How Claude and Gusto automation freed payroll capacity—and what accountants s...

Attention: This is a machine-generated transcript. As such, there may be spelling, grammar, and accuracy errors throughout. Thank you for your understanding! Echo Zhong: [00:00:00] The way I see it is AI creates capacity, but it does not create demand. I think it is the job for us as the advisors to identify, demand and and capture them. Blake Oliver: [00:00:15] Are you an accountant with a continuing education requirement? You can earn free Nasba approved CPE. For listening to this episode, just visit earmark.app in your web browser. Take a short quiz and get your certificate. Hey everyone, I'm Blake Oliver. Welcome back to the podcast. I'm here at gusto next with Echo Zhong. Echo runs her own firm, Zhong and Sanchez. And she's been a CPA for over 15 years. This morning she did the live demo of Gusto's new Multi-client MCP Gusto's pitch is that when software handles payroll, your firm gets its hours back to do advisory. But when echo freed up her own payroll manager, the hours didn't quite go where anyone expected. So today we're going to talk about automating payroll. And then what's next. Echo welcome to the show. Echo Zhong: [00:01:06] Thank you Blake. Glad to be here. Blake Oliver: [00:01:08] All right. So before we get into this demo that you did on stage, which was super impressive, by the way, I want to know about your firm. Tell us about Zhang and Sanchez. Like, who are your clients? Where are you? What do you do? Echo Zhong: [00:01:22] Cool. So we're a boutique CPA firm located here in Cupertino, California. We primarily serve modern entrepreneurs. So they could be bootstrap SMB owners or VC backed startups in the Bay area. We're right across the street to YC. So we actually serve a lot of YC companies and neo companies. Um, so, so yeah, and we're a full service. We do accounting, tax and advisory. Um, and, uh, yeah, that's, uh, that's the firm. Blake Oliver: [00:01:51] So walk me through this task that you automated today, uh, with Claude, like, what is it? What does it do? Paint a picture for those of us like and just assume, you know, like we know what Claude is, right? We all chatted with the chatbot, but like, like, other than that, we don't know. Echo Zhong: [00:02:07] Okay. Um, so in short, this is a task that we have set up in Claude to automate our, um, our pay run every week. And also it, it can update the cash flow forecast automatically for, uh, for the clients that we deliver advisory for. Um, basically what it looks like is, uh, once we run the skill run the task, Claude will go in and check all our inboxes and try to figure out which client has sent in the hours for processing, who hasn't? Or maybe there's, um, there's additional hires that needs to get input or maybe a deadline is coming up. We haven't got the info yet. We got to follow up with the client, basically highlight all these information for us to take an action on. Um, and we can automate that action and approval process so that we can actually get the payroll ready for approval, and then we can either approve it directly in cloud or in gusto, depending on what you feel comfortable with. And then additionally, for, for people who are familiar with the cloud workflow, obviously you can then go from there. You can download the Payroll Journal report in an Excel format, and you can tell cloud to update the cash flow forecast based on all these numbers you've run. If there's additional contacts in the example that I shared, a client of ours, they made a new hire and it's based on it's had the same pay and schedule as one of her existing employees. So we can easily just drop that context into cloud so that cloud can automatically update the cash flow forecast, call out the additional assumptions for us to review and edit before we ship it to the client. So all that is is built into this, this task workflow. So, um, doesn't matter who go in and run it. Like it could be me, it could be anyone from my firm. They run this workflow and it will just deliver consistent results. Blake Oliver: [00:04:08] So you're doing all of this in cloud, like you're not having to go into the gusto interface to do any of this. Echo Zhong: [00:04:15] Correct. If you if you set it up, obviously there's permissions set up where, uh, you can give, you can basically, depending on your comfort level, you can, you can say, I want all the approval to happen in, in cloud. Um, you know, like, but, but basically we're, yeah, we're able to actually execute all of this without leaving the interface of cloud. Blake Oliver: [00:04:37] And you said it's set up so that your team can run it. Is this like a skill that you've created in cloud? Is it a scheduled task? How do you do this? Echo Zhong: [00:04:46] Um, this is a scheduled task. Um, but we also created skills that are that our team member can just, uh, cue the scale and run this process, um, in particular projects if they want to as well, because we share the projects within the firm as well. Blake Oliver: [00:05:04] So it's pulling in the hours. It's, it's teeing up the payroll, adding new hires. And then you said something interesting about like updating a cash forecast. What kind of cash forecast is this? You know, are you using software to do this? Like, how does that happen? Echo Zhong: [00:05:21] Uh, for this particular client, we've just been managing their cash flow forecasts and Excel spreadsheets. Um, obviously there are different tabs on assumptions and numbers. We typically do the 13 month cash flow forecast for this client, uh, just so they can be prepared for, you know, tax payments. Uh, or they can just have clear view on where their cash is going to be in the future. So in this case, um, we, we gave Claude the access to the cash flow forecast templates. So cloud can can locate that template and pull in the contacts and just basically upload, uh, basically update a new draft of that Excel template and submit it for our review. Blake Oliver: [00:06:07] Okay, so you were doing all of this manually before without cloud. Echo Zhong: [00:06:11] A lot of them. Yes. Blake Oliver: [00:06:13] All right. So like, let's quantify this. I mean, if you can do it now, all in that chat interface and you're not having to like pull in these numbers, go into different systems, enter them, update the forecast manually, like what is the impact of this on time? Echo Zhong: [00:06:29] I think there are two aspects to it. One is obviously for, for the time saved because obviously it takes time for you to, oh, find the client folder, uh, drill down to the folder that has the cash flow forecast template, pull it up, and then pull up the payroll report that you just run or like, or pull up that email that I forwarded to my payroll manager on that context and then factor all that in before you even manually updating the Excel sheet the correct way. And then once that's done, there's, there's still a review process that I have to go through. And you know, before, before this is all, you know, done and sealed up. So there's definitely the time factor I think. Um, I would like to say, I would like to think that we're saving at least 80% of the time just because we didn't have to scramble through all these contexts. Um, another aspect I. Blake Oliver: [00:07:23] 80% of the time. So like roughly. Right. So like, what did, what do you think it took before if you had to ballpark it? Echo Zhong: [00:07:30] Uh, depending on the, depending on what we're updating, obviously, if we're just adding a new hire, if we already have a good headcount tab set up, then, you know, maybe, maybe 20 minutes, half an hour, depending on who's doing it and depending on if they found that email that I send them, maybe it just slipped through the cracks. So there's a lot of overhead there. Um, but if it's more complex, for example, oh, they add a new hire. They wanted to do some scenario analysis. Oh, what if I, what if my sales go up or what. Because I hired this new person, my sales should go up. Like, what does it look like for me? Or, oh, maybe we're cutting back. Like sales has been going down. Like, what does that look like for me? If, if we layer in scenario analysis, that could be hours of work. But with cloud, those are those are just tasks that could easily be automated right now. Yeah. Blake Oliver: [00:08:28] So there's lots of different ways to set up an MCP, um, APIs M.c.p.s. Kind of the same thing, right? It's the way these apps talk to each other, right? What do you like about how gusto has set it up. Echo Zhong: [00:08:42] Mhm. I'm, I'm very excited about the Multi-client MCP because I think most of the MCP s on the market right now doesn't support the Multi-client feature. And what that means in practice is instead of like setting each client up, uh, each, each client, its own project and its own connection, I can simply just do one connection to, uh, to gusto MCP and it has, and I can just give it access to all of my clients. Blake Oliver: [00:09:16] All the clients under your dashboard. Echo Zhong: [00:09:17] All the clients under that I have access to. So that, that makes it a lot easier because, uh, not only on the setup because like in the future when you're running agents in cloud, all you need to do is to call the gusto MCP instead of like, if I'm running payroll payroll for 50 clients, if I don't have the Multi-client MCP, I'm going to have to like run the task for 50 times versus if under Multi-client MCP, I can just run it one time. It will check all 50 clients for me. So I think it's a it's a pretty, pretty big game changer. Blake Oliver: [00:09:53] So for any developers listening, go multi-client with your MCP. If you want accountants to be happy. Echo Zhong: [00:10:00] I would be happy. Blake Oliver: [00:10:01] Don't make us set it every an MCP server for every single client. Yeah. Connection. I mean, yeah. Um, so one of the things I've been doing recently with Claude is like, I have it reading my emails and like helping