Removing The Blindfold: How To See Automation Hiding In Your Day-to-Day

Hello, hello everyone, and welcome to another short webinar, uh, where we'll be discussing an interesting topic that I find dear to my heart, which is how to see automation opportunities that are hiding in your day-to-day. This is something that a lot of firms we talk to struggle, and to me it's almost like sometimes I can't relate. It's like there's so many things they're doing day to day that the moment we see it, it's like we absolutely know there's an opportunity to automate it, but sometimes they can't see it. So I tried to put in writing some typical signals you can use for yourself. to identify automation opportunities of your own and for your team. Okay, for those of you that don't know me, my name is Isaac Perdomo. I'm the founder of Opsware, an automation agency that helps accounting firms. We serve over 100. I lead a team of 10 people, and I really like what I do. Let's move on from that quickly. So let's talk about this simple concept. In order to automate something, not only do you need to know how, which we talk a lot about the hows and do all of these tutorials on various tools, but you also need to know what to automate in the first place, right? You need to know that it can be automated. You need to notice that a task is a candidate in order to be automated. And what I've found is that that's a skill that you can build into yourself and you can build into your team. And we'll be sharing some examples today so you can start to train that skill and get that more and more refined and turning on that automation radar, as I like to call it. Of course, then you need to know what tools fit for this particular process. Have you set up those tools? That's where we spend, you know, half a decade building that skill set and speak about that often. But for the purposes of today, I do want to focus on the first one, which is learning to notice. of the things you do on a day-to-day basis, what are some candidates for automation? I'm going to talk about 6 signals to spot attack that's worth automating. We'll talk about déjà vus, drop balls, inconsistencies, data hunting, repeated requests, and tailored distribution. Some fancy names there, but on each one I'll kind of encourage you to think about for yourself what are some of those Things that may apply to you, and I'll share some examples of our own, things that we've experienced in our agency day to day and that we realize, hey, actually we could automate this for ourselves as well. So let's talk about number 1, déjà vu. Like, what do you find yourself doing again and again and again and again? Perhaps something you don't enjoy potentially, or you used to enjoy but you've done it weekly for 2 years and all of a sudden it's become boring and repetitive. So start thinking about things you do on a daily basis, things you do on a weekly basis, things you do on a monthly basis, etc., etc. As an example for us, something that I'm doing now on a weekly basis is sitting down to review the content that our marketing staff has helped us write and put together. Funny enough, I didn't have any automated way to do that and I was just doing that myself, but I started to think about my own thought process like, wait a second, every time I open up a piece of content that our team wrote, I am going through a process. I'm going through a series of steps to, for example, first assess if this is a piece of content that's even worth moving forward in the first place. Trying to understand if this aligns with our voice or verifying any facts, trying to remove anything that's very AI telling that something was written by AI, trying to humanize it a little bit more, put my own sauce on it if you like. I started to build a skill for that particular task so that now whenever I sit down, I can very least drop the link, We use Notion for this. I drop the link into Claude and tell it, hey, I'm about to review this task, run through the process that I've taught you to do, which is it assesses, is this a piece of content we're gonna kill or we're gonna keep? If we're gonna keep it, what are some of the review points that you may want to look at so that then at least I have a starting point to start to go deeper into it? Now, what's a task that you repeat? on a regular basis, uh, that you find yourself doing weekly, monthly, quarterly, yearly, daily. And one of the things I find interesting is like literally in some cases all the deliverables of an accounting firm are recurring in nature. So that's why sometimes I'm like, don't relate with a firm in terms of they don't know where to start or they don't know what to automate in the first because there's just so much that they could be doing, just like list your deliverables, right? You are like doing daily reconciliations, you're doing, uh, weekly this, you're doing monthly financials and closes, you're doing quarterly tax filings, you're doing yearly returns. Like all of those things are candidates for at least you to start to understand and break those processes down, break the steps you do to bring those deliverables to life, to start picking apart which of those steps can you automate and how? Let's move to the second signal, dropped balls. Okay, so every time you find yourself realizing that there's something that should have happened that didn't, that's a signal right there. Is there someone you didn't reply to, a follow-up you didn't do, a mistake that was made? Each one of those can be an opportunity for you to have a routine task automated on the background, an agent, an automation that's gonna monitor to make sure those keeps— those things keep happening, or at least it flags you that they haven't happened. What's an example of that? Um, at least for us recently, we were having troubles with like certain, uh, clients we didn't reply to in time, or like their messages got lost in the inbox or in Slack. So one of our client managers actually, without me necessarily pushing him to do this, he created a skill that now runs, I don't know if it's every week or a couple times a week, to go through and pull, hey, look through all of our inbox, all of my Slack, check if there's any messages I actually failed to reply to that's still an open question that I should have replied to. Go ahead and find that for me and flag it for me to actually look into that. How many times have you have something like that happen where, right, you don't get back to someone? Another example, I'm part of a community and we sponsor and are a good kind of some sort of commercial partner with that community, and we didn't pay an invoice. We simply forgot, and this person reached out like months later to tell us about that. In their case, they probably could set up some sort of recurring AR report of like, hey, here are the invoices that haven't been paid under the last X days, and get that recurring reminder of something that, uh, at least in their case, they noticed, uh, it wasn't— it hadn't happened yet and they should have. So that's, uh, signal number 2 for you. If you find yourself catching anything that slipped through the cracks and perhaps think for yourself something that happened last week, that something that happened yesterday, because this happens all the time. We're humans and it's very normal. Signal number 3, inconsistencies. This is somewhat related but different, which is like, hey, there's something that should happen regularly in the business, perhaps daily, weekly, but sometimes it doesn't. So it's like sometimes it happens, sometimes it doesn't. It's some sort of, some sort of like uneven reliability, if you like. Um, and similar, just Just to share a personal example for us, we have a process internally to meet with our clients every other week, and we have a process to then send them an update every week of what we're gonna work on and what we've worked on on the prior week. Sometimes those client updates went out, sometimes it didn't. Of course, sometimes our client managers are very busy on the day-to-day work and they might not have the time and the bandwidth to actually do that for every single client every week. Again, what did we do? Well, we created a skill. that now goes through all of our clients, all of our active clients, looks at all of our tasks within our systems, which we, you know, use solely to track. And that's where all of our team comments and has back and forths on the work that's getting done. From all of that, it summarizes anything that we've gotten done, anything that we have that we're still working on, any roadblocks that we face, right? We have a blocker here, but we're kind of trying to figure that out. Figure this out in this, on this other way. And then it populates all of that for us in the Slack channel so that now our team just had to copy paste, ensure that that's, you know, a little bit of a human review before sending it to the clients and then off you go. So think about for yourself, what's something that's happening inconsistently in your firm today? Another one, data hunting. I love this one 'cause I love numbers, I love metrics, very analytical guy. And as I'm sure many of you accountants are as well. So is there a metric, a data point that you find yourself having to hunt down on a regular basis? If you have a team and you have some KPIs, is there anything you want to track on a regular basis? More often than not, there's a way to bring that data to you instead of you have to hunt it. And even if— what are sometimes the pushback I get? Oh, it just takes 5, 10 minutes to pull that data point. Yes, but because it takes that long, you don't even search for that number in the first place and you don't look at it. So all of a sudden you're not getting the insight that you should just because this is a little bit of friction between you and actually getting that number. This is an example again, something that I can talk about internally. We rallied the team around being able to ship faster and move work faster for clients, so we started to track time to completion. The idea is we want to track from the moment the task is created until the moment the task is done. How long does it take us and how can you rally the team around improving that metric? For the purposes of that, we created a beautiful dashboard. Let me see if I can share it actually. I don't know if I'll have to share my screen in a different way, but let's see if you can all see this. Looks like you all can. That was probably somewhere around here. Yep, somewhere around December we started to track this metric, and you can see now how every sprint, which for us is a 2-week period, we can see how much tasks took to complete. And for the most part, you can see the trend has been going quite down, and we have now a target to ship things on average every 10 to 14 weeks, which we can see we've been kind of like on track to have that rough average. Over the last few sprints. So again, think for yourself, what is a metric that you need to keep digging for? Some examples that we've built for firms are any sort of like financial split. So maybe, of course, you'll see your revenue accounting software, that doesn't need any fancy thing, but sometimes people want to see revenue by partner, by month, revenue per hour, like some sort of effective rate. Revenue by pod. You want to perhaps split the revenue between recurring and one-off. As I share, there's usually other metrics that you want to track around perhaps NPS, client satisfaction, response time, on-time delivery if you're tracking due dates and whether tasks are being done on time. Sometimes there might be like a speed within a time window of like how many, what percentage of returns are done by this time, what percentage of closes are done by the by the 10th, by the 15th, by the 20th. Perhaps some sort of capacity utilization, units of work, how many returns of the total amount that we have pending we've done. Like there's all sorts of metrics. You can get creative here, but there's, I wouldn't say even almost always, there's always a way to dig for data and have a dashboard as real-time as possible showing you those metrics. In an automated fashion instead of having you having to look for them. Another area we see sometimes it's a little bit more tricky to, to, to track, but, uh, some form of quality sometimes firms want to track, whether that is number of revisions or number of review points or revision rates, or sometimes we've seen firms tracking like how much time for this deliverable was spent on the review stage. There's kind of different ways you can go about it. And also, of course, sometimes comparing your budgets to your actuals. So that could be time budgets versus actuals, dollar volume versus actuals. As I mentioned earlier, like how many— what percentage of work you want done by a specific date compared to your targets, sort of like trying to get a sense of what are the targets you have and how far off or how far above those targets you are on a regular basis. This is probably my favorite, if I were to be honest, but let me share 2 more with you real quick. One is— another one is repeated requests. What do you find yourselves requesting from external parties again and again and again and again? The most common thing here that we see is just some sort of client onboarding questionnaire. On the tax side, tax questionnaires that you're having to request all the information you need. Uh, sometimes on the audit side there's a series of documents you're collecting. On the accounting side there's a series of documents that you're collecting on a regular basis. So anything that's client requests or perhaps vendor requests, information you need to capture from an outside party, there's almost always a way for you to build a tool that allows you to— allows the client to fill that out and have them— have you remind them on a regular basis. So that's another big area that we see. It's like, what do you keep asking for? And this could only— not— I mentioned here external parties, but this could be even internally. Perhaps there's things you're asking from your team and asking, putting systems in place to capture that information automatically. It's always a big win as well. The last one I have for you is tailored distribution. What information do you keep having to like repackage, repurpose in a different tone and for a different context. Again, going and sharing an example for us internally, one of the things that of course responsibility of our client managers is to take the context of a meeting and make sure now that our project managers and our engineers are aware of any changes that the client might have asked for, and then perhaps there are some action items for the client. So there's also things that within that meeting need to be shared with the client as well. So our team has created a, again, like a skill, a recurring agent that every day goes through the meetings for the prior day, get a sense of like, is there any new requests that has been done here that has been requested from the client, picks that up, shares it with our team, and puts it on the appropriate task that he finds is relevant for that request. Request, and then also create the client actions from the meeting that are things that we may want to follow up with the client on that they committed to doing to move our work forward. So think about what are you repackaging and rewriting for different audiences on a regular basis, not just in the context of your services team, but also perhaps even your marketing department. It's very common where you're writing a particular piece of content Recently, actually, I was testing Victor versus ClawTag. This is like the within Slack agent and like proactive AI employee tool that recently came up and I was testing it and then I told it at the end, hey, I want you to write me based on my experiments and my tests here, a version for LinkedIn, a version for Twitter, a version for a community, and it was able to do that for me. Of course there's some level of editing that comes after that, but that's another big area that we find opportunities in, another signal for you to turn on your automation radar to get the sense of what— where does this apply for you specifically. With that out of the way, just 15 short minutes here, feel free to— if you're looking here on LinkedIn or YouTube, click that subscribe button or follow because we're gonna be sharing more and more Uh, like quick little brief webinars like this. And if any of you happened to be available here in the live, I see 6 beautiful people. Uh, feel free to comment on wherever you're seeing this, and I'll be happy to answer your questions. Let's give people a minute or 2 to see if anyone here would want my insight on a particular question. Otherwise, I'll let you all go and enjoy the rest of your day. Okay, let's see, did this conversation trigger any particular idea for you? Maybe yes, maybe no. Maybe I didn't accomplish my goal, or maybe we'll see some Questions come up after the live. All right. Well, no questions from what I can see. So as I mentioned, thanks for joining. I will be doing these short and brief automation webinars more regularly, so subscribe for more and I'll see you in the next one.