The Accounting Technology Lab

Brian F. Tankersley, CPA.CITP, CGMA: Welcome
to the Accounting Technology

Lab, brought to you by CPA
Practice Advisor, with your

hosts Randy Johnston and Brian
Tankersley. What I was gonna

what I was gonna ask, I guess,
is that you know, Sasha, you've

introduced a lot of things, and
this whole paradigm shift is

really intimidating to a lot of
us that have been in this

profession for a long time, you
know, how do you know? Short of

taking the ice bucket challenge
and just jumping in with both

feet, how do you get started
down this journey effectively

without you know without
completely shutting the door on

all the stuff you've built?

Sasha Orloff: I mean, I think
about this stuff all the time as

it comes to like our sales
process and our optimization

process. Yeah, I would say so.
If you're listening, like if

your listeners are there, and so
I guess if we have a bunch of

firms that use Puzzle that start
taking the last week of the

month, you might have more
competition on your podcast,

because people are going to want
to figure out what to do with

the last week of every month.
No, Brian, back to your point.

Yeah, I think the natural
tendency I think for people that

are new to AI is they think
about they give it a task that

is kind of frivolous or not
really you know helpful, and

then they go, oh, like that's
kind of cute. I think give it

something really hard and
obscure to see. There, like

think about what in your last
month of closing process for

your clients, what was something
complex that just took you a lot

of time. Now, is it going to get
it right out of the box?

Probably not, because you're
not. Most people aren't used to

describing their problem in an
articulate way with all of the

different steps, and I use the
analogy of like telling my kids

to go clean their room, and I
say go clean the room, and I

know what that means, but like
what are the chances that they

go and they clean the room to
the same level that I thought

about? And so when you think
about how do you articulate to

AI, I have to say, go into your
room, scan the room. Take your

bed, move your sheets and your
duvet, cover it over. Put the

pillow at the top so it's
perpendicular to your bed at the

top. Take your clothes, take the
shirts, fold up all the shirts,

put them in the shirt drawer.
Take the underwear, put them in

the underwear drawer, socks and
sock drawer. Or I could come in

and say, hey, just if I say
clean your floor and make your

bed, and I open up the closet
door and everything comes

tumbling out. Well, they said
you said clean the floor, right?

So I think a lot of you point
me. So take something that is

hard and just put it into AI.
Don't use any client

information. Just like try
something obscure. Take your own

personal data. Take something
that's low risk, but something

hard and put it in, and learn
how to really articulate your

judgment in a way that is just
like you would explain to a

brand new super smart person who
doesn't know anything about

accounting. Try that. That's
what I say now in Puzzle, which

not here to mention like a sales
purpose again, because

everything is opt out. You can
go build agents, and we put a

bunch of safety guards in there,
so nothing ever changes in your

ledger unless you type in YES.
Now, of course, you can always

undo it and like fix it because
it's still ledger. Like you can

click around, but we make it so
that you can just throw all

sorts of crazy stuff. You can
choose from lots of templates

that we build that are generated
on skills and do it now. If you

don't want to use puzzle because
that's like too much, like I

don't want to switch ledgers and
do something, you'll notice that

one of the limitations of the
industry, because of how things

were architected 30 years ago,
is the API or the interface

between softwares, how they talk
to each other, is a CRUD API.

You can create, read, update,
and delete transactions or

journal entries. Now that is
helpful. You know that's

interesting. Let's take out some
transactions. Let's do something

with them, and let's change
them. Obviously, accounting is

way more complex than just
changing a categorization of a

transaction or something. But
you know you can take around and

you can try individual things.
Just take a transaction. You can

put just a transaction because
you can literally copy and paste

it into ChatGPT or Gemini or
Claude or Anthropic, and just

say, change. Here's a chart of
accounts. Just anonymize a chart

of accounts. Upload a
transaction system. Say, hey,

transaction description came in
like this. How would you

categorize this?

That's a very low risk way, a
low touch way to start playing

with it now, of course, that's
frivolous and not that

interesting. But like, if you're
just like, I've never done this

before, copy a transaction
description, copy your chart of

accounts, paste into an AI,
scrub for any personal

information about your clients,
and just say, hey, what category

do you think this would? What
category do you think this

should go to? Again, so listing.
Yeah.

Brian F. Tankersley, CPA.CITP, CGMA: So
what I hear you saying is that

you're going to generate
prompts, and a prompt in your

mind is really analogous to
almost a checklist. And you're

going to set up a you're going
to and so that's how you're

going to define this is how you
clean your room or this is how

you this is how you reconcile a
bank statement and so forth.

There, okay, well, good. I think
you've I think clarifying. These

terms, because there are a lot
of terms like agents and prompts

and things like that, and you
know when a lot of us enjoyed

the Matrix movies, and you know
when we see agent, we see Agent

Smith, and he's the bad guy, and
so it's helpful to have that

perspective from you about this,
because again, you remember

you're setting it up where nope,
nothing happens unless you say

okay, and I think that's the
important thing that we have to

remember with this: is that you
are still the boss.

Randy Johnston: And Brian, you
heard him say earlier he has the

AI pill from Silicon Valley. So
I think in our time together

that you know we've discovered
that autonomous means that the

books can close themselves, and
you find the mistakes later. But

the alternative is, you know,
the person deciding before it

posts through these agents, and
the profession really doesn't

need AI that removes the
accountant; needs AI that makes

the accountant better and the
ledger able to close itself. So

I think that's really nifty
because you know. Bottom line is

AI can't be accountable by
itself. So, Sasha, any parting

thoughts for our listeners today
that we that you think is really

important that we know?

Sasha Orloff: I think that
everybody should start exploring

and testing with options
themselves in the market, you

should constantly be evaluating
how do I get better, because

other people are, and if your
firm isn't or you aren't as

maybe the leader of your firm,
you're going to set the culture

change that it is okay to play
with this again in the safety

way. Like, don't just shove all
of your client information into

some random AI tool, like you
said, but like you should play

with this. You can't. You can
play with this. Give some

guardrails of like some baby
steps that people can. I do

fundamentally believe that firms
that adopt and start learning

are going to be the ones that
are going to make way more money

than ever before, and we're
seeing it in almost every other

profession that has adopted AI,
from lawyers to engineers, other

professional knowledge work,
they're making more money and

they're delivering a better
product to their colleagues, to

their companies, to their
clients, and it is not too late.

This level of automation we've
been hearing about in the news,

we read about on Twitter the
promise of productivity in

accounting. I would argue, I
think fairly credibly, it's only

really started this year. It's
been somewhat novelty. It's been

somewhat problematic last year.
I would say this year, right

now, we're finally starting to
see the actual productivity with

actual case studies and actual
implementation. That is a

lesson. It is not too late. Just
get started, and have fun and

enjoy it. If done the right way,
it doesn't have to be Agent

Smith. It can be Q from James
Bond. It can be the helpful

technologist that is here to
make you the best version of

yourself.

Randy Johnston: Beautiful. So,
Brian, any other thoughts,

questions?

Brian F. Tankersley, CPA.CITP, CGMA: You
know, I've been reflecting on

this, and it reminds me of a
podcast I listened to recently

from John Kiriaku. That's a it
was a CIA analyst that he was

describing the experience of 911
in the CIA offices, and

everybody was looking up at the
towers in flames, and finally at

the end, somebody at the back
yelled, "Will somebody please

lead? Okay, and then action
started happening. Okay, and I

think that's the challenge to
our profession: is you know you

may be comfortable, you know you
may have started like I did in

the days when people smoked at
their desks and had a 10 key on

their desk in a five 714 column
paper, but I got to tell you,

those people that refuse to
change from that world, they're

in the distinct minority, and
they're not, and they're

struggling to make a living now.
Okay, so I think it's a matter

of it's a matter of of you
getting out of your comfort zone

and providing the leadership in
these uncertain times with this

evolving with these evolving
things here, you know you don't

have to buy any particular
solution, but you need to get

your hands dirty, okay? Because
the problem here is that you

know if you don't know anything
about cars and you walk into a

car dealership, they're going to
have their way with your bank

account, okay? The thing you've
got to do here is you've got to

get you've got to get enough so
that you can develop the filter

of this works, this doesn't, and
you can lead. Okay, and so I

would just say you know for the
love of God, you don't have to

do what we tell you to do, but
please do something. Okay,

that's how I'd say it.

Randy Johnston: So, Brian, thank
you for that. And I was

thinking, with Sasha's taste,
maybe you should have talked

about like carbon fiber bikes
rather than cars. But I don't

know what he's got because we
didn't talk about that. However,

Sasha, your time is valuable. We
thank you, as we appreciate you

being here as a our guest, and
we. Wish you continued and

greatest success. So, until you
are able to join us in another

accounting technology lab,
Brian, Sasha, and I all wish you

a very good day.

Brian F. Tankersley, CPA.CITP, CGMA: Thanks,
guys. Thank you for sharing your

time with us. We'll be back next
Saturday with a new episode of

the Technology Lab from CPA
Practice Advisor. Have a great

week.