Summer snapshot of AI in accounting: A conversation with Roman Kepczyk

All right.

All right, welcome back. Everyone to another

episode of the Spark A I show. We're in the

middle of sort of transitioning and, and

transforming the podcast. And so we're, we're

gonna look a little different today than

maybe you you expect if you're a regular

listener.

So we're gonna be moving forward, we're gonna

be doing more interview style. So we're gonna

be bringing on some of the, some of the folks

within the industry who who are involved

in A I and technology across the space. And

so today I've actually got Roman Kch with

me. Roman is the director of Firm technology

strategy with right works and we're excited

to have him on board today. So Roman, hey,

how's it going?

Doing? Awesome TJ. Thank you very much for

having me on board.

Absolutely. So, Roman, I know you've been

involved in a lot of events lately. And that's

really what I want to talk to you about because

I know that this sort of the summer event

season is a really good time to, to get exposed

to a lot of the, the sort of what's going

on in accounting. Before we jump in though,

maybe just a quick introduction. Tell me,

tell me a little bit more about sort of, what's

your day to day look like?

Oh, sure. Pretty much as firm technology

strategy. My role is being an outsourced

it, partner to accounting firms, ranging

anywhere from 20 members up to 2000, I've

actually worked with about 30 of the top

100.

Basically going in and really looking at

their adoption of technology, looking at

the workflows that they've implemented for

tax and audit production. And now we're doing

a lot of work on CAS, believe it or not.

Which, you know, has a lot of a, I built

into it as well.

Yeah, that's great. So, yeah, speaking of

a, I, I know that you had a chance to attend

several events over the last several weeks

and I know you're, you know, involved with,

firm consulting and things like that. So

you're, you're really well exposed to sort

of what's going on, in the world of accounting

technology. And, I would love to, you know,

I know, I, I heard that you were at engage,

I know you were at CPA F MA a week or two

ago.

Maybe just a quick high level overview of

some of the events that you've attended recently.

And I guess the, the poke into that is just

sort of how, how much is A I taking over

the agenda at those events?

Well, and it's funny like we were at BDO

the Midwest conference yesterday in Indianapolis

and on all through the programs you mentioned

pretty much A I is by far the number one

topic, different variations on how different

departments utilize the tool. What is practical,

what is hype out there?

There's a lot of speakers that are coming

from the outside who have expertise in specific

components of it.

But overall, we're all in the same place

just trying to figure out what is the practical

tools that are happening and using it for

A I?

Yeah, so you mentioned that sort of practical

versus hype and I would love to, you know,

maybe just take a minute, you know, just

real talk for a moment, right? So what, what

are we seeing in terms of, of real practical?

We had this conversation just internally

yesterday with our team, you know, somebody

was asking us like, what are we seeing in

terms of real like adoption with some of

these tools, you know, are there, are there

products out there who are actually generating

revenue on this or are we still just seeing?

Is it still a lot of in the investment and

exploring the capabilities and really still

trying to find where the capabilities meet

up with the problem point? So I guess, yeah,

give me a little insight into sort of how

you're seeing that, that hype versus practical

application right now, what I see from the

big guys, you know, the marriage of players

like Thompson cch into it is they're bolting

on applications, they're in the phase of

trying to roll out something and talk about

it, but they really, I, I haven't seen great

delivery beyond the things that we talked

about in the past, which is like improved

research tools, you know, for tax.

It's really kind of limited at this point.

Where we've seen the more interesting innovations

is from the stand alone products, products

like Iwm Black or the development of the

companies like I'll say bot keeper who are

adding components into it, canopy or carbon,

those, they're adding specific components

to make it easy to use. I'll say the, you

know, generative A I solutions like chat

GP T to basically generate emails and things

like that. But it's not, like I said, it's

not major changes in time savings is what

we're seeing for those.

It's just, it's a streamlining and it will,

it will happen, it will get much more effective

in the long term, but for right now it's

early rollouts.

We are seeing some really interesting discussions

is like, you know, I mentioned Iwin, it's

spelled weird A Iwin.

We've seen some firms say that they've used

that to automate their payables process or

different components of that. And they've

seen significant savings because it's like

I said, saving a lot of time.

So, I expect to see, you know, similar tools

to that being adopted by the Big guys and

then rolled out probably in the fall conference

season.

Yeah. So it sounds like, I mean, that's kind

of the feedback that I gave on to our team.

Is there still, we, I'm seeing a lot more

of the you know, I was at a firm growth forum

out in San Diego a few weeks back and I saw

a number of new companies that I hadn't heard

of before, you know, in the, in the A I space.

And there really is sort of this difference

between you know, so I'm, I'm here representing

spark and we're sort of a broad based tool

very much a chat G BT but, but sort of more

built for the accounting firm. So you've

got the broad based tools that allow you

to access A I generative A I sort of broadly.

And then you've got, you know, some of the

things you were just mentioning are more

leveraging gen A I and machine learning to

solve for a very specific problem, right?

Whether it's practice management or whether

it's tax prep or those kinds of things with

Ian and Black or tell me what, what you're

seeing in, in the space as it relates to

the difference between sort of that broad

base versus the point solution.

Ok. Well, when you talk about the broad base,

you know, chat GP T, you know, we see like

B digital having a chat GP T version that

they provide to the BDO firms, but not for

the alliance firms. And so, what we think

it is going to happen is on the GP T side

is there will be a development of private

GP TS which when you think about spark, it

really is a private GP T.

It's, we've created internally. We allow

people to use it and it's a safe place to

go in and play with the system, understand

how to use it.

To learn to memorize prompts. I mean, that's

one of the big discussions we have with every

group is how do I have useful prompts that

are repeatable and all that. And that's what

I see. Like the private GP TS such as, you

know, like I said, our spark doing that.

I anticipate firms will develop their own

private GP TS eventually for their specific

niches.

You'll recall at our right now conference,

Radika Dirks talked about, you know, 80%

of what we CPA S do is obsolete. Now, the

other 20% the unique thing is what we got

to figure out. Well, firms have niches, whether

it's auto dealerships, whether it's restaurants,

whether it's medical profession. And if they

can take their captured knowledge, basically

their intellectual capital that they've developed

their spreadsheets, their questionnaires

and all that build it into a private GP T

like spark, then they can query that in a

very safe environment and they can build

on it and actually rely on the results. So

we see that on the GP T side, I think the

part that kind of gets unnoticed is the components

that are the early components such as robotic

process, automation, machine learning, what

we call, you know, augmented intelligence

types tools. And I see these generative tools

making those much more effective.

And so, we're in, like I said, some firms,

it's a little bit complex to use those tools

and we have to develop specific experts. Think

about like Power Bi I, you know, it's, it's

a hard program to learn. Once you learn it,

it's great take special skill. But I think

the generative A tools will allow us to query

and make requests in those, you know, everything

from Power B I to excel, which wes at to

teams especially and actually get great value

out of it.

On the other end, yeah, when you talk about

Power Bi, I, I think one of the benefits

we're seeing even now, you know, we talk

a lot about crawl, walk, run, right.

So there's this idea of like the crawl is,

you know, maybe, maybe right now we're not

connected into Power Bi I, but you can leverage

the power of the models to even just ask

it. Like I'm trying to do this in Power Bi

I, how can I think about this? You know, and

it can really help you frame the problem

and then walk you through the solution. How

do you build out the dashboard or the whatever?

So you, you, you know, another theme that

we have is expanding your expertise. So I

am not a power bi expert, but I can actually

do a lot more in those tools than I ever

thought possible because I just take it to

spark. And I say, here's what I'm trying

to accomplish, help me like walk me through

the steps to build this out in a way that

makes sense, you know, and it just, it does,

it just walks you through the process and

what's cool is you can capture that and give

it to, you know, other team members.

So now they know how to do it because they've

got a live example and they have a place

to put in their query into it, using your

structure that you just taught them on and

it works the same way. So the other thing

that I think is really cool that we see when

you said practical, the first thing that

came to mind is the transcription components

of it.

You know, most of us use teams and zoom. But

what's really great is, is you have tools

like auto A I fireflies, you know, teams

does have a product that does transcription

but it's not as concise.

We use Clari copilot at right works. And

the quality is amazing on how it can summarize

a meeting, how it can give you the action

steps at the very end.

Even can help you catch up and do assignments

for those projects.

Yeah. Yeah. Those tools are making a lot

of progress. I was actually in a webinar

yesterday with the, and it was the, the VP

Summer snapshot of AI in accounting: A conversation with Roman Kepczyk
Broadcast by