Modernization Deep Dive: Trial Balance Agent

See how AI can help modernize the month-end close in JD Edwards. In this Deep Dive, Kevin demonstrates the Trial Balance Agent, a digital worker designed to continuously monitor financial activity, surface anomalies, identify potential issues before close, and recommend corrective actions. You’ll also see how the agent fits into a broader financial digital workforce alongside AP and reconciliation agents.

Transcript

So yeah, welcome everybody to our last deep dive session for our Modernization series.

As I said, I just recorded it, so we'll send this out.

If you missed any of the other 8, I believe we're on #9 here the last one.

They're all up on demand on our on the series page.

So if you go to our website, the events page, you should be able to find that it's where you signed up for this.

So today we got the trial balance agent.

And again, we'll keep it short, about 30 minutes.

We'll include a demo.

Please ask questions.

We can address them during the presentation or at the end, wherever it makes sense.

So with that, I will go ahead and turn it over to well, actually, sorry, let me before we get started, I'll I'll talk through ERP suite.

So I don't know, I don't know if everybody's been here or not.

I haven't looked through.

But if we've already heard this before, you know what we do, but it could include hosting managed services, JD Edwards, what you'll see today with AI and agents for the enterprise.

We got all kinds of different consulting and, and help and advising within JD Edwards or, or even outside of JD Edwards and then our EPM practice.

So if you're struggling with implementing EPM or running an old EPM or don't even have an EPM, that's something that we can help you with.

Now with that said, I will turn it over to Kevin to kick things off here with our agenda.

Thank you, Scott, and thank you everyone for joining us today.

It's great.

It's yeah, like I, I can't remember when we started this, Scott, It was like in August or was it in, I mean, I know we were in the middle of the summer, right?

And now we're now we're getting to be the fall, right.

So it's a, it's been a great 8-9 weeks with you guys.

I think it's been a little more cuz we missed last week for in focus will be week 10.

So it might have been as late July.

There could have been late July, right?

So I, I, I really appreciate everyone kind of hanging in there and, and, and working with us through this, you know, through this kind of journey of modernization for JD Edwards.

So today, you know, if you've been to any the other ones, this will be my third right that we've talked about the AP agent, our financial reconciliation agent.

And now today we're going to talk about the trial balance agent, right?

The cool part about is they all kind of work together, right?

And we'll, and we'll touch upon that as, as, as we look at the agent itself.

So, so the plan today is we'll just talk a little bit about about that, you know, why this is important, right?

How the modernized JD Edwards Corporation closes their books today and then give you a live walkthrough of the agent itself.

And then talk about where that this all fits into the whole digital workforce and what goes next.

And like always, we'll be here for question and answers.

So me, you're probably sick of me, but I like I said, I appreciate you guys hanging in for me with me.

You know, Kevin Van Horn, I'm the managing director of AI here at ERP Suites.

I, I've got to update the slide because I'm actually on 36 years of JD Edwards experience.

I started with as a customer in 1990 and I've, and I've been with it ever since.

I was one of the original members of the AEC group.

I was a guy who architected the home building module there and and then I stayed on at Oracle and LED the largest sales cloud sales territory for Oracle up until New Year's Eve 2019.

And then I, I left to get back because I wanted to give it back more into the JD Edwards ecosystem, right?

So like the other agents, you know, we, we have a very kind of dedicated and, and detailed approach of how we handle these, a these agents, right?

You know, as we've kind of discovered and working with our customers, you know, these agents are very personal to organizations, right?

It's how we stop thinking about agents just being software code and being actual digital workers, right?

So on the, you know, the first step in the, in this process is an assessment, right?

We go in and assess your, assess your organization, your business process, look at the the gaps, you know, what's causing problems, what, why where this agent can fit.

You can really think about it is is kind of, you know, developing that job requisition for that digital worker, right?

You know, where is this digital worker going to fit into your organization?

Then we build, then we do, you know, you know, and whether that's an agent like the trial balance, which is already built or a custom agent.

This is where we either build and, and or deploy, right?

This is where we on board that digital agent, right?

We work with you, we make the connections, we we test the agents processing and, and do all that.

So that's kind of that traditional role and then we optimize, you know, we're not one to just kind of dropping an agentic AI agent off in your organization and walk away and let you deal with without you deal with that.

We run this very much like we run our managed services customers where we look at the agent, we look at the, its performances on a monthly, weekly, daily, monthly basis and we provide resources as part of the of the implementation and onboarding process to help refine and, and, and enhance that agent during its life cycle.

Oops, I apologize, I went one too fast.

And that kind of flows into this, this whole idea of, you know, it's, I don't call them agents, I call them digital workers, right?

Because that's truly what they are right there.

You know, people think, they think about agents and, and their mind goes to a piece of code.

I'm really thinking about, you know, a digital worker that plugs into their into their organization and works with their people to help make their business processes more effective.

The difference is our digital workers are always on.

They want 24 hours a day, seven days a week.

They don't take vacations, they don't take weekends off or holidays.

They're extremely goal oriented, right?

One of the things during the assessment will create the goals for the agent for a trial balance.

You know, that might be, hey, how do we reduce the, the time to close, the monthly time to close from six days to two days, right?

And the agent will work towards that goal all the time.

One of the things that makes agentic AI different than say, digital assistants is its ability to learn and adapt and to evolve, right?

And that's what these agents do.

They work together, seamless collaboration, no friction, you know, whether that's with other agents or with your people.

And they're just incredibly scalable, right?

And and, you know, you can deploy one agent to do a certain job, but you can grow and grow and grow as an organization and that that agent can continue to handle that throughput.

So why does closing take, you know, so long, right.

You know, and and you know, and this is the you guys live there.

Anybody out here on, you know, on this call and and this webinar is looking at this going.

Yeah, I mean, it takes a long time, right?

You know, the, the reason is most of the stuff is manual, right?

Most of the problems you're trying to solve are are, you know, you have to look for things manually, right?

And it's hard, you know, sometimes those things are hard to find, right?

You've got, you know, in this example, we've got four companies and but nobody knows which one's the problem, right?

You know, you day one it start, you start triaging, right?

Exporting pivot tables, spreadsheets, eyeballing accounts, running reports, right?

10s of thousands of accounts.

Nobody reviews, you know that nobody reviews in the trial balance before that, all right?

The break usually happens at the end, right?

The problems always happen at the end, which causes those things to push to 3456 days, right?

I've talked to customers who it takes 1012 days to close their books at the end of the month, which is, you know, amazing, right?

What we want to try to change is how do we turn that into an agent who's looking at, you know, reconciliations and the kind of the monthly close on a day-to-day basis so that when you get to the end of the month, there's no surprises, right?

Presenting you issues on, you know, day -3 you know, before it, before you know, you find the problem day +3, right?

Look at it by company and understand the statuses of each of those.

You know, it will show you the problem.

So accountants don't have to look at 10,000 accounts.

They have to look at 2 and manage that, right?

You know, we're looking at these things helping solve problems in minutes, not days or hours, right?

And I think that that will be the key.

All right?

So how does it close today?

You know, start triaging on day one, you know, you know, after the triage, you people start scouring Ledger's exporting files, going into, you know, running pivot tables, all that kind of stuff.

But ultimately what ended up happening was there was one problem, right?

One thing that was out of balance that was blocking the clothes that nobody knew, right?

It was just that they had to search for and find, right?

So the break was not a hard fix.

It was, it was just, it wasn't a problem.

It was just a very hard problem to find, which is what makes the closing so hard and so long to run, right?

It's that, you know, unlike the AP agent and really kind of unlike the reconciliation agent, the trial bounds deals with time, right?

A time constraint, right?

You know, how do you close your books in a very timely manner?

So when things happen on a such a condensed schedule, that's when problems start arising and problems start pushing to the next day, to the next day, to the next day.

So let's let's actually go and look at at the, at the agent itself and, and we'll follow that kind of difference throughout the at the follow it through the agent and you can kind of see how that handles it, right?

So let me minimize this and if you've seen this before, you know, here we, here we go.

Here's this is an example of our, our trial balance agent and really any of our agents, right?

You know, we always have a kind of a landing page for each one of the agents.

And the interesting part about this, and I've been asked this a lot, is that, you know, these landing pages, these dashboards aren't for the agent, they're for you, right?

They're for the human workers, right?

They're to be able to see what the agents are doing and what the agents finding, right?

They really kind of give give you visibility that's not available within JD Edwards today, naturally.

I mean, the data is there, but you don't have a that one place where you can see things like, you know, what's going on.

Is there an unposted batch in error?

You know, you have to go to that screen.

Is there a normal account?

Is there there is something missing, a journal entry?

You have to go to all different places to be able to go find that, right?

Whereas here we give you a dashboard so that you can start seeing some of the things that are going on within that in real time.

Again, this is for the human worker, not for the digital worker, right?

This is an example of what's going on.

So as you can see here, you know, all the cool stuff, you know, the agents connected, it's connected into JD Edwards.

You know, it's, it's, you know, orchestrations are, are, are tied in.

They're working.

But I have now I have a dashboard, right?

And you know, the first one is looking at clothes.

You know, we're in company #1 for, you know, the January of 2026, right?

If I wanted to go to Company 50, I could run that analysis and it could give me, you know, different information about each one of the companies.

I'll go back to number one, run the analysis again.

It's doing the work, right?

And it's telling us that, hey, we're not ready to close, right?

We've got 2, we've got 2 blocking errors here that need adjustments, right?

We, you know, we've got unposted batches in error and we've got an active, an inactive account carrying non 0A non 0 balance in that that needs to be adjusted.

You know, we're giving some warning messages about abnormal account balances and missing journal entries.

And then it shows you things like, hey, what are the accounts that I need to look at right, prepaid expenses.

I've got a, you know, a, an old clearing account that needs to be resolved and we got some accrued expenses that is probably missing a journal entry here because it's one of it.

This, as I mentioned before when we were talking about these agents, they learn, right.

So they know if accrued expenses gets a, a journal entry every month, you know, if it gets the accrue with the, the accrual at the end of the month and it doesn't see that journal entry here, it could say, hey, we're missing something, right?

We're we're probably missing a journal entry that, that, that we need to book, right?

So you're, you've, you haven't done your accrual, the things in balance, but you've probably missed something here in order to do it.

And that's, that's really kind of key, right?

Because it's learned, it learned that, right?

It learns that, you know, you've got to do these accruals at the end of the month and it knows that what these balances on these accounts should be looking like, right?

And things of that nature.

So if I look at this and some of these prepaid accounts, you know what's going on here, you know, I have a, a little trial balance over here, right?

That's telling me that, hey, I've got a, I've got a, a credit in this account when it's expecting a debit here.

So this is something you need to look at.

We can look at things like, you know, what are those unposted batches, right, that are going on, right?

You know, this one, you know, batch #3 is unposted, but it, you know, and it was done by, you know, a Morris and it's for 24150 bucks.

It's just unposted, right?

I need to go in and make sure that that gets posted so it updates.

But this one was actually kind of an interface that came in, but there's an error with it that really needs to be resolved, right?

And the cool part about this is it's, it's going to learn about this, you know, when this happened, right?

You know what I mean?

Not, not like it's, you know, this isn't something you're going to find three or four days later down the road.

As soon as that, as soon as that unposted batch is in the system, it's being tracked.

And you'll know about that throughout the process.

I mean, in, in theory, you know, if you were on the other call about the reconciliation agent, you know that that was my whole pitch there.

It's like reconciliations are an important part of the the finance and accounting department's jobs, but often they get pushed to the end of the month and then they get part, they become part of this.

Well, the reconciliation agents happen happening all the time.

It's looking at reconciliations every day of the month, right?

Every hour of the day during that to help you, so you can look at that, you know, snapshot of your, you know, financial and accounting health on, you know, on any of the the bank, the AP, the AR, the intercompanies, the inventory, the fixed assets, any of those at any point of time.

Same thing here, right along the top.

We have a bunch of different things that we can check out and you know, and it gives us the things like the trial balance, which is what you saw on the first screen.

But you can see this here and you can actually look at it and say, hey, you know what's going on with this one.

You know, if we drill back, you know, there, you know, there was a, you know, you know, this is the account, this is the period.

It's, it's got a negative $200 opening with no debits and no credits.

You know, it's something wrong with this account because usually I'm expecting something to be there, right?

So any one of the accounts and you can, you know, by your own regard, you know, pick which accounts you want to check out from a trial balance.

If you want to look at all of them, you can.

If you want to look at the type of accounts, you know, cash or ARAP prepaid, any of those, you can do that as well.

You can even look at them by business unit.

That way we can look at anomaly.

We have a screen for an anomaly detection, right?

And it can, you know, it can rate and score those anomalies by, you know, hey, these ones are kind of a medium 1, you know, but this one is a severe one that we need to look at, right?

Like, you know, this batch #4 is an error, right?

And it needs to be fixed.

You know, that account we just looked at has a credit balance when it shouldn't be that.

And we've got an inactive account that's carrying a $50 balance that needs to, that needs to get through, right.

I can look at some of the other ones that are out there like, you know, here's an unposted batch.

You know, that's a medium issue because all you need to do there is go in and post the batch, right?

And, and it will be good, right?

So here, you know, so you can actually look at the anomalies by severity, you know, and if you see over here it says corrections drafted, I go back up to the top.

The system then will be smart enough to try to look at the the errors that you're having and recommend journal entries to correct those problems, right?

So here, for instance, it's asking, you know, what did you, what did you want to do with this abnormal balance and prepaid expenses, right?

You know, we can, we should debit that since it's in a credit mode and we can credit, you know, we could send a credit over to this.

And this is what the journal entry would look like, you know, prepaid expenses in the asset account and what is expected to end appearance with a debit balance.

It closed at -200 with no activity.

So we need to do something to be able to fix that, right?

And I think it's cool that it actually tells you why in real words, in real terms, right?

You know, here's another one that will, you know, you know, this one is this old account, it's a legacy account that we should just move it over to the, the cost of goods sold and be done with it.

And then that account can close, right?

And then here's my accruals that are missing.

And this is what we, you know, this is what we can usually expect of that.

And these are the entries right now.

The cool part is it's asking you, do you want to do this?

Do you accept this, right?

And I could right here or submit it for approval, right?

So it will actually then reach into JD Edwards and make those corresponding journal entries for you, right?

So you don't have to go back and forth between screens.

You can actually just send those journal entries right from here right into JD Edwards.

And then they would go through that normal approval and posting process just like any other journal entry would.

You have period activities right of, of each one of your accounts and you can of each one of your companies and accounts.

You can kind of see how these things are matching up.

And then any account movements and where your asset balances are, you can look at this from a consolidated standpoint.

You know, you can look at it from company, you know by from a company position and see which ones are blocked.

You know this one our, you know, our fifty, our company 50.

This is that $450.

Oops, that $450.

Sorry, I didn't want to go there.

That was the first time I've done this in the in, in all of the and all of the the webinars we've done so far.

So you can actually go see each one of the accounts in there, but from a company perspective and see it that way.

You can look at KPIs, right?

You know, ultimately, if you remember in the beginning I talked about these agents being very goal oriented, right?

That's what this is.

That's what the, these are the goals for this agent, right?

You know, you know, we've set a target that the child bounds preparation time for each company is under 15 minutes and we're hitting 8 minutes, right?

We, we set a goal that, that we should be able to touch this generate recommendations to solve problems above 95% and we're at 96%, right?

We want to get account and effort below an hour and we're coming in at 40 minutes, right?

And, and we want to be able to reduce spreadsheet use by 84%.

And here we go, right?

If, if any of these were below, they'd be highlighted in red and you'd see them and they would, you know, and put on a monthly basis that the agent would work towards that goal, right?

And to try to fix it and try to figure out what it could do better and what it could learn more and able to achieve that.

And the last thing I wanted to show you is that the, the, each agent, whether it's AP reconciliation or trial balance, has what's called observability, right?

This is part of our agent.

Again, this is not needed for the agent.

This is neat.

This is for you, right?

This is to be able to see everything that the agent's doing, right?

Where is it from an, an orchestration perspective, You know, what are the some of the the data objects that it's hitting, you know, with some configuration, what did it do?

How many you know, how often is it running, you know, and when it did, how many documents or how many anomalies did it find?

So you can see everything that's happening down to the second for this agent.

So you have complete auditability of every agent.

OK, I'm going to go back to my dashboard.

I'm going to pull up my presentation.

I need to get need to get rid of this one.

OK, where is my screen?

Oh, sorry.

OK, So you know, basically what did end up happening there is that, you know, we had a $450 difference that was traced to one line in a journal entry.

There was only one of the four companies ready to close because there was all sorts of issues, right?

You know, we had multiple accounts that needed to review.

We found 4 anomalies and we drafted fixes for each for two of them, we had 8 minutes.

We, our trial balance preps are down to 8 minutes versus, you know, 15 versus hours before.

And we were down to touch this race of almost 96%, right.

So, you know, the, the key is, you know, and, and, and one of the important parts to remember, and I know we're getting towards the end and I'm almost finished, I promise, is that we, this is the third agent in our kind of our, our financial digital workforce family, right?

We had, we had our AP agent, our reconciliation and now the, the child balance agent, right?

The, the interesting part is all of those work together, right?

The AP agent feeds the reconciliation agent.

The reconciliation is always looking at my 4 elevens versus my 9 elevens.

And now my, my trial balance is looking at my reconciliations and making sure everything's in line from that perspective.

But the key is that wherever we start, wherever we start with you, that first agent helps build what we refer to as our platform, right?

And it's, it's the JD Edwards agentic platform, right?

Not only just the connectivity into your JD Edwards environment, but it learns your JD Edwards environment.

It knows your companies, it knows your chart of accounts, it understands your business processes.

So that when we, you know, let's just say we started with the trial balance agent, if then you wanted to move to the reconciliation agent after, so you it supported it during the month.

Now that just the month end close, the reconciliation becomes so much easier to implement because it's just plugging into that existing JD Edwards agentic platform, right?

Our development organization has stated that the second, third, 4th and all down the line agents are 70% less than what the first agent was, right?

Because you only, we only have to build the JD Edwards agentic platform once and then each agent then can learn from it, right?

It can plug in, it understands the how it connects to JD Edwards, but more importantly it understands your JD Edwards, right?

It understands your business processes and you can kind of, you know, continue it there so that you can truly start building that kind of digital workforce, right.

You know, the ones we have today are AP agent trial balance, reconciliation, financial planning.

We have a bunch of distribution agents around sales order entry and inventory allocation or back order management.

We've got some community industry specific ones for the home building industry around communities and bids and home closing.

But like I said, you know, every agent after the first one lands on that existing foundation and can be plugged in very, very easily.

So what happens next, right in the in the last minute of our webinar is what we would love to do is work with you.

You know, this is one of the things that's kind of, you know, it's free.

We would work with you to give you an assessment of where you are watch agent would be more valuable and how we could help you.

And then we could actually work on a, you know, getting you a prototype of that agent so you could see the impact of your JD Edwards organization.

So that was a lot.

Thank you.

I appreciate your time and you know, if you have any questions on I can hang on a little bit and and answer any questions.

But you know, I mean, you know, really what we're trying to achieve here is help really modernize your JD Edwards really trying to help automate manual work and, and doing such with digital workers that can truly be part of your organization, that can learn, that can think and that can truly make, you know, high quality recommendations to really help your workers perform their jobs better.

So on that note, thank you so much.

If you have any questions, please let me know.

All right, Scott, good stuff coming from our agents and the AI platform.

I do not have any questions yet.

If anybody has any, please ask away.

You can use the chat.

Raise your hand either way.

OK, I guess not.

OK, I will send this recording out to you so you can watch it again or share it with your team tomorrow.

I do not see any questions, so we'll go ahead and end it here.

All right, thank you guys.

It was a pleasure.

This modernization series has been great and we'd love to having you.

So have a great day and we look forward to working with you.

Thanks.