WSW - Podcast Intuit Canny Feedback - Matching in the Bank Feed
===
[00:00:00]
Welcome and Catch Up
---
Dan DeLong: Welcome to another Workshop Wednesday brought to you by schoolofbookkeeping.com. I'm elated that I finally got the volume to just didn't cut off, as as, as it has been for the past few, times that we've been trying it [00:01:00] the workshop in, in, inside of School of Bookkeeping.
How are you, Rachel? It's good to see you.
Rachel Dauchy: I'm good.
A2X Camp Highlights
---
Rachel Dauchy: I'm back from A2X camp.
Dan DeLong: How was that? What was the... what was that, the... Was it a invite-only thing, or did you- It was.
Rachel Dauchy: Yes ...
Dan DeLong: so you are-
Rachel Dauchy: It's,
Dan DeLong: You are A2X campers.
Rachel Dauchy: Yeah. So it's really everybody involved in the e-commerce accounting world, A2X is. It's funny though because it's s- we can use these connectors between systems, really it's used a lot for Shopify to QBO- sometimes looping in other things such as an IMS, but it's not exclusive to e-commerce. Any business that's using some kind of POS system, for... We have salons and spas and things like that, and we use these for services. So it's really any [00:02:00] electronic payment that is being used, not just e-commerce, can be used with something like A2X.
But A2X is a really r- popular tool and really everybody in the e-commerce world uses it, this was very much an e-commerce related retreat and it's... was e- everybody in the e-commerce space. Very, big firms. Yeah. Very well-regarded, and that's why I was like hi. I'm here." But I... No, I met the most incredible people and there's some really skilled QBO e-commerce trainers that were there too, like Theresa Slack, and you might be familiar with her, and and just folks like that.
Dan DeLong: Yeah.
AI vs Automation Tools
---
Dan DeLong: What is their distinct advantage now in this world of AI where I'm sure there is a Claude connector to Shopify or a sh- Claude connector [00:03:00] or whatever A- some kind of AI connection or, you know, some other, you know, accountant-based platform that might be able to connect to these things and allow you to massage or manipulate the data yourself if you want in a way that you want.
What is what is A2X, Bookkeep, Webgility, Synder all of those, places or, softwares that serve as a Give me the give me the downloaded order data and the payout data and I'll get it into QuickBooks, when you're doing a- How do, how
Rachel Dauchy: do they
Dan DeLong: stack up and was there any kind of was there any talk of that at the camp?
How do we, you know-
Rachel Dauchy: Yeah ... What are we up against? Everybody knows in the e-commerce space that the let me just connect it from Claude doesn't really apply in the space. [00:04:00] It does and it doesn't. It does in that you can you can probably grab the data here and put it in AI, but the setup, it's not necessarily straightforward bookkeeping anymore because we're putting systems together.
And so a lot of times it's we need to make sure that the financial data needs to come over from system to system and because maybe they're multi-channel. And so in this way we've got invoices syncing over, but with this way, we've just got a consolidated entry coming over. So when you have a lot of systems that need to talk to each other, and a lot of times it's because maybe they've got two locations and that immediately takes them out of Shopify, so they've gotta do something in a different system Everything isn't gonna work together harmoniously with AI.[00:05:00]
You- it's really good for reading the data and maybe matching up this one connection, but it's not gonna do all of the things on a feed. That's the most important thing. A lot of it is gonna require you to upload it, and or run it through manually. So that's what a lot of people don't quite understand, is it's not quite there yet with a fetch or a feed in the way that we would use an automation tool for.
So H2X is definitely under the umbrella of an automation tool, meaning it's gonna grab it and post it automatically. So there's, if you were to think of it as a Venn diagram of there's maybe some things that, that kind of both systems can do, or both approaches can do, but it's a little where it meets in the middle.
So we really have to design the whole entire thing, and really at this point, there's nothing better that works than [00:06:00] an old-fashioned automation tool. And so- That's really the best way at this point that you're gonna automate most of the system. If you've got really, let's say I just have one Shopify store and I'm only dealing with one set of payouts, and maybe only one third-party gateway, then meaning Shopify payments.
And they're not let's say they're not also using PayPal, and they're not also using authorize.net, and they're not also using some other plugged in affirm or something, Klarna or whatever, which a lot of people have a lot of different things in there. They don't understand what they're doing.
They're adding all this complexity. And so in that situation, the biggest risk is duplication, and I... You have to set the automation up so you're not doing that, and then all of your data is just coming in pre-filtered. Again, you can't really set [00:07:00] that up in Claude. It's not meant for a fetch type of scenario Now I'm gonna stop talking about it 'cause I could yap about this all day.
Dan DeLong: Yeah, so we're not even on topic of what we were, gonna be
Rachel Dauchy: talking about today. I know. That's what you get for asking me a question about that. But it's really, interesting, and it was so funny 'cause this is one of a lot of different solutions for people that do this kind of bookkeeping and accounting.
But really if anybody goes to conferences and they know the folks from A2X, the Jamies are really, incredible people, and kind, and do informative sessions at things like QuickBooks or Intuit Connect and that kind of thing. They're very knowledgeable.
Dan DeLong: Getting you choked up.
Rachel Dauchy: Yeah. Sorry.
Dan DeLong: But it, plays into this larger [00:08:00] conversation, I think, that accountants and bookkeeping pros are, dealing with today, and we actually kinda stumbled on it on last week's QB Power Hour, which is now The-
Rachel Dauchy: The Power Hour, Shanna
Power.
Dan DeLong: There is... Yes, there is no QB, but that doesn't mean that we won't talk about QuickBooks. It just means we're abiding by Intuit's marketing guidelines before they come wagging their finger at us. Tyler who is a, who's a big proponent and, advocate of double keeper whatever you- wanna call it. He brought up a, he brought up an interesting point, and it plays into the topic we were gonna talk about today about bank feeds. And I wanted to share that clip in the chat here. So it's, it really highlights that, that part of the [00:09:00] conversation.
Using Intuit Canny Feedback
---
Dan DeLong: But, really what it boils down to, and what we wanna talk about first is, in Canny, which is the Intuit-sanctioned, feedback channel for accountants, right?
A lot of times they'll you'll, hear this message especially from the Intuit folks, put in feedback, But when you do it from within QuickBooks, it just kinda is this text over the wall, and you, don't get the closed loop.
Otto agrees that he's... He likes the feedback. Your dog likes the feedback here in Canny as opposed to putting it within the product. But this is specific for accountants. Oh, and I'm realized I'm not sharing my [00:10:00] screen. Where'd it go? Probably throw this over here. And make sure I'm choosing the right one.
There we go. Okay. Now it should be showing my screen. So it's just intuit.canny.io, right? And it's, focused on QuickBooks accountant feedback people who work with QuickBooks and are power users of QuickBooks and in the accountant space. You can go in here. You can just go to the website and then hit log in.
You just create your own login there. But one, one of the cool parts about it is you can see the roadmap. So there's a roadmap section of what's under review, what's planned, what's released. There's two main boards. One is for bank feeds and accountant agent feedback, which we have a lot of feedback for that.
And then the other hot topic [00:11:00] is modern view of reports. So you can always go in here and see what they are and-
Rachel Dauchy: Dan Are you
Dan DeLong: reading something?
Rachel Dauchy: Oh my gosh, I'm laughing at that comment. Terrible. What? Can't find anything.
And- You need to stop listening to the hipster developers with ski hats and beanies
Dan DeLong: Yeah. And and you can upvote this, right? So if you log in here you can put this under, It, it's currently listed as under review, right? So immediately you have that insight into these these feedbacks that other people put in, right? If you agree with this notion- ... You can upvote it, right?
So that it gets... So that your voice is, And that's, part of the challenge when you just put in feedback in the, the inside the product, [00:12:00] is you don't know if the feedback you're putting in is the same feedback as someone else, right? So that's the, cool part about this Canny board here, is that it helps you, piggyback I guess is the best way to put it.
Combine your voice with other people who've already done it, right? And then as people from Intuit w- will- Have this come to their attention. Project managers, pro- product managers, people who are driving these initiatives, even developers can go in here and, reply, right?
And cl- close that loop. So this is... If you're gonna put in feedback and you're an accountant, definitely have this bookmarked Canny or intuit.canny.io. [00:13:00] Because one of the things that we want to start doing here is combing through some of these things that are either under review or planned and, and/or released because that is our closed loop, from Intuit, right?
So as feedback is put in, it'll get under review, it'll be, documented internally with them, and then when they put something out they will put it on the, release schedule, and you can keep up to date because there's a lot of things going on all at once and the speed of which things are changing is every day it's faster, right?
And, and-
Rachel Dauchy: they, 'cause they've got a lot of- Hard to keep up ... ski hats and beanbag chairs over there- ... and they're drinking a lot of Red Bull
Dan DeLong: exactly. And for nothing else, that's a [00:14:00] good l- that's good for a laugh. Yeah. To go in
Rachel Dauchy: and- That made my day.
Dan DeLong: Yeah.
Rachel Dauchy: Oh, I can't find anything without Rachel, so-
Yeah
I get it. Yeah. I keep her.
Blue Match Confidence Badges
---
Dan DeLong: One of the things that I noticed on, Facebook is that people were asking about the blue match, right? So this was something that was relatively new in, release. When you're in the bank feed and there is, a match, there's now a new color to the match.
And so with no fanfare this change happens, and you're not sure, w- what, is blue? Why is it blue as opposed to green? Which is better? Do you want blue? Do you want a blue match, or do you want a, a green match? This is where you can find out those, those answers by going in in, in taking a look here.
[00:15:00] So this is what we're really talking about, is the matching confidence badges. So basically when it finds a potential match or multiple matches to something that's come down in the bank feed, it will match and indicate that it's not an exact match or i- its confidence level is less than an exact match by putting that match figuratively in blue, right?
It's basically a confidence score. So matching is good, right? So there's a whole sequence of events when something gets downloaded. It looks at the r- the register that it's in and tries to find a potential match. If it finds one and only one, then that's gonna be a green confident match.
You still have to click the Match button so that it, [00:16:00] connects those two. But if you have a common transaction, a common dollar amount that's unmatched and there's two potential matches, then it's less confident as to which one to match it to because dates, are customizable, right?
You can change a transaction date. So the posting date versus when you actually entered the transaction- Yeah ... that might, there may be multiple transactions for the same amount around the same date, so it's not sure and it's not quite confident. And that's what the blue signifies. Is
Rachel Dauchy: that, is this kind of a- replacement for the two possible matches
Dan DeLong: Yeah.
Yeah. So now you've got
Rachel Dauchy: this- Because what was the problem- ... this example ... with the two possible matches? It was very straightforward. It was go in, c- cor- find the correct, select the correct date, [00:17:00] end of story, move on. Yeah. This to me is opening up a murky, I don't know. I don't
Dan DeLong: know. Yeah, see, I'm not...
I'm trying to, I'm trying to look at this the difference here between invoice 1001 for this date. The highest confidence band is what you, what can graduate to auto-post over time when those matches are consistently accepted, right? So what was happening in the past with the two possible matches is it would just do that, right?
It would just give you the list, and you would have to choose which one it is. The-
Rachel Dauchy: But that would force you- Next- ... to make sure you knew what you were doing.
Dan DeLong: Yeah. Exactly. However, what this is doing is training the machines, right?
Rachel Dauchy: Yeah.
Dan DeLong: In this particular company, that you're doing things with, that if it's a, if it's a consistency.
And that's what AI and the [00:18:00] accounting agents are, are- Supposed to be doing right? I
Rachel Dauchy: guess. Is- I just So these two, John Smith and Better Homes, are s- related somehow?
Dan DeLong: Yeah, I
Rachel Dauchy: mean- Because only by date? 'Cause they're not related- Yeah, QuickBooks- ... by anything else.
Dan DeLong: Yeah, 'cause, 'cause there could certainly could be, a...
This could potentially be a, a, a deposit that is matching to multiple invoices, Right? Maybe that's a, it's a payment, that was for multiple invoices and that's why it potentially comes- Yeah ... into this. So it's gonna start learning. All
Rachel Dauchy: I'm seeing is person who doesn't know going- Yeah
click. Matching something ac- you know, across, especially if it's a- payment on invoice and it's not matching to an- you know what I mean? This is matching straight to an invoice, not to a payment. [00:19:00] You know what I mean? Yeah. It's just to me, opening up a can of confusing worms.
Dan DeLong: Oh, yeah. Yeah. It could certainly go off the rails for sure.
Rachel Dauchy: Ugh.
Dan DeLong: Now and I,
Rachel Dauchy: and I think- I'm just thinking negative Nelly, but I don't- Oh, yeah ... I don't love that.
Training Machines Concerns
---
Dan DeLong: It is... We're at a point where humans and machines need to interact, and this is one of those ways that that Intuit is hoping to advance, its own functionality by, pointing the the humans to tell the machines what, these, all mean, right?
Because like-
Rachel Dauchy: To me, that sounds like that should be their job
Dan DeLong: Sh- their meaning who? The, person in the bank feed?
Rachel Dauchy: No, the Intuit developer.
Dan DeLong: It, can-
Rachel Dauchy: You make it match exactly. [00:20:00]
Dan DeLong: Yeah.
Rachel Dauchy: I shouldn't have to train it. It's like going to cook your own whatever dinner.
Dan DeLong: Absol- absolutely, and I think that is probably, the, promised land, right?
We, I had a... I took part in an all hands event when I worked at Intuit in Tucson. Scott Cook and then CEO Brad Smith came down, so that, that tells you how long ago this was. And they did this whole d- design for delight thing, where it was coming up with a problem trying to fix it, that sort of thing.
But one thing that they mentioned was they were talking about Google searches, right? And we all know when we search back at, back then before AI Overviews came into the picture, you wanted to be as a business, you wanted to be on page one, right?
If you, wanted to optimize your search, [00:21:00] for your website so that you showed up on on page one. But they posed the question how many search results does Google want to give you really?
Rachel Dauchy: 473 quadrillion, 327 trillion.
Dan DeLong: Their answer was one, right? Oh. Because Google, wants you, wants to present you with what you're looking for, right?
That is their product, relevant searches, right? So what you put in versus what you get should be one, and they're trying to get to that point right now with this AI Overview, right? By, curating what you're searching for, and putting it together, and coming up with a, with an answer.
And then you're then supposed to continue the chat in the AI Overview type thing. But the problem is there's so much [00:22:00] garbage.
Rachel Dauchy: Yeah. But you know what this indicates to me? This indicates to me that they're so far off that they can't even match something like this. That using, you know- databases, that makes me feel like then you can't AI your whole entire operation.
Good luck.
Dan DeLong: Yeah. Yeah and at this point the machines aren't there yet, right? And it's not, what do they say at Scaling New Heights? They, called it, general intelligence. That's... When we get to that point and the machines are able to s- think and reason that is when, it comes to the point of okay, are these machines better than humans at doing these things?
And it's still getting there, right? So it's still
Rachel Dauchy: maybe- I'm not even necessarily worried about them thinking and reasoning. I'm more [00:23:00] worried about operators harnessing it in m- a more dangerous way. Prompting it in nefarious ways. But I'm also seeing that if it can't even pull certain things from a relational database, how are people thinking that they're gonna go in and, manage an entire- invoice or order to cash operation on, in Claude.
I don't know. I just, to me, I just can't see it.
Dan DeLong: Yeah.
Rachel Dauchy: Right now- I can see lots and lots of different pieces and parts of it pieced together, but someone is still prompting and doing a whole bunch of stuff, "Oh, just put it in here. Just put it in here. Just put it in here." you still have to just do it.
You still have to do it. You have to still put something somewhere. The whole point of automation is a [00:24:00] little different. That's a more you set it and forget it.
But, Yeah. Ideally you could set it and forget it on all kinds of automations and go in at month-end.
Dan DeLong: Like you were talking about with, with, A2X, right? That's- Yeah ... a tool that you're using and you're gonna go in and, set up those mappings the way that you, the way that you want to, do that.
Rachel Dauchy: Yeah and in a perfect world, I'd want this to be auto-matched, auto-added, auto whatever- with me just mapping it ahead of time. I don't wanna spend time on training. I don't want to- train it. No.
And so I don't have time for that. So it's if it isn't able to auto-match it with a s- very small margin of error so I can go in at month-end and I can streamline everything and do it, then I, do not see this as an advantage at this point.
Dan DeLong: Yeah. Yeah and I think this is [00:25:00] This is one of those things that is, because it's putting it into this confidence level of matching, it's just a way for the, the machines to let the humans know w- how to, interact with wi- with this data.
Eventually it'll get to the point, right? As w- as, as we are helping or as these processes are being, Managed then it'll get to the point where it just does it for you 'Cause that's-
Rachel Dauchy: Well, this is what it really is. It's a let's take accounting professionals out of the mix, 'cause we don't need trained accountants that understand how to go in and f- and find the correct match and utilize these more m- murky, mysterious functions in QuickBooks.
We need to make [00:26:00] this clear and understandable for every bu- business owner so they can go in and DIY their bookkeeping, and we have no more need for specialists that understand a more obscure, difficult software.
Dan DeLong: Yes and no, I think, right? I, agree with what you're saying. And then this brings me back to our conversation that we had with Tyler on the Power Hour about this sort of stuff, right?
Because this is where your value as an accounting professional comes in, right? Being able to handle all this, right? To your point of, "Hey, why am I training Intuit's machines to do this for me to ultimately replace me?" Because they've made no qualms about of how they're they're [00:27:00] impacting and treating the, the independent accountant community.
They, want, Clearly they want, they wanna, corner the market on the term expert, right? And,
Rachel Dauchy: and that- Yeah. But that's very clear and obvious they're trying to remove us from the mix.
Dan DeLong: Yeah.
Protecting Firm IP with Double
---
Dan DeLong: So this this, conversation that we had on the, the Power Hour, led it to at this da- this day and age and I know you aren't in the bank feed for your clients. You have a team that works works on the client's bank feed. But how much time are your i- is your team spending specifically in the bank feed for a client? Probably
Rachel Dauchy: Way, way too much. Yeah, for sure.
Dan DeLong: So what the, what Tyler brought up as a good point is that [00:28:00] things like platforms like Double and we talked about comparing that to a new player in the market called Accru, they have, they essentially are a middle ground platform between, for the firm and the clients that firm serves, whether they're on QBO, Xero- other GLs that are out there. Right now both of them, sweet spot is with QBO Xero for Double, but to Tyler's mentioning on the Power Hour, it wasn't, as robust as a feed inside of QuickBooks. But what he was trying to say, and that and, you can watch, the clip, but, this allows accountants and firms to do the work outside of [00:29:00] QuickBooks and not- Yeah
And not le- not train the machines, to ultimately replace them or not, not do, not have all their hard, work go with them if you separate from the client, right? Like just
Rachel Dauchy: I know. You know what I mean? And that's when I want to go back into Double. We used Double for a long time, and now I can't use Double anymore.
It's a long story. But yeah, exactly. I wanna go back to using Double.
Dan DeLong: This allows people in the profession to have their intellectual property, their IP- Yeah ... right? Their workflows, their tasks, right? 'cause if you're going into your client and you're setting up all these rules, and you're going into your, client and putting in recurring transactions that they do, and you do all that for two months, and then the client says, "All right, we got this.
See you later," right? They, they- Yeah ... they break their [00:30:00] engagement with you. They take their ball and go home, or they take your ball -
Rachel Dauchy: Yeah ...
Dan DeLong: and go home, right? That's a very good
Rachel Dauchy: point
Dan DeLong: So these platforms like D- Double, Accru others that are gonna be out there, we're actually talking with, with a former Intuit employee who's created a, a, a full charge soup to nuts dashboard for clients agnostic of QuickBooks, called Wesley.
and, and these things are gonna come into play and that's their advantage is that- Yeah ... You can still do your work for your client in their ledger of choice, but the how you do those things can stay with you and your team so that if they leave, they don't take all your hard work and go with them because they're just gonna- Yeah.
Well- ... either they're either gonna do [00:31:00] it themselves or they're gonna find somebody else that maybe does less or whatever they're looking for as far as change is concerned. But they're not gonna take your hard work and give that work, that leverage to somebody else so that they don't have to work as hard.
Yeah.
Rachel Dauchy: Now that's another reason I need to add my list of why I need to switch back to Double.
Client Portal Switching Story
---
Dan DeLong: So what happened? Do, do... Can you talk about it a
Rachel Dauchy: little bit? The firm that I acquired they were... So we had used Double. We had first used Client Hub as a client communication hub, and I really like Client Hub, but I needed the, y- I needed additional things such as what's offered in Double, and Double also works as a client communication, where they upload stuff into folders and whatnot.
So- We had switched to Double, but then when I acquired the other firm, they were using Client Hub, and those... I couldn't switch the client portal [00:32:00] with those folks. And Ah ... 'cause they had some stuff switched, and it was a lot of switching. But, so in the event of wanting to consolidate everything and not have this one por- portal for these clients and that portal for that client, I switched my clients back to Client Hub because they had used that before.
We had used it before, and I was familiar with it, but it just does not have the same kind of month-end capabilities and 1099 tracking capabilities that Double has. And, it's always been an issue. My team hates it. So at some point my plan was maybe Client Hub will level themselves up to have something like this, or maybe enough time will go by that I can switch the other group- Yeah
back to Double. And that's- I can't do it this year, maybe next year. I don't know. We'll see.
Dan DeLong: What might be a, a better option for you is incorporating something that [00:33:00] doesn't interfere with both of those, right? So that's where as we were comparing Double with Accru. Accr- Accru is relatively new in this AI task space, right?
Where Double has entered that, right? They first started as a client collaboration portal for, for- Yeah ... clients. And then added this, And a reporting
Rachel Dauchy: tool, sorta.
Dan DeLong: And it's now entering into, you know- Repe- repeated tasks that- Yeah ... you want to do for that client, right?
So- and it's
Rachel Dauchy: a, one of the best products that are out there.
Dan DeLong: Yeah. Now, Accru is starting with the repetitive tasks, and they don't have a client collaboration, right? So maybe- that's the tool that you might wanna investigate, is something that does that instead of switching your clients, 'cause who's gonna, who's gonna pay the price for [00:34:00] switching clients is the clients, right?
Yeah. So if you're changing the client collaboration tool that they've already used or did use in the past, they're just flip-flopping between which one do I go in now?" And, yeah.
Rachel Dauchy: Yeah,
Dan DeLong: that might be something- And then,
Rachel Dauchy: and then- ... I should definitely check it out. Yeah, 'cause- And they're just gonna text you anyway.
And then you have the risk of, then you have the risk of, okay, adding on a new tool, oh, that's gonna change- Yeah ... 180 degrees in two months, 'Cause the technology moves so fast that I just don't necessarily ha- I can't... I I've had so much whiplash lately- with tools that I'm like, I don't know, I just kinda wanna put a, "We're using this for now" on here until I, I see something that's worth a massive switch.
But no, I def- I would definitely like to check that out.
Dan DeLong: Yeah. Now, we, Oop, let me stop sharing here, 'cause we're, kinda done talking about the matching piece of that.
Accru Cohort Announcement
---
Dan DeLong: So we're gonna have [00:35:00] Accru is actually wanting to leverage School of Bookkeeping as their educational partner, because they are- Whoa
bringing you into the space. So what we're doing with Accru is we are doing a cohort. Very similar to this type of format with the workshop but it's solely gonna be just for the use cases of, Accru, right? It's gonna be a paid cohort where there's, some skin in the game, where you have to pay for the, the option to get in.
And then be able to review the recordings that we have. And I just put the, the link for the interest of that, because that's gonna occur right after Intuit Connect. Once we've get back to back to business or back to- wherever we happen to be. I don't know where I'm gonna be, by, by that time.
But once we get set settled there we're gonna do four live sessions [00:36:00] of the things that you can do in Accru, and manage those things in with your clients. Now, if you're already using Double- Oh ... I don't want you to switch or, do anything- Yeah ... like that. But if you have a paid account for for Accru, the cohort is free, right?
So there is no- Interesting. Yeah ... Potential option to do that. I've been-
Rachel Dauchy: Here's a little pricey. I'm looking at it right now. It's, dif- But I mean-
Dan DeLong: It is a different pricing model than Double, right? So-
Rachel Dauchy: Yeah ... with
Dan DeLong: Double- Yeah,
Rachel Dauchy: it's a d-
Dan DeLong: yeah ... Double is a- That's what
Rachel Dauchy: I'm guessing ...
Dan DeLong: Double is a, per client, right?
Which makes it real easy for, for people to kinda test it out, right? I'll, I'll- I'll test it out with this client, right? Like, how well does this work? And then you can branch into and go up to using the Ask Double and the Double AI, if you wanted to. Whereas Accru is more of you get a, a, a maximum number of clients to work with, and you pay a platform [00:37:00] fee for that, and then there's nothing on top of that yet.
I'm working with them to talk to them about a proving ground version where you have maybe three clients or something like that. Maybe a certain, you know- trial time period where you're able to really test it out, because that's typically what accountants will do, is they wanna...
They don't wanna move everybody unless they are sure of that.
Rachel Dauchy: So that- Wow, it'll even do a cleanup diagnostic. Oh, boy.
Dan DeLong: Yeah. Yeah, so- Oh,
Rachel Dauchy: boy. The see, this is what happens. Now I'm looking in here going, "Oh, I gotta schedule a demo." Dan, why are you doing this to
Dan DeLong: me? Sorry, Rachel, but I... That's, what I'm always on the lookout for is, the, shiny ball,
Rachel Dauchy: right?
That- I do not have time for this right now. But I... This is really intriguing, and maybe I'll do the core ho- [00:38:00] co- cohort. Yeah. I can't even tell you how
Dan DeLong: I don't have time for this right now. That would be probably the easiest thing to do is join the cohort. Without the, without it obviously we're gonna use these use cases and between live sessions it would be best to put them to practice, with an account. And, there are trials and those types of things, and we'll work on that, so you have the, your choice, and then that will give you more indication of is this right for you. Because there are so many different options nowadays. It's like you were saying at the, at, at scaling, it was like you just don't even wanna go to the exhibit hall anymore because you're just distracted by the
Rachel Dauchy: new shine.
No, I don't even go in there. No, I can't. I have to have blinders. Actually, this year I'm not going to Intuit Connect
Dan DeLong: Yeah. It, that's probably best.
Rachel Dauchy: Yeah. I'm taking [00:39:00] I'm getting my master's right now. I'm taking audit and it's this semester and data analytics, and at first, on audit, I was like, "Oh, this is cool.
This is easy. I'm gonna sail right through." And then, yeah, then I just had this week's assignment and I was like, "Why did I say that?"
Dan DeLong: Now,
Rachel Dauchy: speaking of- If any of you on here do audit, it's-
Dan DeLong: Yeah ...
Rachel Dauchy: something.
Dan DeLong: Yeah. It's all timing, right? It's when the, time is right to do these things.
Rachel Dauchy: Yeah.
Dan DeLong: And I think that is where we're at with with AI and accounting, right? Like, when is the...
A lot of people are in different places in their journey with AI, right? It- Yeah ... from is AI a glorified search results or, is it, is that something that you ask, or is it something that you've created agents to be able to do some of [00:40:00] these repeated tasks, right?
Rachel Dauchy: Yeah.
Nitwit Prompt Stitching
---
Dan DeLong: And that's one of the things... We talked last week. You weren't here 'cause you were camping with the Atech- A2Xers. We had, Joel come in from Nitwit and he did a nice deep dive on on his tool, which is essentially, grafting prompts- To multiple, AI agents or AI LLMs, even your own, right?
So that way you don't spend a lot of time. And I think the easiest way that I can explain it to, to people is that my son created a game. It's a it's like a Minecraft type of caveman game, and he was showing me, showing it to me when we visited in a couple months, and what he did was he ta- he had a chat with ChatGPT to basically formulate the ideal prompt to send to Claude [00:41:00] to actually make it, right?
Rachel Dauchy: Yeah.
Dan DeLong: Why do that? Rather than wasting his tokens and his usage, usage-
Rachel Dauchy: Oh, smart ...
Dan DeLong: With Claude, he did it in a free a free portal or a free option. Yeah. And that's really what Nitwit does is it allows you to send- ... those prompts to multiple LLMs, and then it can knit... That's where the knit comes in.
Oh. It can stitch together the commonalities and identify the differences, and then you can either share that with your client or share that with somebody else, share that with your team or take that and then put it in your CoWork or, or- Yeah ... Some other thing-
Rachel Dauchy: That's really
Dan DeLong: cool.
That's really cool ... To get the most out of your prompts because, as we come into this age of goldfish memory with AI where it just forgets the last thing you told it to do- [00:42:00] Until that gets to a point, this is great because you can repurpose those honed in s- crafted prompts and be able to use them rather than where did I put them? Did I put them in a Google Doc? Did I put them in ... Where did I put them?
Rachel Dauchy: So- I, a lot of, like, when I took a AI class, there was like you can create a prompt library and, like, all this stuff- Yeah
and/or a prompt database, which I did. But even that is still, a lot of it, you still have to do it. It's a lot of work. Yeah. If an o- and the- If you can save it in some kind of way, then cool,
Dan DeLong: yeah. And the thing that really is the, the challenge is that when you send that prompt to ChatGPT and then send that same prompt to Gemini your results are gonna vary, right?
So you don't know which is the best one to use in a circumstance. [00:43:00] So this allows you a side-by-side comparison of those different models. Does it
Rachel Dauchy: have a dashboard where you can see the things? Yeah. Oh, wow.
Dan DeLong: Yeah.
Rachel Dauchy: And then it saves them in a, in the th- okay.
Dan DeLong: Exactly.
Rachel Dauchy: Wow.
Dan DeLong: So-
Rachel Dauchy: Interesting. Oh, I need that too.
Thanks a lot, Dan.
Dan DeLong: So we're, we are starting a cohort next week on that one. So-
Rachel Dauchy: Okay ...
Dan DeLong: So from the 20- on the 28th, and we'll have a cohort or live sessions Monday and Thursday at noon with Joel to go through each of those, Use cases as, as, as we t- we talk that through. Okay. But I'll send out an
Rachel Dauchy: email- I don't know if I can do the nitwit one, but I'm gonna definitely do the other one.
Dan DeLong: Yeah. And, if you are not, if you're not signed up for the cohort we're working on an option or a membership option where you can access the live sessions after the cohort's done, right? So [00:44:00] you don't have to necessarily participate in the live cohort, but you get the benefit of those, those sessions.
We're looking on ways to right size, right time, right? Because not everybody's ready to do a live cohort or has the ability to do that, in this, at this time, at any time. So we wanna make them available for, for the community at large.
Wrap Up and Next Week
---
Dan DeLong: So that kind of wraps, wraps up our workshop today.
Next week we are gonna be talking about something that I haven't even looked at yet, so I got a week to figure out what in the world this is. Playbooks is a, a new is- newest feature of Intuit Accountant Suite. I'm not sure what flavor of Intuit Accountant Suite, if it's on Core or Accelerate.
I imagine it's gonna be part of Cel- Ex- Accelerate. I'll do some double-checking [00:45:00] before next week. But it's essentially a way to document your standard operating procedures for you and your team. When we do this, this is what happens, right? That's what playbooks, indicate to me.
So we'll talk about them next week on the Workshop Wednesday, and hopefully everybody has a great week ahead. We'll see you next time. [00:46:00]
