Episode Transcript
Transcripts are displayed as originally observed. Some content, including advertisements may have changed.
Use Ctrl + F to search
0:30
So we are continuing our
0:33
using AI month,
0:36
right? Last time we talked
0:38
about Intuit Assist and. Talking
0:40
a little bit about that sort of thing. So we're continuing
0:42
that, that series this month
0:45
with using AI and chat GPT
0:47
for your accounting practice. We're glad you joined us
0:49
for the QB power hour and
0:53
Michelle. Introduce yourself
0:56
as if we need no introduction.
0:58
Welcome everybody. My name is Michelle Long.
1:01
I am a CPA and the owner of Long for Success,
1:03
a trainer for Intuit for a very
1:05
long time, contract trainer, and author of several
1:07
different books. And invite
1:09
you to join us in the Facebook group and continue
1:12
the discussion. And currently
1:14
enjoying having lots of buzz
1:16
around Kansas City Chiefs with
1:19
Taylor Swift in the house
1:21
recently. And just enjoying
1:23
watching that little love story going
1:25
on. It's just fun gossip going
1:27
on and stuff. We've been enjoying the
1:29
little romance in the air
1:32
around our stadium. Let's put it that way.
1:35
As if you need, as if you needed more
1:37
information or more stuff floating
1:39
around with the Kansas City Chiefs.
1:42
I know. Everybody. Yeah.
1:46
It's fun. Anyway,
1:49
enough of that.
1:50
Go ahead, Dan. All
1:52
right. My name is Anda Long, owner of Dan Witt, worked
1:55
at Intuit for nearly 18 years,
1:57
co hosting today over as well as the
2:00
Workshop Wednesdays over at schoolbookkeeping.
2:02
com tech editing duties
2:05
for the QBO for Dummies series. So trying
2:07
to keep them honest and actually we're going
2:09
to have the author of the QBO
2:11
for Dummies series come on
2:13
the Power Hour next month. He's
2:15
actually come up with a way he's got some,
2:17
if you've never read the book he's
2:20
got some really powerful chapters on
2:22
Excel, and he's actually come
2:24
up with a way to be able to download
2:27
QuickBooks online data into
2:30
Excel and have it usable without
2:32
having a desktop version of QuickBooks interesting
2:35
Excel hack when it comes to
2:37
when it comes to QuickBooks, so we'll have him come on.
2:40
As well. Now, has Michelle
2:42
joined or has has Heather joined us or
2:44
is he still at a prior meeting?
2:47
I sent her the link and
2:49
we are expecting Heather to join
2:51
us. So Dan, you go ahead with the in
2:54
the intro and I will continue
2:56
trying to reach out and
2:58
get Heather.
3:00
Alright, Heather is is obviously busy.
3:02
A busy individual.
3:05
He Tina, also doing the the
3:07
happy hour with Liz and Heather.
3:10
She has, in the last year, become
3:12
the director of education and media over at
3:14
Woodard. So this is
3:16
this is obviously taking taking her in,
3:18
in, in different directions and we
3:21
were hoping that she would join us as well
3:23
because Michelle, you had done a
3:25
a session at Scaling New Heights with
3:27
Heather, correct?
3:29
That's correct. Yeah, Heather and I did
3:31
this together. It was. A version of this intro
3:34
version of this and it was very widely
3:36
received. It was standing room only. This
3:38
is just very popular with so many people
3:41
trying to learn about this because it
3:43
is, what's going on right now. And
3:45
Heather and I have done it together a couple
3:47
of times now. Anyway we are getting
3:49
Heather. She will be here in a little bit, but Heather
3:52
is just an amazing individual, very
3:54
up to speed on this stuff, and we
3:56
will have Heather joining us in just a little bit,
3:58
but she did an amazing job with
4:01
scaling new heights, and as a reminder,
4:04
Scaling New Heights Call for Papers
4:06
is due September 30th, so if you'd
4:08
like to speak at next year's Scaling New Heights,
4:11
get your topic submissions in by
4:13
September 30th, and we'd love to
4:15
see you
4:15
there. All
4:17
right, and so hopefully she'll be
4:19
rolling in on two wheels as
4:21
soon as she's she's available. She did say
4:23
that she had just wrapping up a
4:25
meeting right before we launched here
4:28
today hopefully she'll be on time. Finding
4:30
the link to join us and and hearing
4:32
what you, what we have to say. So a little bit of the details
4:34
of the QB Power Hour it's
4:36
every other Tuesday at noon Eastern our
4:39
upcoming webinars. You can always check the website
4:41
for upcoming events, as well
4:44
as topics that we will be discussing
4:46
on the QB Power Hour. We're not eligible
4:48
for CPE. That was Hector's thing
4:50
and he's been gone for a while. But.
4:54
You can always go to the qbpowerhour.
4:56
com slash resources website
4:59
to download PDFs, recordings
5:02
of the podcast, as well as other resources
5:04
as well. And in the chat, I put
5:06
the link for today's session
5:09
as well as the download
5:12
there. If you have specific questions
5:14
about. What we're going to be talking about
5:16
here today, please put that in the Q&
5:19
A. There should be a little button at the bottom
5:21
of the Zoom webinar for the Q& A.
5:23
And then the if you have general
5:26
comments please put that in
5:28
the chat. But if you have a question about something
5:30
that we're talking about during the webinar,
5:32
please put it in the Q& A so that we can see
5:34
it, address it, and and answer
5:36
it. Speaking of Heather, there
5:39
she is! Yay!
5:42
I'm gonna make her a host so we can see
5:44
and hear her.
5:46
Oh, thank goodness.
5:52
My outlook was being very unkind
5:54
to me because
5:57
it sent your email, Dan, directly
5:59
to my deleted folder. So I'm checking
6:02
all the things I'm like, could it
6:04
possibly be in the deleted folder? And
6:07
there must be some patient that it's not working
6:09
properly.
6:09
While Heather's working on her
6:12
video and getting that
6:14
going, I'm going to go ahead and go over the agenda.
6:16
Dan, did you do the housekeeping side? Yes,
6:20
I did. I did. Okay. So for
6:22
our agenda today, we're going to review
6:24
a little bit of the chat GPP and AI stuff
6:27
that we've talked about briefly, or
6:29
previously, and we're going to tell you where you can access
6:31
some of that. And then we're going to give you some
6:33
examples of how you can begin
6:35
using AI and chat GPP
6:38
and all this stuff. How you all can begin
6:40
using that in your firm today. Heather
6:42
and I are going to share some very concrete examples
6:45
and give you some things that you can start doing
6:47
today. So you can benefit from this
6:49
today. We'll talk about for example,
6:52
using it with financial statements and how
6:54
you can start giving them some advisory
6:56
type of things in there. Heather's going to share
6:58
social media things with you. I'll share
7:00
with you how you can train your own chat bot if you want
7:02
to. And we'll also discuss the impact
7:05
on the accounting industry. So
7:07
that's where we're going today. Heather, you
7:09
still can't get that webcam working, huh?
7:11
Nope.
7:12
So I'm just gonna sit back here and show that
7:17
you can.
7:18
Oh, look at that. Yay. I'm gonna fall. I'm gonna fall on my
7:20
sword. It was actually me. Because
7:23
the pop up was on my screen
7:25
on the other window, and I was staring intently at
7:27
you, and I
7:29
was like, what, why can't you do this, and
7:31
it's over here saying, would you like to make
7:33
Heather a co host, and that's what
7:35
was the problem.
7:36
I am now a co host, so thank
7:38
you for
7:39
that. And, yeah.
7:43
Okay, so I'm going to launch the first poll,
7:46
and then I'll stop sharing so that you too
7:49
can start to take. Take
7:51
over here. So let's go ahead. We're going to launch
7:54
it. How have you used chat gbt
7:56
being clawed or others so
7:59
far? We just want to take a little litmus
8:01
test as the Where you are
8:03
today, and somebody asked me a question about
8:05
the prior session that we did.
8:07
We will have it in the there is a link
8:10
inside the handout. So if you do download
8:12
the handouts, you can get to that link,
8:15
but also just searching AI in
8:17
the QB Power Hour website, we'll
8:20
bring up that prior session
8:22
that we had last time.
8:25
There you go. And we had a question
8:27
in there from Nancy about is there
8:29
an AI tool that will read
8:31
multiple web articles? Oh,
8:33
no, this is one
8:35
to restart. And I didn't want it to. Sorry.
8:38
I wanted to know about, wanting to
8:40
get an idea on if there's
8:42
a AI tool that will read multiple web
8:44
articles and consolidate them into
8:47
one single new article that addresses
8:49
all various points. So
8:51
I think she's wanting to stay up to date
8:53
on what's new, it sounds with
8:55
AI and things like that.
8:57
I can answer that, Michelle, because actually I just
9:00
did that
9:00
today. Okay, go
9:02
ahead, So I just
9:04
did that. Go ahead today, Nancy. And
9:06
actually chat. B t can do it even though
9:08
chat b t said it couldn't. And we'll talk about
9:10
that during the webinar when it actually lies
9:13
to us. there are there
9:15
are add-ins plugins
9:17
that you can add to your chat.
9:20
G P T. And the one
9:22
that I used this morning
9:24
to do and I did something a little
9:26
bit differently, but this premise is the same.
9:28
I was creating a digest of different articles that
9:31
actually I had written that I wanted to pull into a single
9:33
article. And the way I used
9:35
it was I had used the scraper.
9:38
Plugin in chat GPT. So it would actually
9:40
scrape from each of these websites.
9:43
And it does have a limitation. So there
9:45
are limitations to it. The other
9:47
way that you can do that, if you would
9:50
like Nancy that has worked well for me
9:52
is if you just copy and
9:54
paste the content from the article
9:56
into a text document or word document,
9:59
you can actually use Claude,
10:01
which Michelle's going to talk about
10:03
and actually share those files and it can
10:05
do it too. But As we'll talk
10:07
about later, you've got to make sure you're proofreading and that you know
10:09
the content, because it will lie.
10:13
That's correct. You do have
10:15
to know your stuff, because you
10:17
need to double check it, because it could
10:19
be inaccurate. Yeah, that's
10:22
exactly
10:23
right. My wife was just
10:25
on a call the other day and she,
10:28
about blog articles, and
10:30
they were talking about you can't
10:32
just rely on chat GPT
10:34
to spit out a blog article
10:37
for you because there's
10:39
metadata in there. So it knows the
10:41
Google will begin to, push
10:43
down the ones that are actually written
10:47
by machine learning.
10:48
And Dan, I recently read an article
10:50
that was saying that actually has been debunked.
10:54
Oh,
10:55
that it doesn't
10:55
know. And I actually
10:57
know a person against AI.
10:59
I know an individual that actually
11:01
wrote an article from scratch, like the old
11:03
fashioned way, and her
11:05
articles were flagged as being created by
11:07
AI. So it does. So
11:10
whether if they are running that engine, it's
11:12
not accurate. So it's, they
11:15
say, they may say they can do it, but they're not actually,
11:17
it's not a proven thing that they can do.
11:21
Interesting. That's good to know. And
11:24
I love you too, Nancy. Okay.
11:28
All right. Let's go ahead and get into this because we've got
11:30
so much to cover here. So first of all, let's
11:32
review some of what we did before. So the
11:34
QB power hour that Dan and I did back
11:37
on. August 1st covered
11:39
a lot of the basics that Heather and
11:41
I covered previously at Scaling
11:44
New Heights and on another webinar that we
11:46
did with CPA Academy and stuff.
11:48
So there's the link to that YouTube video
11:50
that you can watch where we covered some of the
11:53
intro stuff on what's a different
11:55
screen, chat, GPP 3.
11:58
what are plugins and all of
12:00
the basics and things like that. So you can watch
12:02
that through that link right there, using
12:04
AI as a tool in your practice. Some
12:06
of the new features that have come out recently
12:09
since that October 1st is the
12:11
ability to give it custom instructions.
12:13
In other words, I can tell it, I'm a
12:15
CPA, working with QuickBooks,
12:17
I'm a tax professional,
12:20
or you can tell it things about yourself.
12:22
Custom instructions, stick to the
12:24
facts, don't use any interpretations
12:27
or anything. So you can give it custom
12:29
instructions. So every time
12:31
you're talking to chat GPT, you have to use
12:34
those custom instructions. There's also
12:36
a code interpreter and Heather, you
12:38
use, a lot about code and things like
12:40
that, but you can actually use chat
12:42
GPT to help if you're doing coding
12:45
to write code or to debug
12:48
your code. And to help you when you're
12:50
doing coding and things like that. And I know Heather
12:52
does a lot of no code things. But there's
12:54
now a code interpreter within the chat.
12:56
G P. T. I've got links
12:58
in here for you. Barred is
13:01
with Google, so you can use the
13:03
Google version is barred.
13:05
There's quad to now.
13:07
The quad original one without,
13:10
but now quad two is out. And as Heather
13:12
mentioned ChatGPT pulled
13:14
Bing in early July, and when
13:16
it did that, there was a, in my opinion,
13:19
what that's worth, but
13:21
in my opinion, ChatGPT's
13:23
performance suffered greatly when they
13:25
pulled Bing out, and I didn't
13:27
like how ChatGPT, the
13:30
results that it was giving me, and I
13:32
started experimenting a lot more with
13:34
Bing, and I didn't like it all that great.
13:36
I just started using Bard recently,
13:38
but I started using Claude and
13:40
I learned about Claude from actually Hector. He
13:43
had mentioned it and I thought I'll check out Claude. So
13:45
anyway, I checked out Claude. I really
13:47
liked it, so I have been using
13:49
Claude a whole lot more
13:52
and now Claude two is out. It is free
13:54
as well. If you are in another country,
13:57
you may have to use A V P N to access
13:59
it. I know it's in the US and one other country.
14:02
So you may have to use a VPN to access it. Also,
14:05
Intuit Assist was just
14:07
announced. That is going to be AI that
14:09
is available within QuickBooks,
14:12
TurboTax, Mint not Mint,
14:14
MailChimp, I've
14:16
combined all that, MailChimp, and
14:19
Credit Karma. So Intuit is
14:21
putting their own AI within
14:24
all four of their products in their suite.
14:26
And Dan and I talked about that previously,
14:28
and there's a link to that one there. So
14:31
that's where you can go and watch some of our previous
14:33
discussion and get some of the history
14:35
there so you can catch up on some of that
14:38
if you need to. Okay? Heather
14:42
or Dan, do you guys want to add anything on some of this
14:44
review before we get into the details today?
14:46
I wanted to I wanted to ask Heather,
14:49
because we had talked a little bit yesterday, Michelle,
14:51
about the artificial
14:53
of the artificial intelligence and
14:56
I was using a analogy of,
14:58
we seem to be in the, if
15:01
we're talking about artificial sweeteners and
15:03
comparing it to artificial
15:05
intelligence, we're in the sweet
15:07
and low phase. Of artificial
15:09
intelligence, right? We're not at
15:11
the stevia point yet. Would you
15:14
agree that is like where we're at
15:16
with AI today?
15:19
Or would you say that, hey. We're
15:21
already at TV can't even tell in
15:23
some
15:24
cases.
15:26
Oh, yeah, no, I think that you can
15:28
tell I think you can definitely tell
15:30
the difference between the AI platforms
15:32
especially if you're using if you're using
15:34
them. Consistently, right?
15:37
And you're informed about what
15:39
it is that you're using it for. So that's
15:41
the big danger. And, Michelle, we just talked,
15:43
we've talked about this at length. That's the big danger
15:45
of these tools is that don't
15:48
ever assume that it's smarter
15:50
than you. And don't ever
15:52
assume that it's an expert at
15:54
something you're not an expert at. Because
15:57
what it does, the way that AI
15:59
works is especially
16:02
chat GPT, which is a generative
16:05
language. Something.
16:07
Help me with that. It's there, but I can't find
16:09
that last word there. Model. That's
16:11
the word I was looking for. What it does is
16:13
it predicts. They've fed it all
16:15
of this, all of these sources from
16:18
the English language from books, periodicals,
16:21
conversations call logs,
16:23
and it has trained it How
16:26
to respond based on predictive
16:28
analysis of what the next characters,
16:31
not sentences or words
16:33
or any context because you have to remember it's not
16:35
a person it doesn't have any logic
16:37
going on in its brain, but it is using
16:40
an algorithm. Right. That is
16:42
going to predict what the
16:44
response should be based on the patterns
16:46
and all of the content that it's been fed
16:49
during its training. So once you understand
16:51
the limitations of that,
16:54
then you can start to look at it in a way that you
16:56
should look at it, that it's not an expert.
16:58
It is a tool that is basically
17:00
using code underneath it to
17:02
predict a correct answer. But
17:05
those predictions can be wrong. And so if
17:07
you're asking about tax law and nothing about
17:09
tax law, do not take whatever it
17:11
says and copy and paste it somewhere for people
17:13
to rely on, because it could
17:15
be incorrect. You've got to
17:17
be confidently wrong. You
17:19
can be confidently wrong. And you
17:21
should assume that you're wrong. I think that's
17:24
the thing is that it's great for things like rephrasing
17:27
what you already know.
17:29
Right. Awesome.
17:32
If I'm
17:33
writing something like I need this to be more professional,
17:35
I can say, here's what I'm thinking. And it can
17:37
say, here's a really beautiful
17:39
way to say it. But I'm not going to ask
17:42
it how to perform brain surgery. Because
17:44
I don't know how to do that and it could
17:46
tell me something that isn't correct. Right,
17:49
right. Sorry about that. Ran. When we, I'll get
17:51
off my soapbox. In our prior
17:53
in, in our prior session, we had Michelle
17:55
really drove the point home, treat
17:57
it like an intern, right? Like you can
18:00
have a really
18:02
high performing intern but
18:05
you still want to. Proofread
18:07
their work and go over what it is.
18:10
Is that still your philosophy
18:12
Michelle and Heather, as far as where
18:15
it is today? Because, tomorrow or two
18:17
days from now we could have a different
18:19
opinion.
18:21
Yes. Yeah. Trust, but verify.
18:23
Trust but
18:24
verify. Trust but verify. I
18:26
love that, Michelle. And I think, Nancy's
18:28
asking the difference between cloud and other AI tools
18:30
is they do the same thing. They're just
18:33
got a different, maybe their algorithm behind
18:35
it is a little bit different. And
18:38
that's where To your sweet and
18:40
low analogy or Danology
18:43
we're going to start to see the leaders come
18:45
out is where we're going to start to see more
18:47
accuracy and the,
18:49
computer wizards are going to start to
18:51
introduce code that is more accurate
18:54
than others. And that's where we're going to start to see differentiation
18:56
within the
18:57
market. Right. Yeah
19:00
and somebody just posted out there that
19:02
they like to use it for to start IRS
19:04
abatement letters, and that's a great idea. We
19:07
talked about doing communications and
19:09
using it for that in a previous webinar,
19:11
and we gave you the link for that. But
19:13
I wanted to talk today, and we talked about this
19:15
briefly before, the financial statement analysis.
19:18
And so what you can do here is you
19:21
give it the client's financial statements for analysis
19:23
and recommendations. But what you want
19:25
to be extra careful of is you do
19:27
not give it any client confidential
19:30
information. Do not give it the client
19:32
name. Double check that there's no
19:35
identifying information in
19:37
the account name. Make
19:39
sure there's no client name
19:41
or anything that could be identified
19:44
in the name of the client. So
19:46
what we're going to do is we're going to copy the account
19:48
name and the numbers. No
19:51
identifying information there. Just the
19:53
names of the accounts and the, the months
19:56
and the numbers. So no identifying
19:58
information. So you're not disclosing
20:00
anything that could identify the client.
20:03
And then what we can do is we can
20:05
say, hey, give me a trend analysis
20:08
and the The growth and net profit margins
20:10
or if you were doing it on the balance sheet You can
20:12
look at the liquidity and leverage ratios
20:15
and stuff and then you can ask it for recommendations
20:17
for improvements so you can ask
20:19
it for Some advisory
20:21
type of recommendations that we then
20:24
can provide to our clients and
20:26
this is really super cool so because
20:28
cat gpt can be slow sometimes
20:31
i've already gone ahead and done this for
20:33
you So that I can just show you
20:35
instead of, waiting for it to generate
20:37
its response. So what I generally
20:40
do is I use a Word document
20:42
so that I can do things there and then copy
20:44
and paste. So what I started
20:46
with is I had Perfect here because
20:48
I like to use Prompt Perfect, a plugin
20:51
to help with your prompts. So
20:53
I'm saying I'm going to give you
20:56
the monthly finance, monthly P& L statement
20:59
of a company, no name,
21:01
perform a trend of financial aid analysis
21:04
of the key profitability ratios, the gross
21:06
and net profit margins, and provide insight
21:09
in three recommendations for improvements,
21:12
limit your response to 3, 800 characters,
21:14
and then I copy and paste this
21:16
Financial statement with, again, no
21:19
identifying information. Now,
21:21
the cool thing is, you can also,
21:23
with Quad, just upload
21:26
an Excel spreadsheet. But again,
21:28
make sure you delete the
21:31
name. If you're going to do
21:33
an Excel spreadsheet, delete the
21:35
row that contains the name. Okay.
21:38
Also, don't have your
21:40
spreadsheet named Craig's
21:43
Profit and Loss. You could just have it Company
21:45
P& L. Make sure that if you
21:48
are uploading that to Clod, you
21:50
can upload a document. All right.
21:52
So if you were doing that, make sure you don't have
21:54
any identifying information. But what
21:56
I've got, this is not a great one
21:58
because you've got a lot of zeros and it's a sample
22:01
company, so there's not great information,
22:03
but what you do then is you go put this
22:05
into chat GPT.
22:08
So up here, you can see I enabled
22:10
plugins. You need to enable your plugins
22:13
before you start your chat.
22:16
So I had enabled three plugins.
22:18
I enabled prompt perfect. As
22:21
well as Wolfram, because Wolfram
22:24
will do the math. Tag GPT alone
22:26
can't do math, so you want to enable that plug
22:28
in. We talked about those things in that
22:30
previous webinar. You can go back and listen
22:32
to the details on that. But I want to show
22:35
you, it shows you when it uses that
22:37
plug in. So it'll tell you that it used
22:39
that. And then, again, I
22:41
let it do this previously. So I'm going to
22:43
scroll down and look at our results here.
22:46
So let me just get this moving down and
22:49
you'll see I copy and pasted all those
22:51
numbers in there with no identifying information.
22:54
And then it says, okay, here's what we've
22:56
got. It gives you the calculations,
22:59
double check those to make sure
23:01
that it's calculated those correctly. Then
23:03
it gives me some insights and
23:05
it says, okay, there's some seasonality.
23:08
Here's some high expenses. We had negative
23:10
income. Here's some recommendations.
23:13
You can optimize your seasonal operations.
23:16
You need to look at some expense management,
23:19
investigate your negative income. Then
23:21
I said, okay, give me some additional recommendations
23:24
to improve our margins. It gave
23:26
me all of these additional recommendations.
23:29
This is where we're getting into some
23:31
advisory type of thing. Okay.
23:34
So I can give my clients some advice.
23:37
Okay, so I can do some advising
23:39
for my clients to help them improve their
23:41
operations And their profitability
23:44
so this can help you to provide
23:46
some advisory services to your
23:48
clients Help you get ideas
23:51
that you can then provide to your clients Now
23:53
the nice thing about this and the reason
23:55
we initially told it to do
23:57
it with 3800 characters was
24:00
so that you could copy and paste that
24:02
into your financials of QuickBooks,
24:05
because when you go into QuickBooks and
24:07
you click on your add notes up
24:09
there, if you go up there and you click on
24:11
add notes, you then could
24:13
copy and paste all of that right
24:15
here into the notes of the financial
24:17
statements. So you could give that
24:19
to the clients with that provided
24:22
now One of the things that I like to
24:24
do is say, provide
24:26
me with those advisory,
24:29
with those recommendations without the
24:31
Calculation so that those
24:33
calculations aren't taking up part of your
24:36
4000 characters But this
24:38
is where you can use this to help provide
24:41
You know advisory services or
24:43
some good advice to your clients
24:46
quick and easy without
24:48
taking a whole lot of time. Now, Hector's
24:50
right tool, he is integrating this
24:52
with his right tool, which is
24:54
another big time savings for you.
24:57
The other thing that I wanted to point out
24:59
to you, let me get back to
25:01
my slides if I can find them. I have too many
25:04
things open here. Let
25:06
me just tab through here because the
25:08
other thing that I wanted to point
25:10
out to you If I can find
25:12
my slides, here we go is so
25:15
this is again This is a great way to add
25:17
some differentiation to your services
25:19
for your clients. The other thing is use
25:22
digits. com To do
25:24
this and when instead of doing it the way
25:26
that I just showed you copying and pasted
25:28
into Either chat gbt
25:31
or claude or you could try being
25:33
or any of the other ones instead of doing
25:36
it that way And using
25:38
chat GPT and copying it
25:40
back into the nodes, you're using the right
25:42
tool that Hector has, you can
25:44
use Digits. com. Digits
25:47
link is a third party app
25:49
that will connect with QuickBooks Online
25:53
and Digits uploads all that information
25:55
to their Digits. com.
25:58
And if you look on the right of this screenshot
26:00
over here, They have
26:02
the A. I built into
26:05
their app. So it's in
26:07
a secure environment in the
26:09
digits dot com app. It's
26:12
got the A. I built into
26:14
their app and
26:16
you can do the stuff up there. You
26:18
can see they've got this client portal down
26:20
here for you to manage your clients and
26:23
you can run reports and you can do your AI
26:25
type of stuff over here and you
26:27
get five free companies
26:29
so you can do this type of analysis
26:32
and you can do some of these reports and stuff
26:34
here with Digits. com.
26:36
So go check out digits. com for
26:39
some of this financial analysis. It's
26:41
really cool. So
26:43
check that kind of stuff out. It is a great
26:46
way for you to up your game with your clients
26:49
using social media to get a jump
26:51
on these things. But remember what we said,
26:53
trust, but verified double check the calculations.
26:56
Make sure you use Wolfram. To
26:58
do the calculations in the first place
27:01
and you know read through those recommendations
27:04
to make sure they make sense Before
27:06
you just copy and paste them into the financials
27:08
But that's a great way to add value
27:10
for your client now heather
27:13
has some great stuff there
27:15
On social media. So heather, do you
27:17
want to share your screen and share this
27:19
with them? You're muted
27:24
And i'm muted
27:26
Yes, I'm going to share my screen. I'm going to share
27:29
some of my secret
27:31
things that I do as
27:34
the director of education media at Woodard.
27:37
So I use this
27:39
in a couple of ways. So I use chat
27:41
GPT to create
27:45
social media posts. So what I can do
27:47
is I can ask it to create a social media
27:49
post based on information
27:52
that I give it. I can and the way that
27:54
I do that, let me see if I can find my social media
27:56
post generator. So right here
27:59
in my social media post generator, I started
28:01
by asking it to create three
28:04
social media posts. Right
28:07
for a particular article and I
28:10
asked it for one for Facebook, Twitter
28:12
and LinkedIn. I asked it to use emojis
28:14
and the hashtags that you
28:16
see here and I asked it to tag me.
28:19
So what I can do is I can
28:21
just feed it an article.
28:24
And have it create this.
28:26
So let's do one, let's go ahead
28:28
and I'm going to copy this. I'm going to create a new chat.
28:31
And this time, what I'm going to use is
28:33
I'm going to use plugins, just like
28:35
Michelle did use this prompt.
28:37
Perfect. And the scraper to create this.
28:40
So I'm going to go ahead and put in prompt prompt
28:43
or perfect. Sorry. And
28:45
then I'm going to paste what I just asked.
28:47
And then what I'm going to do is I'm going to go over
28:49
to the Woodard report. And
28:52
I'm going to grab a URL
28:54
of an article so we'll grab this URL
28:57
from Norm Axelman, and I'm
29:00
going to copy it, I'm going to paste
29:02
it right in here, and
29:05
then I'm going to click I'm going to go ahead and
29:08
ask. chat GPT to do its
29:10
thing. So the first thing that it's going
29:12
to do is it's going to use Scraper to go
29:14
to that article out on the web and
29:16
read it. Now, sometimes if the article
29:18
is too long, it'll give me an error
29:21
and say, Hey, I can't really read it. And then
29:23
I get a little stuck in the mud. If
29:25
it's just a regular 750
29:27
word article, it typically works really well,
29:30
but it's going to go through and it's going to take all of that
29:32
information. And then it's going
29:34
to create the social media posts. So
29:37
it went through and it found it. It's
29:41
adding the emojis. It
29:44
actually, I didn't tell it that Norman Axelman
29:46
was the author, did I? But it
29:48
was able to read that from the article.
29:51
And it's
29:52
actually... But Siri heard you and told Alexa
29:55
and then bopped down to... Correct.
29:58
Correct.
30:01
So a couple of things, because this is a
30:04
new one, my, in my established
30:06
social media generator chat with chat
30:08
GPT, it's learned
30:10
that tagging at each Saturday actually
30:13
creates that link here. It didn't
30:15
because this is a new chat. And one of the things about
30:17
chat GPT is if you continue the same
30:19
chat, it continues to learn from
30:22
you as you interact with it. So it
30:24
did it and went ahead and it created an individual.
30:28
Social media message based on
30:30
that. And if I were to click this, guess what happens? It
30:32
goes right out to it. So what I can do
30:34
now is I can actually come over here and
30:37
I can copy this
30:40
and I could go out to Facebook and I could
30:42
just paste it and it's going to paste the
30:44
link. It's going to pull up the picture of the article and
30:47
I'm going to hit. And now I
30:49
don't have to do this on my own. Now
30:51
I've actually supercharged this in the way
30:54
I've served anybody. One guess on how
30:56
I've supercharged this particular
30:57
workflow. You
31:01
got it, Dan.
31:03
Exactly. Right. So what I
31:05
can do is the, what I've actually created
31:08
is we use meet Edgar at. We
31:10
use bdigera at Woodard. It's
31:12
a tool I've used for a really long time. There's other great ones
31:14
out there. But what I've done is I'm
31:17
able to create his app that actually
31:19
watches the Woodard Report RSS
31:21
feed. So I use RSS by
31:23
Zapier, which is a built in tool
31:25
inside Zapier, and I
31:28
have told it to go ahead out to
31:30
the Water Reports RSS feed,
31:32
and I've asked it to watch it for when I post
31:35
a new article. Then
31:37
my next step is to go into
31:39
ChatGPT, and I could also use
31:41
the developer version of this, which is OpenAI, which
31:43
is their developer.
31:46
API version of this. And that's actually what I
31:48
use. But you could use it right in regular
31:51
chat GPT. And then I can
31:53
ask it to do the same thing. So
31:55
I can go ahead. Oops, that's the wrong one.
31:57
Let's go ahead and
31:59
put in the prompt, right? And
32:02
then what I'm going to do is I'm going to go through
32:04
and I'm going to grab that URL, right?
32:07
And then I'm going to ask it to go through
32:10
and write this, go
32:16
ahead and write this this social media posts.
32:20
And let's see how it does. Now, what
32:22
I'm going to tell you right now is that it doesn't
32:25
have the ability, it doesn't have access
32:27
to that to that plugin.
32:29
So I'm not sure if it's going to come
32:31
back and just use the title to write
32:34
it, or if it's going to use
32:36
or it's going to just use the URL. So
32:38
let's see what it did. So here's
32:40
the content. Here
32:44
we go. So it could not read
32:46
the article, right? So we couldn't
32:48
read the article here. So it's just
32:51
being very general about the Woodard
32:53
report, right? It wasn't able to see what was
32:55
there. So one of the things that I could do
32:58
to change this is I could
33:00
go in and instead of putting the
33:02
link in, What
33:05
I could do is I could
33:07
actually put in the description
33:09
or I
33:11
could put in the content, right?
33:14
Either one is going to give it the content
33:17
of that article. I would probably use
33:19
the description because I may time out with
33:21
the amount of data that's actually in that,
33:23
but let's go ahead and retest the step and see
33:25
what happens now that I've pointed to
33:27
from the trigger step, the content of
33:30
that particular article when Zapier, that RSS
33:32
feed triggered and Zapier pulled it in, it
33:34
pulled in the entire content of the article.
33:37
So let's see how it does here. We'll
33:45
just give it a moment. And
33:51
it's taking a little bit longer because it has to
33:53
read through some more stuff. All
33:57
right, so let's see what it did here. All
34:04
right, so this actually pulled a different
34:06
article and what it's doing is it's actually
34:08
looking at my in the news article,
34:11
and it actually did a great job, because
34:13
here it is talking about the latest tax relief
34:15
measures which was the new. The
34:18
new measures that they just put out for the
34:20
hurricane that went through so giving Massachusetts
34:22
and Maine some relief. It's also highlighting
34:25
the article that I wrote about the new members of the Intuit
34:27
Partner Council, and it
34:30
is tagging me and there we go. So
34:32
now what I would be able to do with this is
34:34
I would be able to step put another
34:37
step. To either send it right
34:39
out to Twitter, or what
34:41
I do is I send it out to my meet editor account
34:43
and it actually adds the scheduled
34:46
posts at that point and I could do this with Hootsuite,
34:48
I could do it with, any one of those buffer
34:50
any one of those social media managers that
34:52
is supported on the Zapier platform. So
34:54
you can see what this means guys. that
34:57
I can create the zap and
34:59
then don't tell Joe Woodard, but now I can go
35:01
take a nap because it's going
35:03
to do all my social coaching for me, right?
35:06
Or I'll go ride my horse. So anyway,
35:08
my point is that this is a way that I can leverage
35:10
technology. I am not plagiarizing here. These
35:13
are all our content, but I'm lever
35:15
leveraging this technology
35:18
to help me. do my job
35:20
and free up time for me to do other things like focus
35:22
on the amazing content for scaling new heights,
35:25
right? So that is one of the
35:27
tools in my toolbox. The second
35:29
one is an SEO
35:31
generator. So after I've written an
35:33
article, right? Let's
35:35
see if we can find my SEO. After
35:38
I've written an article, I
35:40
can actually go through and I actually
35:42
use a text expander tool, this magical
35:44
up here. To say, Hey, I
35:46
need a 155 character medical
35:49
description for an article and SEO keywords
35:51
that I can put into HubSpot so that we do
35:53
really well on on the Googler.
35:55
Right. So I'm going to go ahead and do that. And
35:58
now what it's going to do is it's going to create
36:00
the meta description. It's going to
36:02
create the ss e o keywords,
36:05
and then I'm able to go through and add those
36:08
within within our HubSpot. So
36:11
what I will tell you is that normally
36:13
it puts us in an array format, meaning that
36:15
it needs, it's in and I can actually tell it to do
36:17
that if I wanted to, but this would help me to
36:20
copy and paste that code directly
36:22
into HubSpot by using
36:25
another hotkey, which
36:27
in magical, which I'll talk about in a minute. Which
36:30
and it didn't do it, hold on, which
36:33
basically cops copies the code that
36:35
I need to then insert these SEO keywords.
36:38
So what I've done is I don't have to remember the code
36:40
snippet that I have to put into every single article
36:43
to make sure that it's going to have really good SEO
36:45
results. And then what I do is in the array
36:47
form and I would just copy this and paste it into
36:49
the keywords here. So what
36:51
I'm talking about is this text expander
36:54
called magical which is free, by the way.
36:56
Where I can create these little keywords
36:58
or little snippets that will pull text
37:01
into wherever I am on the interweb.
37:03
Okay, any field on the interweb. So there's my
37:06
meta. There's my keyword
37:08
and then these are sales taxes sales tax
37:10
keywords. This is my calendar
37:12
link This is the call for speakers
37:14
link because everybody asked me for the call for speakers
37:16
link 800 times, right? So this
37:19
is a way for me to just type in a code
37:21
no matter where I am on email or anything else
37:23
And it will actually add these This
37:25
information. So what
37:27
I'm what I use this for and the way that it relates to chat
37:30
GPT is you can use something like magical and
37:32
there's others out there to put in your
37:34
prompts that work really well. And
37:36
then all you got to do is put in the dash and the letters
37:39
that you've assigned to it and it's going to remember your
37:41
prompts and Michael's sharing that he uses shortkeys
37:44
which is the same thing. So
37:46
that's just two of the little secret
37:48
sauces that I use.
37:50
In my job to help me
37:52
do that heavy lifting. When
37:55
I'm working with social media for the
37:57
Woodard report. So hopefully that's helpful to you guys.
38:02
Now you're
38:03
muted, Michelle. Oh,
38:09
we lost you.
38:16
I'm muted to while we're waiting for Michelle
38:18
back, I was going to address this question by
38:20
Donna that says, does each of these
38:22
site, do each of these site the original
38:25
article so we're taking the articles,
38:27
and we're asking it to rewrite it or
38:29
whatever does it cite the article, and
38:31
if it doesn't, isn't that plagiarism. And
38:35
so my answer to that Donna
38:37
and Michelle and I had a great conversation about
38:40
this at scaling new heights, is
38:42
that maybe so first of all it does
38:44
cite the article, so it will
38:46
actually cite the articles. And
38:49
then the other thing to keep in mind
38:51
is that at the end of the day, it
38:53
comes down to you and how you use it.
38:55
Right? It comes down to your ethical
38:58
scope, right? And what is
39:00
comfortable for you. So in the case
39:02
where I was creating those social media posts, I'm
39:04
using my own content, or I'm using
39:06
content that was supplied to me by a writer who has
39:08
authorized me to publish their content
39:11
on the Woodard Report. No ethical issue
39:13
here at all. Where there would be an ethical issue
39:16
is if I went onto somebody's website, copied
39:18
their content, threw it in ChatGPT, and said, rewrite
39:21
this article, and then published it as my own. In
39:23
that case, I would say yes. 100 percent
39:25
plagiarism. You can't do that,
39:28
right? Or I don't think personally,
39:30
and that's a personal, these views are
39:33
expressed, are the personal views of Habitat. I
39:35
believe that is plagiarism. So I would not,
39:38
I wouldn't do that. What I would do
39:40
is I would go out and do my research and find articles
39:43
that are giving me the facts and figures that
39:45
I want to present to my audience. And
39:47
I would copy those links into
39:49
the article. Write the article as
39:51
a starting point, go in and I call
39:53
it heatherize it with my own personal
39:55
views and opinions, and
39:58
then I would make sure that I linked out to every
40:00
single article that was cited that
40:02
I used to write that article. To
40:04
me, that is using this
40:06
tool in the way it's intended to be using it is not
40:08
an ethical concern at all.
40:13
Very good.
40:14
And we can hear. Yay.
40:17
Okay, I'm going to go ahead and move on. Is
40:19
that okay? Okay,
40:22
so are you seeing my screen on create
40:24
your own chat bot? Okay.
40:28
All right. So here's something else that you all might
40:30
want to do If you'd like to this
40:33
is super easy Let's say your
40:35
firm has a lot of your own articles
40:37
like heather was just saying, you know You've written a lot
40:39
of your own content you know you your
40:41
firm has you know This wealth of
40:43
content or you just want
40:45
your you know, your clients coming to your
40:48
firm for questions and asking
40:50
questions, Dan's got a lot of school of bookkeeping
40:52
content or the advanced certification
40:55
or tax codes or accounting regs
40:57
and rules and, whatever,
41:00
if you would like to create your own chatbot,
41:02
it is Super easy. I am
41:05
not a techie person and I could do it in
41:07
a matter of minutes. Like with bot press
41:09
or others, you can get started for free
41:11
with no credit card or anything. And
41:13
then you can upload your own knowledge
41:16
base information with documents
41:18
like PDFs or websites.
41:21
If you got a whole site map and everything,
41:23
so you can create your own information
41:25
and then create a conversational workflow.
41:28
And then people can come and ask. The
41:31
information and access that
41:33
chat bot. It is super easy.
41:36
So for example, when you say you want to create a
41:38
new chat bot, it will ask
41:40
you, how do you want to do it? And
41:42
you can do it from documents, websites,
41:44
start from a template or do it,
41:47
handholding. Now I already started
41:49
one called QB. And what
41:51
I did is in for
41:53
the KB, the knowledge base, I
41:56
uploaded a bunch of the advanced certification.
41:59
PDFs. So I uploaded all
42:01
of those knowledge based articles,
42:03
and you can see I've got those here. That
42:05
only used like 20 percent of my stuff.
42:08
Over here, I can test it out. So
42:11
you can see I asked it a question over here.
42:13
So if I wanted to ask it something like, how
42:16
to donate inventory
42:19
in QBO, I can ask
42:21
it a question and then it's
42:23
going to go to my KB,
42:26
my knowledge base and look for an
42:28
answer first to provide
42:30
that answer. And so you
42:32
can very quickly and easily,
42:34
and it'll say answer found in
42:37
the knowledge base. So if you
42:39
want to create your own knowledge
42:41
base. And have
42:43
that available, you can do that. And then you
42:45
can have a new knowledge base out there.
42:48
And you go through and you create your
42:50
own workflows. And so out
42:52
here, you have your workflows.
42:54
Start here and if, if you get an answer
42:57
or not. And so anyway. It's super
42:59
easy to do that and go through
43:01
and create your own workflow So if that's
43:03
something that you want to do That
43:05
is super easy to do and that is something
43:08
that you can do So think about doing something
43:10
like that if you would like to and I know
43:12
there's lots of other uses for chat
43:14
gpt out there and these other ones.
43:16
For tech support purposes you
43:19
can use it, if you've got a question when it comes to
43:21
tech support, I was able to fix our
43:23
refrigerator. The refrigerator repairman
43:26
basically told us, don't call
43:28
me anymore because
43:30
I can't fix it. And, you're just wasting
43:33
your money. And so anyhow,
43:35
I was able to use actually Claude because
43:37
Cat GPT didn't give me what I wanted, but
43:39
with Claude, we were able to fix
43:42
it and we got the silly thing making ice
43:44
again and I fixed the codes on the front
43:46
of it and everything. So Claude helped
43:48
me fix our refrigerator and get it
43:50
making ice again. And it's
43:52
still running and making eyes. So
43:55
there you go. Claude helped me
43:57
with my tech support on my Samsung
43:59
refrigerator. So it can
44:01
help with tech support. There's lots
44:03
of other usages, uses out
44:05
there. If you want to research a potential client,
44:07
let's say you've got a new client that's coming to you.
44:10
You can research that industry
44:12
to find out more about that industry.
44:15
You can use it to do data manipulation.
44:17
When Heather and I were doing this for Kaylee
44:19
New Heights, I was in the car, Laura
44:22
was driving and I was typing in there, okay,
44:24
I'll do this section. Heather will do that. This
44:26
one should take five minutes. This one, 10 minutes.
44:28
And I was just typing this up in
44:30
a notepad on my phone. So I
44:32
would type like intro five minutes,
44:35
Heather, this one, Michelle,
44:37
10 minutes, right? So I just typed this
44:39
in a notepad. And what I did
44:41
then. Is I said, please put
44:43
this in a table format with these
44:45
headings, title, name,
44:48
time, whatever. And I said, put
44:50
this in table format and
44:52
it took that text and put it
44:54
in a table format. It took
44:56
from text without any commas
44:58
or anything and put it into
45:00
a table format. In chat
45:02
GPT for me so it can do
45:05
data manipulation. It can debug
45:07
your Excel formulas. It can do all
45:09
kinds of different things. So give it a try.
45:11
You'll be amazed at some of the
45:13
things that you can do. It's just
45:15
it's amazing what you could do with it. It
45:18
can clean up your bank transactions
45:20
and do your expense categorization
45:23
and stuff. Somebody wanted to know if they
45:25
can convert a PDF bank statement.
45:28
No, I don't think it can do OCR
45:31
and convert your PDF into
45:33
an import that you could import into QBO.
45:36
There's other third party apps that will do that
45:38
kind of stuff. Now, a couple of other
45:40
updates that I wanted to share with you.
45:43
On how it's going to impact how we
45:45
work and in our industry,
45:48
and this is where it's really getting amazing.
45:50
When you look at your suite, whether you're
45:52
using Microsoft or Google,
45:54
and this is like Apple or Android,
45:57
if you use Microsoft, and you're using
45:59
Word and Excel and PowerPoint
46:01
and Teams and all that. They
46:03
have Copilot, Microsoft Copilot.
46:06
If you're using Google and you're using,
46:08
Gmail and Google Drive and
46:10
Google Slides and YouTube and all this
46:13
they have Google Bard and they're
46:15
building it into all their suites. But anyway,
46:17
both of them are building AI
46:20
into their suites. You got Google Bard
46:22
and you have Microsoft Copilot. They're
46:24
building this AI in
46:26
all of their suites, just like Intuit Assist
46:28
is going into all the Intuit Suite,
46:31
it's going to be within everything. There
46:33
is a really quick little two minute video
46:35
on Google barred and how they're building
46:37
it within all of their stuff. It's
46:39
going to change how we work and
46:41
make everything so much easier
46:44
when we're doing things, when we're working.
46:47
And just doing everything that you do,
46:49
it's just going to make everything so
46:51
much easier and faster. So
46:54
it's just going to change how we work and
46:56
like with Intuit Assist and the
46:58
MailChimp, it's going to assist our clients
47:00
with marketing. We can start helping
47:02
the clients with the marketing and everything. These
47:05
are things to think about how it's
47:07
going to change how we're
47:09
doing things and how we're working
47:11
with our clients and everything. Now,
47:13
a lot of you know me, a lot of you
47:16
have heard Casey on these QB
47:18
Power Hour videos. And Dan,
47:20
I showed him this slide and I said, Dan, do
47:22
you, what do you think of these images here?
47:24
And Dan says it looks like your dog. You
47:26
took a picture of Casey at the lake. It
47:29
does. That looks like my dog
47:31
at the lake, right? No,
47:34
I went into Bing and I said,
47:37
create an image of a Bishan
47:40
on a blue wave runner at
47:42
the lake with trees on
47:44
the shoreline. And in 30
47:47
seconds or less, Bing
47:49
created four images for
47:52
me. Here's two of them. It does
47:54
look like my dog. So
47:55
Casey is an internet troll.
47:59
It's just, but Sean, but these
48:02
are being, these
48:04
are computer generated, AI
48:07
generated, and that's where you get the term,
48:09
Dan one time asked me, what does it mean,
48:11
generated? What does it mean AI generated
48:14
or something like that? It AI generated.
48:17
Something new. It created this
48:19
image. And that's where
48:22
the small businesses are really
48:24
going to benefit from this is not only
48:26
can it generate When we're communicating
48:29
and it's generating our words for
48:31
our emails that we're sending to people,
48:34
it now can generate these images. So
48:36
for small businesses that are selling products,
48:39
they don't have to go out and do a photo shoot
48:41
anymore. And you saw that in one
48:43
of Intuit Assist's videos on
48:45
MailChimp, where they had a chair
48:48
and they said, change the background. And
48:50
it changed the background for them. We
48:53
don't have to now pay for photo shoots
48:55
and graphic artists and things like this.
48:58
You can just say, Hey, create this image
49:00
for me. And if you don't like it, you tweak
49:02
it and you tell it things. And this is where Dan
49:04
and I were laughing yesterday. We
49:07
all had a part in training this. You
49:09
remember all those capture images
49:11
for years before you could log into
49:14
QuickBooks or you could log into anything and it was
49:16
saying mark all of the images that have
49:18
a school bus in them or mark all the images
49:20
with a stoplight and you had to mark
49:22
all those images. We were training
49:25
a I to recognize things
49:27
in these images. And now
49:30
we can use a I on these
49:32
images to remove things or to
49:34
create an image for us. And
49:36
I created these images with being
49:38
so if you go to being
49:41
and. Tell it to create an image. It's
49:44
free. It's easy. It's
49:46
that simple. So now, remember
49:48
when the QuickBooks and all these
49:50
third party apps, all of a sudden these small
49:52
businesses had the benefit
49:55
of improving their processes
49:57
and automating their businesses through
50:00
QuickBooks and the third party apps. And we
50:02
could help them to do that so
50:04
they could compete with the small, the bigger
50:06
businesses. Now these small
50:08
businesses are going to have this opportunity.
50:12
From the A. I. And the ability to
50:14
do more marketing and
50:16
more of this type of stuff. And we can
50:18
help them with that. We help
50:20
these small businesses get to the cloud.
50:23
We can help them implement A.
50:25
I. And how to do this stuff in
50:27
their business because it is
50:29
going to change how we work. It is
50:31
going to help us continue to improve our
50:33
automation and our efficiencies because
50:36
our bookkeeping services. It's
50:39
a commodity service. We
50:42
really have to differentiate ourselves now
50:45
and we have to look beyond basic
50:47
bookkeeping services. In my opinion,
50:49
again, you know what that's worth. Anyhow,
50:52
so we really want to focus on our relationship
50:54
with our clients and those advisory services.
50:57
And so anyway, that's my thoughts
50:59
on it. Dan and Heather, what are your thoughts
51:02
and what do you guys want to add?
51:05
Go ahead, Heather. I mean,
51:08
I couldn't agree with you more. Michelle,
51:10
I mean, not only do we need
51:12
to embrace this
51:14
technology but we need
51:16
to, in order to stay competitive, but we need
51:19
to embrace this technology in order to just
51:21
stay viable. Right? Because,
51:23
As this technology it's, I
51:25
feel like where we are is it feels
51:27
like we're just at the, if you think of a hockey stick,
51:29
it feels like we're on the bottom of the stick, but
51:32
we're not we're way past that. And
51:34
this technology has been, it's been in development
51:37
for decades. at this point
51:39
and it's come to market. So what we're going to see
51:41
is now that all of this technology is readily
51:44
available, we're going to see a huge increase
51:46
in the momentum of this. And so
51:48
I think that accounting professionals definitely
51:50
need to learn how to use
51:52
this within their practices. And
51:55
also I think our clients are going to be relying
51:57
on us to educate them
51:59
on how to use this technology
52:01
as well. So I couldn't agree
52:04
more, Michelle.
52:05
Yeah.
52:07
I think that also one other, one other
52:09
aspect of all this is that, a machine
52:12
doesn't worse,
52:14
right? And like typically, what we're
52:17
seeing today, if this is the saccharine,
52:19
the sweet and low flavor, I can tell
52:22
this is not artificial. I
52:24
can tell this is an artificial intelligent thing,
52:27
right? Whatever that happens to be, because I
52:29
think that's where we're at this point when,
52:32
if you're just getting Into the
52:34
technology itself, trying to decide
52:36
okay do I go with Claude or do I go with
52:39
Bart or do I go with chat TBT
52:41
or what's the best sweetener out
52:43
there? It's
52:45
not going to get worse. Like a machine
52:47
doesn't forget, right? Like it's an elephant in
52:50
that it. It will, it's only going to
52:52
get better, right? Like it's going to increase
52:55
as we get now, who's
52:57
going to be the Facebook
53:00
of AI, right? Who's going
53:02
to be the meta. We don't know yet.
53:04
Right. That's, we're witnessing this
53:07
as it comes out and as it plays
53:09
out. Right. So that is, that's
53:11
part of the, that's part of the challenge here is you
53:14
go with Bard because
53:16
I'm a Google suite. What if. Copilot
53:20
is the wave of the future for AI, and
53:22
now I've backed the wrong course, now I've got to learn something
53:24
brand new because I went
53:27
down one, one technology path. I mean,
53:29
the worst case scenario is you have, you
53:31
take those applied learnings and bring
53:34
them to something else. And I
53:36
think that's what we are with streaming services,
53:38
right? There's things that you can get in Netflix
53:41
that you can't get in Hulu and so on and so
53:43
forth. So it's pick your poison. And then now
53:45
you're ending up, you cut the cord
53:48
or you cut the cord because you don't like
53:50
cable and now you're spending more money on your
53:52
streaming service. Which
53:55
app or which AI is the right way
53:57
to go? I think Martha asked a question. I'm going
53:59
to pass that out to you is which one
54:01
is better, which is better engines
54:03
for different uses is
54:05
cat GPT for writing clog
54:08
for math, which, what do you two think?
54:12
I prefer Claude for writing is
54:14
my preference. But you know, it really,
54:17
it depends on what you're doing, which one
54:19
you're going to like the best. And
54:21
that's why I try both of them. And
54:23
sometimes I copy and paste it into Word
54:25
and blend the two together and choose
54:28
what I like best. And I see Heather saying, yes.
54:30
Heather, what are your thoughts? Yeah,
54:32
I mean, when I'm providing,
54:35
so Claude, I go to,
54:38
when I'm providing documents,
54:40
Just because it's easier. You can provide up to
54:42
five documents that it can then ingest.
54:45
And, it, that seems to be easier
54:47
and more reliable than when I'm using the plugins.
54:50
So the plugins can sometimes be a little bit quirky.
54:52
I do the flexibility of all of the
54:54
plugins in chat GPT, but
54:57
I've had, I've seen. I've
55:00
seen undesirable
55:02
results from both, and I've seen great results
55:04
from both. And, one of
55:06
the things I think is really important for people
55:08
to remember is that this
55:10
is experimental at this point. Both Claude
55:13
and ChatGPT are experimental
55:15
platforms, and they're open source,
55:17
meaning that anybody can use them. And so
55:19
the developers behind the scenes are
55:21
learning how we interact with it and trying to make it better.
55:24
And as we all know, as cutting
55:26
edge, technology
55:27
users that sometimes what we
55:29
think is a great idea isn't. And
55:31
so there's certain times where, you
55:33
know, chat GDP and Claude will feel
55:36
like it's taken a step backwards. So
55:38
I think it's too soon as
55:40
Dan said, with the sweetener, I
55:42
think it's too soon to give you an answer
55:45
on that. Definitively. I think you
55:47
have to experiment with it yourself. Yeah.
55:50
And they're going to keep changing, every
55:52
week, every month, they keep changing. I just
55:55
heard, I think yesterday I
55:57
think Chad GPT now has the
55:59
voice, just like you can say, Hey
56:01
Siri, you're going to be at Hey GPT
56:03
or Hey Chad, or Hey Chaz, or
56:05
whatever it's going to be called. I don't know. But
56:08
you're going to be able to talk to it now.
56:10
Instead of just typing into it, which
56:13
will be interesting, so
56:15
it will be interesting for sure. Yeah.
56:17
Yeah, it's just it's
56:19
an exciting time and go
56:22
ahead, Dan. Sorry. I was just
56:24
gonna say I always freeze up when I need
56:26
to speak to something as
56:29
opposed to a person, because
56:31
I want to say the right thing and then I say
56:33
the wrong thing regard.
56:36
Yeah. The one thing I would keep
56:38
in mind is years and years
56:40
ago when the PC came
56:43
out, everybody started using QuickBooks
56:45
or DAC Easy, or M Y O B
56:47
or whatever the accounting program was. Everybody
56:50
was like, oh my God, that's
56:52
gonna kill our business because the accountants
56:54
used to do all the bookkeeping and now
56:56
the clients can do their own
56:58
bookkeeping and this is gonna kill our business.
57:01
They won't need us anymore. That
57:04
was the farthest thing from the truth
57:06
because they needed us more than ever. It
57:08
just changed our relationship with the clients
57:10
and it changed how we worked
57:13
with the clients Not the fact that they
57:15
needed us and I think this is the same
57:17
type of situation It's changing
57:19
our relationship with the clients because the bookkeeping
57:22
is being automated All of the
57:24
basic bookkeeping is being automated
57:26
and that is going away What
57:29
we're doing with the clients is not gonna
57:31
going go away. We still need to
57:33
have a relationship. They still need us.
57:36
So we need to focus on that relationship
57:38
with the client more than ever. We
57:40
need to make sure that they know we're there
57:42
for them to help them grow
57:44
their business.
57:45
Yeah, my I just met with a client today
57:47
and his main question is, when
57:49
am I gonna run outta money? Right. When
57:53
are my bank accounts going to run out of money?
57:55
That's not something that, I mean, you
57:57
can look at a report and just extrapolate
57:59
it out. We need a little
58:01
bit more.
58:06
And that's where we can use chat GPT
58:09
and quad to come up with ideas. How can we
58:11
improve cashflow? Give me a hundred ideas
58:13
to improve cashflow. Then you can go to the
58:15
client. Here's our top three things that
58:17
we can work on for your business
58:20
to improve your cashflow. And then you've
58:22
got things that you can work with the client
58:24
on, but you can get ideas from AI
58:27
to help you to help the client. Absolutely.
58:29
And The other thing that I would just say is that chat
58:31
GPT can't look across
58:34
a zoom screen or a desk
58:36
or whatever, and empathize
58:39
with your client the way that you can,
58:41
and it never will be able to do that. And
58:43
that's really what our clients are buying from
58:46
us is that partner in their business
58:48
that's going to be there when things are great,
58:51
and that are going to be there when things aren't so great,
58:53
and they are Don't care how
58:55
we get the information. I mean, obviously we want to be
58:57
ethical, but they just want
58:59
us to give them reliable information
59:02
empathy and, trust.
59:05
So I agree with you, Michelle.
59:07
I think this technology is just another tool in
59:09
our toolbox and that,
59:12
we're just going to bring more value to our clients
59:14
because of it.
59:16
Yeah. All right. We are
59:18
a little bit over, we appreciate you for
59:20
sticking around a little longer and then then
59:22
we we have allotted. Heather and
59:24
Michelle, awesome to see you both. Love
59:27
you guys. And
59:31
hopefully you can join us next next time on the Power
59:34
Hour. Next month we'll be like I
59:36
said, we'll be talking about downloading QuickBooks online
59:39
data and using it in Excel. Maybe you can use
59:41
ChatGPT to to do some of
59:43
that. I don't know. We'll see. But again,
59:45
thanks for joining us and we'll see you next
59:47
time on the QB
59:48
Power Hour. Thanks, everybody. Have a great
59:50
day.
Podchaser is the ultimate destination for podcast data, search, and discovery. Learn More