances,

which I think are quite, quite cool.

Quite fun.

The other thing which has been
quite interesting as a group is we

reached out to our, IT support, our
IT service provider to get reporting

on what apps were being used.

and we did it as a, just really
interested, thinking that, you know,

in the Copilot offices, everyone would
be using Copilot and in the ChatGPT

offices they would be using that.

And instead we found that there
were a wide array of different

tools being used across all the
offices that we didn't realize.

Some of them, it shows the data being
uploaded and downloaded so you could see

where information was being uploaded and
tools that were otherwise not secure.

It's led us to double down, not only
on our training, but also on our

feedback requirement that if you are
using something else, that's, that's

okay as long as you're using it
with the right guardrails in place.

But you do need to feedback to
us in terms of what you're doing.

And the more information we can
gather across the group in terms

of what people are working with and
how they're using it is I think,

really valuable for the future.

I don't think we would've thought of
that, we would've wouldn't have considered

that as a concept if we were just sitting
in Auckland doing our Auckland thing.

So I think that's been
really, really cool.

The learning and the speed at which
you're, you know, kind of staying current

with us from rapidly evolving space,
plus the agility, that you're, you

know, you've got inside of your group.

I can see how there's so much value
in, in that collaboration you've

got with those other firm owners.

you mentioned Tech Stack, and so
as a lead in then I, I know that

you and your team Megan, are using
AI inside of some of your tools.

So not just with Copilot, but also with,
you know, Syft and Aider and Karbon AI.

And overall, you've reported that
AI has resulted in some really

significant improvements in both
efficiency and service delivery.

And so if we think more about the, on
the +MORE Group's values of impact

and agility, togetherness, empathy,
and personal fulfillment, how has AI

helped you live up to those values?

it's an interesting question because I
think the pace of change, relating to

AI helps us, but it also provides us
with quite a lot of challenges in terms

of living into those values as a firm.

from a client delivery perspective,
winning back time and having delivered

process improvements mean we carve out
more time to spend with our clients.

Spend more time on the phone.

We can ask questions, understand their
businesses and the challenges, and

deliver better information to them, which
lives into our empathy and agility and

impact in terms of our client delivery.

But at the same time, we have to be really
mindful of retaining that human element.

There are ways that we can automate
all sorts of moments that are.

Sort of opportunities for a touch
point with our clients and a chance

to find out what's going on with them.

So for example, we can automate an
amazing set of management reports out

of Syft, and we can send them to the
client with a bunch of insights that

are AI generated and we haven't had to
do any sort of real work within that.

But I think that's really obvious to
our clients and it's a really lost

opportunity to give them a call and say,
hey, I've, done these management reports

and I can see a bunch of stuff on there.

You might wanna chat about, you
know, do you wanna come in or

do you wanna go for a coffee?

So, need to sort of be mindful
that it doesn't, we don't become

just a, a churning manufacturer
of financial information.

And that's, it can be tempting
when there's so many clients

that need so much help.

But yeah, it's really important to us and
it's something we talk about regularly.

From a team perspective, I would
say that rather than AI helping us

live up to our values, I would say
our values have helped us remain on

course and how we harness AI and tech.

So trying to maintain our change
management program so that the team

don't feel either left behind or
overwhelmed, or that they feel really

excited about what we're doing.

So we talk a lot about impact and agility
and empathy where people might be feeling

a little bit uncomfortable with the pace
of change, and I think that's become

more and more of an issue even in the
last year, that the pace of change and

the number of options we have available
to us Being that our internal strength

as a firm becomes really important.

We cannot leave our people
behind through this process.

Yeah, that's excellent.

It's good to draw attention to that
because with things changing so

quickly, you can, you know, have,
have some of your team who's just not

comfortable or isn't keeping up and
you have to be really aware of that.

So love how, you know, you've got that
flip on where the values have played

into your AI adoption in your firm.

Thank you so much, Megan.

It's been such a pleasure to chat with you
and to see you again and, appreciate your

time and for joining this episode today.

For our audience, you've been listening to
the 2025 Karbon Excellence Award podcast.

Every episode we ask Karbon
Excellence Award winners, a series

of questions on how they raise the
bar for the accounting profession.

For more advice and strategies on how
to grow your firm, visit the Karbon

magazine@Karbonhq.com slash resources.

That's it for today, folks.

Bye for now.