091 Your Accounting Firm Meetings On AI

I need some good news. I need something to get excited about. Let's talk about how AI is about to make our client meetings more productive than ever before. Kind of like what we did when we talked about your practice management system on AI. We're going to talk specifically about how it impacts meetings. What do we already have today that can make us more productive? What's like really close, like next six months and what's a couple steps further down the road? Let's talk about it. Come on in. So you may have heard there's been some controversy recently about a change to Zoom's terms of service that basically said they can use any of their user data to train their own AI models. Uh, and obviously that's concerning. So user data in the context of Zoom being like meeting transcripts, meeting recordings, pretty much any data they have about their Zoom meetings terms of service seem to indicate that they could use that for their own sort of internal AI purposes. So that was concerning. They came out the next day and actually issued a clarification to say that there's an opt in. But what isn't still clear is the opt in is like, are you going to opt into letting Zoom summarize your meetings and stuff like that. So what isn't still clear is whether you can opt into their AI features without giving them permission to use all of your data to train their models, like, in order to leverage summarized meetings. And some of the cool AI stuff that AI is already enabling that we're going to talk about. Can you get that stuff without Zoom, then using your data to train their own sort of internal AI models? So that's still unclear, but really this is not too dissimilar to, you know, kind of how firms are going to start approaching meetings in a much more meaningful way to capture that context in a way that's going to be more helpful for your team, for your firm to create less wasteful back and forth with the client. Meetings themselves are an incredibly rich source of context. And while I don't want my virtual conferencing software provider to be like, leveraging all that stuff, I do want to leverage it within my firm. So let's start with stuff we already have. We've talked about this a bit, but the stuff that you can plug in today that is super useful. What kicked all of this off, like the big kind of jumping off point for how AI has changed meanings was OpenAI Whisper, which is a speech to text AI model that overnight made auto transcripts auto captions really good. We've had Auto captions for a long time. Like the captions that it generates, for example, on YouTube videos back in the day. And they've gotten a lot better. But in the past you would see the auto captions and they were like hilariously bad. But that stuff has gotten way better. There's been a couple more models that came out this year that are every bit as good, if not a little bit better. So the notion that you can have a meeting and in a really quick, really inexpensive way, get a full accurate text transcript, that kicked off a lot of these products where without that quality of transcription, that product wasn't really helpful before. And the first products we saw were specifically around like summarizing meetings. So you have a meeting, but you have a recorder, like a recorder app sitting in on that meeting. Afterwards it will send out a summary to all the people involved. It'll even send out like action items based on what was discussed. Fathom was a tool that was doing this early days. I think Hugo may have been doing this now. Zoom is in the process of building this into the product. Microsoft Teams has a premium version of Teams that will now do this through sort of AI stuff. I don't know if they've actually rolled it out yet or if they're still just talking about it. We did another episode on why you should start recording your meetings and how that's kind of becoming normalized now. Admittedly it is still kind of weird when you're in a. Especially in a one on one video conference when there is a third person there that is just a recorder. That's a little weird. But I do think that there's ways to work through that and set that expectation from the beginning. Check out that episode if you're still kind of stuck on that because there's a huge number of benefits to capturing those meetings. Not capturing meetings is like, what if the organizational policy was to just basically delete emails like after they had been resolved? Like you're. The meetings are such a rich source of context that honestly you want to have for years. And I think there's like easy ways to kind of come over the struggles with that. We talk about more in that episode. But then a really cool unlock that Fireflies was the first one I. I saw lean into Fireflies. AI was not just summarizing meetings, but chatting with meeting transcripts. Because if I think back to my early firm running days when I was the junior and I was blocked by the meeting that the senior had with the client and I needed the info that came out of that meeting, a summary, I don't know would always give me that information. Oftentimes it was, did they talk about insurance expense on this specific rental or something very, very specific that a summary wouldn't necessarily talk about. And that's where being able to chat with a meeting transcript is really helpful. And part of the functionality of that chat and why it's fundamentally better than what we've had in the past is chat can pull up from that meeting the most conceptually relevant parts of the meeting. So in the past we've always had captions and transcripts, they've been a thing. And you can kind of do like a control f text matching search to find any specific things. But oftentimes that doesn't work. Like you can't get to the exact right word. Or maybe if it's something weird that was said, like a street name where it doesn't know how to spell it, the captions are going to somehow get that wrong in a way that maybe you can't anticipate. But the value of chat is that you can chat with that meaning transcript in order to request like conceptually was this stuff discussed without having to nail any of the specific text to, to find that thing that you want. And that for me, like as a senior in an accounting firm, massive unlock. Because especially, you know, and the worst of it is tax season when you've got like back to back client meetings and you're perpetually the blocker for the rest of your team and then by the end of those meetings you've forgotten what you talked about in the first meetings, can't make sense of your notes. But even if you're a solo operator, just like there's so many situations where I was like, did we talk about that a year or two ago? Like I have this vague recollection of it. If my past meetings were just searchable, gosh, that would make my life easier. So it's something that we need to just start capturing now so that we've got that context. And honestly the tools that we have to comb through and surface things in that context will continue to get better, which are some of the future developments we're talking about. But what can you do today? You can do the meeting summaries, you can chat with your meetings, you can do follow up like items to dos. But right now all that stuff lives within the tool that you use for, for those meeting transcripts. So it's not living in the tools that you use day to day yet, it's really living kind of in these other systems. Better than nothing. And I do still think you should be capturing these meetings and even down the road you may be able to migrate them to a different system. But one of the limitations today is they're all happening kind of in this siloed place where your other systems can't see into it. But still a huge positive like the Net benefit. I think with these tools there's less pressure to unnecessarily pull other people into meetings. If you think about a meeting where in the past you may have pulled like two or three other people in, now those people can get up to speed and even search and chat with everything that happened in the meeting without having to be there. So it's probably going to save you having to get into some meetings. Another handy thing is somebody shows up late. Some of the services Microsoft Teams talks about. This will actually give you the transcript live so that you can see anything that you missed from before you joined. And then you just hope that they weren't talking badly about you saying, oh, Jim is late again, because that would just be uncomfortable, right? So that's all the stuff that we have today. This episode is sponsored in part by the fine folks at Cloud Accountant Staffing. Do you hire accountants? Bless your little heart. Not the best part of the job in my opinion. Not something I ever enjoyed. Well, listen, you can build your accounting dream team with talented offshore accountants in the Philippines that work 100% full time for your firm. Their accountants aren't freelancing or contracting for multiple firms. They're all yours. They work exclusively for you and are incentivized to stay with you and your team long term. They're not gonna get swiped. Cloud account staffing is 100% dedicated to the accounting industry and founded by a former accounting firm owner that understands your business, knows your pain points. They had to hire some accountants and they said, you know what, we're going to build our own pipeline in the Philippines, going to pull in some super talented people and then open that up to other firms. Basically that's the story I've been talking about a lot about staffing, building more resilient staffing pipelines for your firms. I had staff in the Philippines totally red pilled me to like, oh geez, like we need to globalize the the way that we get our work done. Check these folks out. Link in the show description cloudaccountantstaffing.com hey. This episode is sponsored in part by Canopy, the practice management system. Canopy unlocks the firm that you always wanted. Think about It. Close your eyes. Lean back in that chair. What is the firm that you always wanted? Oh wait, Canopy unlocks it. And they do this by unclunking accounting firms with an end to end solution that makes your tech stack feel a little less. Cause it's end to end. Putting our customers first with world class user experience, support, education and innovation rooted in customer feedback. Working and working well anywhere and for any size or type of firm. Wherever you are now and wherever you're going, multiplying your efforts so your practice requires less proverbial midnight oil. You know, sidebar, if you go to the conferences, Canopy's got like, they always do some like really good little like sorta, you know, the stuff that they use to like trick you into coming to the booth. Well this year they've had like Legos out there, maybe, maybe a double down on the midnight oil thing, you know, maybe like, I don't know, give away a little, you know, little actual midnight oil. I guess it would need to burn too, but that one's free. I think it's a good idea delighting your clients with a modern, easy to use portal that helps you get the info you need when you need it. That is kanopy. Check out the link in the show notes to learn more. Let's talk about what's close, like what's next 6 to 12 months and then what's coming. So what is close? Searchability in the systems that you already use. The reality is when I'm in my practice management system and we're, we've got, you know, files in there and like that is supposed to be the home for everything that I discuss with my clients, you know, like organizational email history. All that stuff needs to be in that one system. And as long as my meeting transcripts are somewhere else, that's kind of a pain. And it's a big source of friction compared to just searching within your practice management system to then have to go out to the suffix system and search in there. So that all is going to get way more helpful once we can pull that stuff into a single system. And for the practice management systems that enable searching into the contents of of files, that is not just if you search, you're not just getting file names, you're actually getting the contents of the file. This is something that we'll have super soon because you could just pull in the text transcripts from that other tool or if we can get an integration that will push the text file that is the transcript out of that, save it as a file in Your practice management system now that's searchable, whereas before it wasn't. I know Client Hub has said that they are working on this, a way to pull in the, the context from your meetings into the practice management system. But getting all that into one place like that's going to be a huge quality of life improvement that we'll have on some platforms, honestly within the next few months. It's going to be the kind of thing where other platforms like will just take years or we'll never have it, but we'll start getting it in a matter of months. For now it's just a pain to have to kind of hop out to that other system. Right. Another thing that is close is how those follow up items are managed. So most practice management systems now are really leaning into requests, client facing requests and being able to set those up to go out automatically and remind the client and all that. We've really been pushing this paradigm of let the machines do the requesting and the human communication for like the humans do all the other stuff. You don't need to be the one following up with them about the documents that you need. We got to let the machines do that. So, so we're not refiguring out what we need to request every week. And this is where the follow up items like the TO DOS at the end of a meeting ought to be pulled into that same request system. So when that transcript comes out, there's a list of TO DOS for everybody. Some of those are gonna be internal to dos, some of those may be client TO dos and you need a way to automatically get those into the system from the transcript. And that might happen totally automatically. Probably not in the beginning. There's probably some sort of approval workflow where it says, hey, it looks like we need to set up these client tasks and these internal tasks, do these look good? And then you can just acknowledge them and then it will ping that client right away after the meeting to be like, hey, here's a reminder of the stuff that we need. But then the other valuable aspect of that is it gives them an explicit place to put that information that you're waiting for. That's something that these tools already will do for you. Creating those follow up items, what we need is just a more intelligent way to push that into the systems that we already use, rather than having this like siloed additional tool that has an entirely separate set of TO dos. Right. Like that's a pain. Especially if you're trying to train clients to reliably use the portal and always Come back to that same place. That's where you really want them doing everything. Another really cool thing that is like a definitely possible today thing, we just haven't seen anybody do it yet is sentiment analysis on the meetings themselves. So we've had AI models that do sentiment analysis for years. And that is you can plug some text into it and it will kind of give you a rating of like, are they mad? Are they emotional? Like a whole bunch of different dimensions on which you can like measure that communication. And interestingly, when you think about how that could be helpful across your entire firm, Sentiment analysis on the client level. So like, what do all the communications look like from that client? That's really helpful. Trends by a client, maybe in the past they've been super cheery. In the last three months, things have kind of gone sideways. You want to know about that? But then also at the client or at the staff specific level, we all have very different communication styles. And if you're trying to maintain like a consistent representation of your brand and you have an expectation of what that ought to look like, that sentiment analysis across the staff person level, it's probably going to be helpful too. Now, like, admittedly there's a little bit of a sort of odd aspect of like surveillance Big brother to this. So there's maybe a little bit of thought that needs to go into that. But I can tell you as a firm owner, if a client's gotten grumpy in the last few months, I sure want to know about that. So that sentiment analysis, that's the thing we've actually had for a while that just gets really useful when we can automatically pipe our captions and our organizational email history and all that stuff into one system to kind of have a barometer of where everybody's at. Last thing I've got here on the list of what's really close, amazingly, is like anything related to language barriers. Arguably the coolest thing about these speech to text models is that they will do this in almost any language that people speak. And they're relatively fast. So most relevant to me in my firm running days was having Spanish speaking business owners and the struggle with like them only really being able to work with the folks we had on our team that could fluently speak Spanish. And this isn't going to be a relationship for like, I think human beings ultimately that can converse in the same language. But the better this gets, like, it's hard to overstate what a meaningful thing that is for, for the world to just be able to communicate seamlessly with Folks in virtually any language, like the sci Fi version of that and Hitchhiker's Guide was like this universal translator that goes in your ear. In Star Trek, everybody just spoke English and Star wars is the same way. But unless we can get the entire universe on board with a single language, this seems like a cool sort of practical way to break down a bunch of barriers. So that's like really close. I think we'll actually probably see that built into the video conferencing platforms and is another thing that's kind of like making the world smaller. I think Covid made the world way smaller when it forced people to work in a distributed way for the first time. And the biggest long term impact of that on how we approach staffing will ultimately be, oh, I just realized that I can work just fine with my team from across town. So does my team really need to be across town? I think Covid actually accelerated like the globalization of how we staff in a really cool way. And when you take that even a step further and you can eliminate language barriers, that's a really cool possibility too. In fact, if you work with offshore folks who like English is not their first language, you know, and you can see it when you realize that person just started pumping their communications through ChatGPT, right? Like, have you experienced this yet? It's pretty interesting. Non English speaking folks more and more are using ChatGPT to better comprehend what you're sending them and more clearly write what they are sending back to you. And it's really, I think it's really obvious when they do it. And just like so many AI things, everybody's like, ooh, is this a bad thing? Or is this a good thing? Or something in between. Honestly, the more I've thought about this because I've got several people I work with that do this now. Every single message I get from them has very clearly gone through ChatGPT. Honestly, right now, like in the context of how I work, I actually love it. I think it's great. I think it's. It shows some initiative from them of going out and trying to, I don't know, like, present themselves in the best way possible to me. And I also think they're generally getting a deeper level of comprehension of the things that I'm saying because it can be explained to them like natively in a language that they understand. Now it could probably go too far where, like, if they don't have any English comprehension, maybe it's difficult for them to determine whether what ChatGPT is telling them is correct. But that's sort of an interesting development that we've gotten recently that has changed. Definitely how I work with offshore folks for whom English is not super comfortable. Okay, that is what's like just around the corner. Like, that is stuff that in the next 12 months we will definitely have access to. Doesn't mean that every single platform will be implementing it. But honestly, the point of having these conversations is to shine a light on the value of this functionality and then amplify this to encourage the platforms that we use to pull it in. That's why I've always really stressed requests. I actually think we've seen a lot of investment go into requests across practice management platforms in the last six months or so. We saw this when we talked AI around practice management. Obviously the world is much bigger than the things we're talking about on this podcast. But having these conversations and users like you having a voice and talking about these things, sharing the stuff that we're talking about, like it genuinely makes a difference. This episode, believe it or not, it is sponsored in part by the fine folks at firm360. You know what's better than anything else? Because you don't know how to make sense of this ad, right? Like, how do I know this better than any other tool that I'm using? How do I cut through the fluff? I'll tell you what's the ultimate validation. Social proof. When your friends like, yo, that stuff's for real. It's the best. That's what I look for, right? So let me show you some social proof on Firm360, who has a shocking, shocking number of five star reviews. How about this one from Kelly? Great project management software that allows seamless communication with coworkers and clients. Think about that. Do you have seams right now? Are there seams in your communication? I bet there are. How about Brandon? This one in June of 23. Wow. We just an amazing company to do business with. With their firm management software and assistance, we have streamlined more of our processes and become more efficient and profitable. Every other firm management software we researched and demo seem to be lacking for accountants until we found firm 360. Which is not surprising since accountants are the ones who founded the company. True story, Cherith. And know the needs of of accounting firms. Every month we get updates and in the year we've been with them, they've adapted and evolved their software as well as they listen to their customers. Okay, Brandon, why don't you do the ad read? That's it. That's the end of the Ad read that was better than anything I had written. Learn more about Firm360 and the link in the show notes gang. This episode is sponsored in part by Client Hub. Listen this week on Tales from the Hub, we've been talking about the importance of a platform that automates your firm's cloud client requests. And recently we heard about how super smart accounting firm adopt a client hub. And their biggest productivity improvement has come from utilizing the Client Tasks feature. Client Hub has taken a novel approach to the client task feature. It's built right into the firm's workflow. In a job and Client Hub, you can have two type of tasks, Client tasks and internal firm tasks. For example, got a monthly bookkeeping job. The firm creates a client task in that job that requests the page the bank statement from the client and that task automatically gets sent to the client when the job is supposed to start. You don't have to wait for Tina to send that thing over and then realize why is this month end close not done on the 15th? Oh Tina, did you ever email them about? Nope. Thanks for nothing, Tina. And of course with the web and mobile apps, Client Hub is so simple for clients to use. It's easy for them to get you the answers that you need. And if they forget, it's gonna pester them with those auto reminders. With Client Hub, super smart firm always has the client docs before they need to even start the work period. Wow, Exclamation mark. That is this week's episode from Tales from the Hub. Check out Client Hub at ClientHub app or the link in the show notes. Okay, what's coming? And this is the stuff where in my mind it's like definitely the next three years. Sometimes it happens faster than I think it will, but it's in the pipeline stuff that we have to look forward to. One really better integration of all communications. Whether it's a meeting, whether it's a phone call, whether it's an email, whether it's team chat. Right now those things kind of happen in siloed ways. And what I would rather have is like a more platform agnostic way of managing all that communication. How do I roll up all of the conversations I've had, whether it was an SMS or a call or a meeting? How do I roll that up into one body, like one CRM that is comprehensive and includes all the communications that client has had with any member of my team or you know, communications that anybody that works for that client, you know, anybody associated with that client, all the comms they've had with anybody in my entire team right now. It's like you've got an email tool, you've got a team chat tool, you've got your meetings app, and they're all happening in different places. But where we need to get to is all that. All that communication centralized somewhere in a single place. So that you've got this clear view of, like, all of the communications that have happened. A biggie that I would love to see. You know, we've got this. This kind of paradigm shift now to using requests in your software to get things from clients, basically assigning these requests to your client. I would love to see more intelligence between a meeting transcript and the request itself. So we can have a meeting, we can talk about the things that I need to get from that client. They get it, they talk about it in the meeting, and then at the end of that meeting, the request is resolved. Right? Because it can see into the transcript, it knows that we discussed it, it can see the answer that the client gave, and it can embed the most relevant part of that meeting discussion into the request itself. So that when you come back and you gotta find that information, there it is, not the entire meeting transcript, but just the part that's relevant to that request, and then that unblocks whoever is working on the project. This is particularly valuable when you work with big teams. You have multiple people who are talking with clients. You think about big audits, especially when you've big teams working with a bunch of stakeholders with the client. You've got all these different blockers relying on the human to then work out that this thing that was said resolves this task. Like, there's a lot of work that just gets blocked because that's. I don't know if it's a hard thing to do, but across a bunch of tasks and a bunch of clients, that often breaks down. Another exciting development is the notion that a lot of this stuff can actually be done live during the meeting. So we've seen with Microsoft Teams Premium examples of the stuff they're doing within a meeting, where if I join late, I can see the transcript that's already happened. But as that live transcript is happening, how can that actually inform me as I'm running a meeting to make that a better meeting? So, for example, let's say we've got kind of a template for how we want to run our discovery calls with clients. And every time we want to make sure we ask eight questions or something like that, there's tools today that will run you through this in A very linear kind of robotic way where always ask these questions and put the answers into this form. But like the more intelligent version of this is it can see everything you've covered in the discussion and dynamically show you like, hey, here's maybe some areas where you need to fill in the gaps, like live as that call's happening. You know, if you do annual intake calls for receiving tax info, like what are the boilerplate things that you always need to remember to ask that meeting tool being able to show you that stuff on the fly to make sure you don't get off the call until you like track down all those things that would be really valuable. Pulling in other organizational information like requests. You're meeting with someone and you don't realize your other colleague needs this bit of information from them. Great. That system knows that surface that request to me as I'm meeting with them. I love that even the ability to like maybe propose follow up questions when they give you a sort of ambiguous response to a request. Because if it can see the transcript as it's happening, it can assess has what was said actually like fully resolved this request or is there any ambiguity there? Man, I would love for like the tool to follow up if there's like a sort of half answer that the client gives for the tool to kind of be suggesting as we're talking, you may need to get more information about this or that. If you think about sales calls and sales best practices and how much coaching goes into that and the things that you need people to cover on every sales call and how that could really dynamically assist them during a call. A tool's ability to see into the rest of your organizational data and fetch things for you automatically. Let's say you're talking about something with a client and they're like, yeah, did we claim this last year or was that 21 or 2022 when we bought that property? Well, if that system can see into all of your other client information, as soon as somebody says that in the transcript, you can have an agent or whatever it is automatically go out and try to track down that information for you as you're talking about it. Right. We've already seen some early browser based agents where you can give it a prompt. Pixie's got one. They're working on where it goes out to zero and fetches all the information that's requested. But we've already seen very early AI agents that will go out and do this stuff when you prompt them to do it. But how about that tool that is capturing a transcript of your meeting live as it happens. And then when somebody frames something that like, is unclear or is a question, that agent just automatically goes out and gets the answer for you. That's going to be super cool. But then even like more abstractly, that tool as sort of an assistant for brainstorming, like as you're maybe thinking things through things with the team or thinking through things with a client, maybe even tax strategies, if it's connected to, you know, an AI model that is trained on tax research and, you know, tax strategies, maybe it's something that you have access to the ability for that model to like speak into the stuff that you're talking about or maybe debating or maybe considering with the client to add in any other considerations, like, yeah, maybe you could hire your kids, but you're an S corp. So like this kind of changes that discussion accordingly. I think there's probably a lot of ways to make those meetings more productive on the fly and then last, just more generally kind of like insights based on the meetings that your team is holding. So to use kind of another industry as an example, we're seeing a lot of fine tuning now around AI models and companies building their own sort of custom AI models for their purposes. And if you use the example of people doing customer support and, and handling tickets from customers, if you take, you know, if you got 100 support agents and you take the five highest performing support agents, you take all of their ticket history and fine tune a model with it, and then use that model to suggest replies for the rest of the team with the goal of that sort of informing best practices for how you respond to those tickets, you're able to support the rest of your team and probably get better alignment around exactly how you want them to respond to this sort of thing. And so as we're capturing our meetings, especially meetings around specific topics like sales, that sort of thing, AI can probably be pretty helpful when it comes to suggesting what you could have done differently, other ways that you could improve upon it, what person A is doing differently than person B. Because if you have a team, most of the work that we do is like pretty siloed. You don't necessarily see all the email interactions, you don't hear all the meetings that they have with clients. And when you have all of that information across the whole firm, you can say, here are kind of some outlier things that maybe this person does, outlier things that maybe even this client does. And then there's probably some helpful insights to be gained from all that information. Now, that's probably one of those things that's pretty domain specific. So, like to build a general tool to do that that would be very helpful for us. I don't know that that's realistic, but specifically around the work that we do, if this is, for example, attached to the tool that we use to manage all of our outstanding requests for tax information, and these are tax intake meetings, there's probably some really helpful insights that it could generate for you there. Especially running a team that is doing these intake meetings all day long. That is that last block. That is the stuff that's like probably next few years sort of thing, and it might come in the next year. But those are more like the big lift things that I think we'll get. And honestly, like, all of these things have a pretty big impact on how we capture and organize the information that comes from these meetings and then how intelligently we can follow up to the things that happen in that meetings. From getting the client to track something down to us figuring out some answer to a question and then pushing that back to the client. I don't know about you, every time I kind of do a deep dive around a specific thing and think about how AI is going to impact it, I come away just being more excited. I don't really come away being more scared. I'm like, oh, yeah, you know, like, I didn't really enjoy loading client tasks into their portal after every single time that we met. That's nice that that's going to make my life easier or being able to search all of the meetings I've ever had with that person, like, that's going to be really valuable for me. It's going to save me time. Doesn't really feel like it's going to put me out of work. But I think the more we explore these things and the specific impacts they'll have on our work, I don't know, the less spooky all this stuff becomes and the more we actually steer our tech partners in a direction that is really helpful and constructive for us so that we can stop wasting time on stuff we don't want to do. Right, that's it for today. We're probably going to do more of these. Like, how does it change? How do proposals change on AI and how does this. And how does that. Probably do some more of these in the future. Let me know what you think and I'll see you tomorrow.