Sign in
Request a demo
Talk with an accountant.
Connect with us:
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.
Christine Butchko: Good afternoon, everyone, and good morning-ish if you're calling in from the Pacific Coast. Welcome to today's session, Demystifying ERP, AI, and Automation: What's Real and What's Just Hype, with HubiFi and Withum. To start, I have some standard housekeeping and NASBA compliance information for the CPE webinar.
This program is delivered group internet-based, with live instruction, polling questions and interactive Q&A. It's worth one CPE credit in the field of study accounting, and the program level is beginner. No prerequisites or advance preparation are needed.
Our official compliance statement: HubiFi is registered with the National Association of State Boards of Accountancy (NASBA) as a sponsor of continuing professional education on the National Registry of CPE Sponsors. State boards of accountancy have final authority on the acceptance of individual courses for CPE credit. Complaints regarding registered sponsors may be submitted to the National Registry of CPE Sponsors through its website, nasbaregistry.org. Our sponsor ID is on screen and will also be in our follow-up emails.
This program has no fee. For questions, cancellations or complaints, contact me at christine@hubifi.com. In accordance with NASBA, we keep responses, evaluations and attendance records for five years.
Four polling questions will be deployed at random intervals, and you need to answer at least three to receive credit. You'll find them on the right side of your screen in the box labeled Polls. They're sporadic and unpredictable to keep you engaged with the content. If you don't answer at least three, you will not receive credit for the program. With that, let me bring up today's speakers.
Lonnie Bloom, CPA: Thanks, Christine, and welcome, everyone. First, thanks to HubiFi and the team for putting this together. The topic is super timely, in a world of AI, automation and complete disruption across the accounting industry and every type of business. We're all asking what AI is going to do to our profession.
I'm Lonnie Bloom with Withum, where I've been for close to fifteen years. I'm in our technology and emerging growth sector, and I co-lead our e-commerce sector. I deal a lot with D2C brands and omnichannel sellers across the country and the world, plus SaaS and technology companies. I'll hand it to Wally and Jason for their intros.
Wally Merkas: Excellent. I'm Wally Merkas, a partner in Withum's management consulting group. My day-to-day is technology strategy, helping customers make more informed decisions. My experience in the marketplace is in ERP integration, which we'll talk about, along with accounting packages and CRM packages. I come at it more from the business process side than the accounting side, so I'm looking forward to crossing over with Jason and Lonnie on how these systems work in the marketplace. Over to you, Jay.
Jason Berwanger: Wonderful. Lonnie, Wally, Christine, thank you. I come from a corporate accounting background on the non-audit side, and I also have a technology background in data engineering and ERP implementations supporting accounting and finance. I've done six ERP implementations and advised on about fifteen others, on selection and business requirements. So this topic is near and dear to my heart. Lonnie, I fully agree: it's an important question people ponder often, as the world feels like it's changing. In many ways it is, and in many ways it's not. We'll get into it today.
Lonnie Bloom, CPA: Thanks for being here to talk about this. Clients come to us and say, "We're having a tough time with our close. It's taking longer than it should. Should we upgrade our ERP? Should we move to NetSuite or Sage? What are the next steps?" The real question is the diagnosis: what is actually slowing down your close? Some companies assume a new ERP is the next step, or that they need to jump into a new AI-native ERP. That's why we called this webinar "what's real and what's hype." Let's slow down and figure out what the decision process should look like. Do we need an ERP, an ERP-lite, or a patchwork of software? Jason, what do you see truly slowing down the close for your clients?
Jason Berwanger: The nuanced answer is: it depends. The symptom is almost always the same: a slow close, and tension between management, FP&A and accounting. The hard part is that the underlying disease differs, and the nuance comes from the business model.
To oversimplify, there are B2B business models with deep contract complexity, not a ton of customers, and high-dollar but low-volume contracts. I'm still a big believer that an ERP can be a great solution for the underlying disease there. You have a single place to manage contracts, or ingest them from a well-integrated CRM, and you manage orders, entitlements, billing and cash application. I've seen a lot of good come from ERPs paired with the right business model.
The other side is B2C or high volume. That's where the ERP has been decentralized into a CRM, a billing system, an order management system, and payment processors like Stripe and PayPal. Same symptoms, but a very distinct set of problems. The ERP ends up being a hub where accounting places manually curated information into a GL, and moving into an ERP in that situation can be a challenge.
I ran into this personally when I joined Root Insurance. We'd signed with NetSuite while still on QuickBooks, and we got NetSuite implemented fairly quickly, mostly on the GL and financial reporting side. Then the CFO and I talked about our success criteria: are we IPO-ready? My answer was that we got a better general ledger, but we didn't solve the problem. We weren't IPO-ready because the third-party systems where customers contracted had no connectivity to the ERP and GL. It wasn't a NetSuite problem. The GL only solved a small, commoditized portion of the problem.
Folks can blow a lot of money, and consultants can lead them in the wrong direction, by saying, "You're just one more customization project away from getting what you want out of your ERP." If you're that second business model, the underlying disease needs a much different treatment than the B2B model.
Wally Merkas: Jay, to pick up on that: the biggest mistake most implementers make, and I say this as an implementer, is not saying no to clients. Not having a backbone. I'll say in meetings that when my kid was eight and asked for a second scoop of ice cream, I didn't say yes. Sometimes we have to say no to customization and keep the ERP simple. Implementers start building and scripting logic, what I'll call undocumented features, throw all of them in, and that's what makes projects go sideways. Jason and I will come back to this later, but food for thought: customization is not always the right answer.
Lonnie Bloom, CPA: I love the ice cream reference. She's probably asking for that second scoop now.
The options: reconciling data from multiple billing platforms; manual journal entries and adjustments; waiting for data from other departments; audit trail and documentation; or other.
Lonnie Bloom, CPA: I'd add from that conversation that data fragmentation is a big issue. From an audit perspective we see it often: disparate systems and different data sources that all have to be reconciled at month-end. Instead of closing in two days, it takes two weeks or two months, depending on where the data comes from and who quarterbacks and reconciles it. Do you have the right people at your company to manage all those disparate systems?
Jason Berwanger: If you pick "other" or don't see what you're looking for, feel free to drop a question or comment in the chat, and we're happy to give our take on that driver.
Lonnie Bloom, CPA: Let's close the poll and break the month-end close into components: revenue recognition, inventory, accounts payable, payroll, consolidation and so on. Jason, Wally, where do you typically see the biggest pain point? Payroll? Fixed assets?
Jason Berwanger: I'll kick it off, and Wally, I'll give you the lion's share of the mic. For B2C business models, revenue recognition and AR are disproportionately the challenge, because that's where you have huge volume, lots of complexity and lots of systems. It's also where FP&A and management have their eyes. And from the AICPA's view of material weaknesses, about 25% of Big Four audits come back with deficiencies, and the number one reason is revenue. That's largely the complexity of ASC 606, then FX gains and losses, then CECL credit losses, all of which revolve around customer receivable and revenue data. Then there's inventory, Lonnie, especially in e-commerce, where cost of revenue has to tie back to the timing of revenue.
Outside of that, having seen enough of these projects: multi-entity consolidation combined with payroll accruals and cross-entity allocations can get nasty. You're often splitting a payroll dollar several ways between subsidiaries, sometimes under intercompany agreements, and consolidating up to a holding company. A lot can go wrong, and it takes both data work and accounting compliance know-how. The accounting team disproportionately bears that burden at month-end.
Wally Merkas: Jay, I think you're right. For me, the biggest issue is identifying where the source is: the upstream fragmentation you mentioned. That's where we spend most of our time. When I'm putting in an ERP and figuring out why the month-end close is slow, I start with an end-to-end view, an in and an out, and walk through everything. Most people look at the close purely from an accounting perspective and forget there are systems, people and process. Even handoffs, from CFO to controller to AR clerk, need to be looked at.
Sometimes it's a simple handoff, but often it's upstream fragmentation. Someone decided to use metadata tags in Stripe, and now trying to report on those tags in the accounting system can be crazy. And I agree on multi-entity and consolidation: add foreign exchange and consolidation challenges, depending on your complexity. People forget there are different standards across the world, even between the US and Canada. Now you have IFRS and GAAP, beyond "just" ASC 606, and I can't believe I said "just" ASC 606, which is complex enough.
So, Lonnie, to your question: I don't know that it's payroll or e-commerce platforms specifically. It's everything upstream of the ERP, and the ERP is the last place you close. That's where we see fragments all over the place, usually from over-customization and over-engineering.
Lonnie Bloom, CPA: From my perspective, revenue is always an issue, especially with multi-element contracts and multiple revenue streams. In e-commerce, inventory: you're operating with one, two or three 3PLs with different reporting, and you have to tie that into your financials, valuation and costing.
And I'll triple down on consolidation. The horror story: a multi-entity consolidation across foreign jurisdictions with different reporting currencies, all in different systems. QuickBooks in the US, Xero in Mexico, different software for an India subsidiary, all pulled together in Excel. Then one person runs that process, a controller, and that controller leaves the company. Now nobody has the instructions to do the reconciliation, and you're scrambling. I've seen that too many times. You have to peel back the layers of the onion to find the real issue.
Wally Merkas: And a little dig, not really at Jason: founders who don't get out of their own way. Sometimes a founder has done something the same way all along, and it's time to move to a more automated, streamlined version. People play a heavy hand here, and we often forget that when we try to clean things up and get the close faster and more aligned.
Jason Berwanger: Guilty as charged. Lonnie, there's an upvoted question in the chat I'm happy to take.
Question (anonymous): We're a high-growth SaaS company, and our transaction volume is making NetSuite incredibly expensive. What are our options?
Jason Berwanger: That's a loaded question in 2026. You have a lot of options. With high volume, many folks try to force-fit third-party systems into NetSuite modules, and that does two things. There's a heavy tax to maintain those integrations, and you're forcing third-party data models into the NetSuite ERP model. That's typically not a good trade-off.
It comes down to two options. First, use other technology to integrate with the third parties, and use NetSuite as your GL and hub for financial reporting: its intended purpose, rather than overextending it to absorb all the data. You'll save a lot on maintenance and on NetSuite volume fees. Second, there are many other ERPs for various business models, industries and sizes, if you have some ERP need and some B2B plus higher volume. The market has options that may fit better than overextending and customizing NetSuite. Using NetSuite as a general ledger for financial reporting has good ROI. But if you push transaction volume into its operational model when your operations are decentralized, you'll run into problems, and it will cost a lot.
Wally Merkas: Quickly on suites, Lonnie: NetSuite and most suites are a mile wide and an inch deep. I like what Jason said. You don't have to use any application, whether NetSuite, Intacct or Microsoft, all the way across. Adding something already built for that problem makes a lot of sense and reduces your overall technical debt.
Lonnie Bloom, CPA: Wally and his team field these kinds of questions all the time, and it's a good segue. As a reminder, if you have questions, type them into the Q&A and we'll try to answer them live.
How do we calculate true impact versus perceived impact? As we discussed beforehand: say closing fixed assets takes four hours, and a software solution gets it down to three and a half. Is that real value? You also get pitched software that automates something, but the automation requires a human in the loop who spends as much time as the manual process did. Can you talk through scenarios you or your clients have faced?
Wally Merkas: Jay, I'll start and you wrap. Someone will tell me they have a one-hour task in the month-end close, and they look only at that task, never at the full process in and out. That one-hour task may block ten people from accessing the information they need. So when you look at hours, whatever the rationale, maybe tying back to the contract on the revenue side, make sure you understand the upstream and downstream impact. That one hour may block more than you think. I have a few more thoughts, but Jay, pick it up and I'll come back with a scorecard.
Jason Berwanger: I agree. The slippery slope many folks fall into is weighing automation directly against only the hours attributed to the task, which is a version of what you're describing. The biggest value we see in our implementations is in revenue leakage, tightening up compliance, and audit fees. There's a lot of value there. And Wally, I agree that how many hours and dollars you're blocking downstream is one of those value buckets.
Often someone says, "This takes me five hours." But the real cost of that manual process is your time and intellect, an opportunity cost for you and for the FP&A team that depends on accounting's numbers to run the business better and see where it's losing and making money. We often have customers who find more revenue leakage and net income improvement in the first couple weeks of implementation than our total annual contract. Once great data is in the hands of accounting and finance, where they can segment and form hypotheses, a ton of value is created, and it has very little to do with the four hours the manual process took. It's about using your time and intellect strategically.
Frankly, a lot of accountants undervalue what they can do for a business. Nobody is better at translating operational transactions into financial information than the accounting group. That's what we're great at, collectively.
Wally Merkas: Agreed.
Lonnie Bloom, CPA: Let's move on to deterministic versus probabilistic data, and how that applies in practice. Jason, I know you've talked about this a lot. I think of AI and large language models as probabilistic. But in Excel, two plus two is four: deterministic. As auditors, we like deterministic solutions. What's your take?
Jason Berwanger: I think this is the main reason we haven't seen as much AI adoption in the accounting profession. Accountants are right: they're ultimately responsible and liable for the output of their calculations and reconciliations. The problem with AI is that it's a probabilistic, even stochastic, solution. By design, you can get a different answer each time. That's not a great foundation for accounting teams to rely on.
What we've seen work is almost a demystified version of SaaS and ERP: a central home for the information accounting has curated, placed where everyone can see it, which is the ERP ledger. But there are now two newer technology approaches people use to get the value they want.
The first is end-to-end deterministic automation. How do I attach every Stripe invoice, billing and payment transaction to the actual source of truth? Show me the needles that don't match. If you can make the match, create the entry; if you can't, tell me which ones failed. Then calculate the journal entries and show me a reconciliation proving you got it right back to the source system. Once you have that, you can use the middle layer on the slide, the automated journal entry output you review and use for your own analytics, and then layer AI on top.
So from a business model perspective, we put accountants in control of the deterministic data inputs: which policies and rules apply to the output. They stay in the loop to make sure reconciliations and alerts are in place. Now there's trust in the data. Then AI sits on top, where you can ask, "What's the history of this customer, who had five invoices over the last six months, one canceled and one disputed?" AI is great at reading that data, because you can always tie the answer back to a deterministic foundation.
As an accountant, you want auditability. From a data perspective, you want provenance and lineage: what source record was the accounting based on, what were the steps, and how did you get there? If you haven't built that middle layer and you just slap AI on top, you'll either end up with no adoption because nobody trusts it, or you'll trust things you shouldn't. You end up like the PwC report where half was right and half was hallucinated.
It's not sexy. It's very first-principles. First consolidate the data, then automate to produce deterministic information with lineage, provenance and reconciliations built in. Then AI becomes a great strategic layer on top for research, analytics and efficiency. The customers we've seen succeed took that approach, in that order, because it's a hard prerequisite to applying AI. Take the top-down approach, and "how do you know you can trust the output?" becomes a question no one can answer.
The options: revenue recognition and AR; expense management and AP; fixed assets; payroll accruals; leases and depreciation; or consolidations.
Lonnie Bloom, CPA: Going back to your point, Jason, about AI as the top layer: it makes a lot of sense. If you try to implement AI with fragmented data that isn't clean or unified, it won't work. We've seen clients change their AI strategy, but it all starts with a clean dataset. Once you have that, you can layer AI on top and do a lot of powerful analysis. I like the hierarchy you outlined.
Jason Berwanger: Funny side story. I was helping my dad find a truck, and we were shopping on AutoTrader, Carvana and the other new sites, which all have AI. We realized truck bed length often isn't a data input in the used truck inventory. We could use AI all day in natural language, but we couldn't find what he wanted, because the inputs weren't captured upstream in the first place. I said, "Dad, this is a perfect example of what we do all day." If features aren't tagged consistently, it doesn't matter how great the AI is. You can't get a good output if you can't walk it back to data you know and trust. That's a rudimentary example, but it's how I explained it to my dad last week. We still haven't found a truck, by the way. Still looking.
Lonnie Bloom, CPA: If we've got our polling answers in, let's skip ahead. This one's for both of you: when should you buy a third-party solution, and when should you customize? Wally, can you start?
Wally Merkas: Build versus buy is always a big conversation. The key is that most folks over-engineer their ERP. The ERP was never intended to do a lot of these things. Some areas you can handle right out of the box. When you look at what logic you need to enrich your data to get where you're going, always ask: has this been built before? Is there a third-party app for it?
I appreciate that a third-party add-on costs money. What people forget is that building software costs money too. Even if you sell software or have engineers on staff, people over-engineer. One of the first and biggest things I correct when I come in is removing scripts and technical debt and buying a third-party app that would have cost a quarter of the price. That's on top of a failed implementation, a struggling integration, or the application charging way too much based on transaction volumes, all of which would have been considered if it had been done elsewhere.
Succinctly: be careful what you wish for, don't over-engineer, and keep it simple. As you build your process, sometimes Excel is still an option early on, or there's a third-party add-on, so you don't do what most implementers do, which is over-customize the application. Jay, that's a good segue: with a third-party app, we can do much of the work there and post an allocation journal entry, or whatever we need, into the ERP for compliance, so Lonnie's happy when he's doing his audit work. What are your thoughts?
Jason Berwanger: I think that's right. I'd also separate SaaS that's just a new place to input something and integrate into your ERP. That probably won't add much value. A third party has to take a meaningful portion of your problem and actually automate it, and unfortunately a lot of SaaS doesn't. It's not that any third party is great. It has to solve the problem end to end. Otherwise it's just another so-called source of truth that isn't one.
The other part is first principles. You have to value the benefit to the organization from automation, compliance and analytics. And you have to measure the true cost, including the cost of customizing and maintaining. The most expensive things I've ever seen are bespoke build-outs of Salesforce, NetSuite and SAP: tons of customizations, multiple layers of systems engineers and analysts, and software vendors making a killing. One of the worst things you can do is underestimate what it will cost to build and maintain a customization.
The final calculation is: what's the value, what's the risk we're mitigating, and what will this actually cost us? More often than not, customizing and replacing are prohibitively expensive, so you need a tremendous potential business value to justify it. It's not sexy to say, but maybe you boosted morale with cool new tech, yet did it meaningfully improve the business? Often it doesn't. If you work backwards from value and put an appropriate cost on creating and maintaining customizations, you have your best shot at a clear success measure for a project like this.
The options: current systems can't handle transaction volume; poor integration with billing platforms; lack of real-time visibility into revenue; manual workarounds are breaking down; or not considering an ERP change.
Lonnie Bloom, CPA: On build versus buy, I agree with everything you both said, and I'd add the people aspect. Whether you implement an ERP or a customization, there's always a human in the loop for reconciliation and financial reporting. Do you have the right people to manage the implementation and the change? There's real change fatigue if you're implementing multiple systems. Are people trained, and how much training will it take? There's a cost there. We've seen many ERPs fail not because the software isn't great, but because the company didn't have the right people to manage the process, or to learn and use the software for financial reporting.
It looks like a mixed bag of responses, with poor integration with billing platforms leading as the driver of interest in ERP customization.
Lonnie Bloom, CPA: Let's move to the next section: story time. What happens when we get it wrong? I'll kick off. As an audit partner at Withum, I work with clients across industries and verticals in our technology sector, many raising their next round of capital to level up. They've grown significantly and are going to market for VC funding, a bank loan, or an acquisition and exit.
My horror story: a client went from about $10 million in revenue to $100 million in one year, in a booming market, and wanted to exit by getting acquired. They needed two years of audits. We said a clean audit could take eight to ten weeks, and we could expedite it, but they needed clean data, reconciliations and an audit package. We sent our audit request list, about 100 items, and waited weeks and weeks while barely getting a single thing.
We realized they had data fragmentation and reconciliation issues. They'd send revenue schedules, weeks would go by, and nothing reconciled. Meanwhile, the market turned. Their industry wasn't hot anymore, valuations dropped, buyers lost interest, and the plan for a big exit and payout was gone, all because of data fragmentation, ERP issues and reconciliation problems. We tried to help them through it, but as auditors, our hands were tied at a certain point. That's a horror story of what can go wrong. Wally, can you share one, and then Jason?
Wally Merkas: Mine is about over-customization, back to that logic, and similar to Lonnie's, this company was working through an exit. They had heavily customized the ERP and a front end. They didn't use the front end normally: they created custom fields, didn't even use the shopping cart the standard way, and added all kinds of metadata tags. When it came to reporting on those tags, two things happened.
First, the engineer responsible for the customization had left about six months earlier, taking all the knowledge of the undocumented feature set in the front end. Second, the back end had a matching pile of custom scripts. We had to extract and reverse-engineer them. For a frame of reference, that cost about $45,000 just to reverse-engineer what I'd call basic scripts, not even complex ones. In some respects, it's better to shut them off and rebuild from scratch.
That's why I emphasize people, process and technology, and not over-engineering. Don't turn your company into a software company on your platform. Let the platform do what it was intended to do, don't over-customize, and use build versus buy. Jason, what have you got?
Jason Berwanger: Well said. I'll take a different angle. We often see great stories from companies that invest the right way, and it frequently shows up as identifying fraud. A customer that implemented in Q4 found over $1.5 million a quarter in fraud. Customers were signing up for free trials, canceling, getting credit tokens, and repeating it over and over, until there was a ton of free service that was ultimately written off. Catching it was a big benefit to the bottom line, along with better accounting practices. There are real dollars to be found from better managing accounting and finance data.
I'd take it a step further: some of the best strategic finance stories I've seen started by solving the accounting problems first, which enabled strategic finance and leveled up both people and teams. At a number of my fractional clients, and at places like Root, we had revenue and customer data more often than a monthly close that looks back thirty or forty days. We'd find pricing or billing issues, and instead of catching them weeks later as a $5 to $10 million problem, the accounting team caught them quickly as a few-hundred-thousand-dollar problem. When you automate the right way, you get an opportunity benefit on top of the automation benefit, and the accounting team looks like rock stars, finding real value for the business, which is what they're best at.
On the bad side, we've all seen a bad quality of earnings destroy value, like in your audit story, Lonnie. And there's the notorious Bird scooters case. It's simple: customers bought gift cards and credits up front on a usage-based model, and the company didn't properly track them as deferred revenue, taking them inappropriately as revenue. How many billions of dollars were destroyed for lack of investment in accounting systems? Even if you're a pessimist, good revenue and financial reporting practices are an insurance policy protecting the value of the company, at 100 times revenue. I've seen both sides, and there's a big gap between the value destroyed by underinvestment and the value created by doing it right.
The options: audit current data flows before considering an ERP replacement; schedule a conversation about automation with our existing ERP; continue evaluating ERP vendors; assess whether AI tools could help with specific reconciliation tasks; or other.
Lonnie Bloom, CPA: Love it, thanks for sharing. Those horror stories and success stories make me think about how to measure ROI when evaluating software. It's not only time savings. It's cost savings, risk mitigation and revenue generation. We do a lot of automation and evaluate tools for our own service lines, and that's how we think about it.
It looks like "assess whether AI tools could help with specific reconciliation tasks" is the front-runner, at 55%.
Lonnie Bloom, CPA: Let's close the poll and move on. Wally, Jason: what can teams do today? We've talked a lot about software, automation and resolving pain points. Is there a checklist, or a few things teams can do to make these management decisions? Jason, I'll start with you.
Jason Berwanger: Sure. I'll give a quick summary, then give Wally most of the time, since this is deep in his expertise. The two main things to get right:
First, understand your data flows: where your customer and financial transactions actually happen. If you have a fixed asset ledger, those were purchases at some point, so where does that happen in an upstream AP process? Where's your revenue, invoicing and payments? Do you have a dispute management platform? Identify every part of your transactional flow.
Second, understand the business model and the accounting policies. At the intersection of those two, the business model and policies on one side and the data sources on the other, you can get a great assessment. Where do I have problems? Where do I need to invest in automation? Where do I have compliance gaps or missing policies and documentation? Where is the data in these systems not great, so it just takes more time? That's an operational tax accounting pays. Maybe everyone loves the AP or expense management system, but it's really hard for accounting to do accruals, AP reconciliations and the fixed asset schedule from it. Get the data inputs and their intersection with the business model and policies right first. Even if the accounting team starts on the wrong foot, with those two worked out, you'll get to a great outcome.
Wally Merkas: I like that. Let me make it a bit more tactical. I see folks nervous about building a checklist and deciding where they want to go: they're scared to dip a toe in the water. Start small, and think about what your advanced technology and process improvements could be; I won't just say AI. Maybe start with ten transactions and, per Jason's point, tie them back to the contract and through to cash. Walk through that flow as a group. Have pizza Fridays and walk the process together, so you're actually looking at your own data, not following a generic YouTube video.
That's one of the biggest issues. I'll get asked, "I saw this YouTube video, or this is from the help files of some ERP, so why isn't it working for us?" Then I look, and their process is nowhere near it. Yes, they recognize revenue, but their process may be completely different. Establish your own approach and understand what you're doing. Beyond policies, which matter a lot, and remember I'm speaking to a room of accountants as the tech guy, there should be a tactical checklist and process that gets you through it.
And keep an error log. I liked what you said about dispute resolution, Jason. I could count on one hand how many folks have a dispute resolution step in their regular process. Instead they assume the month-end close is off and go chase $11 without a conscious thought process for how the close works.
On the automation and advanced technology side: try it, see what's going on, and test it a bit. I'm not pitching AI; I'm pitching advanced technologies. It's much more mainstream than we think, and I still see a lot of people scared to dip a toe in. So dip small and make mistakes. If you haven't made a mistake this week, you haven't tried hard enough. Then keep working through it.
Lonnie Bloom, CPA: Love it. I like the hierarchy: automation before AI. AI is everywhere now, a shiny thing in a box that nobody really understands, and everyone has FOMO and jumps to "I need AI in my organization." We have to pause and assess: what's my issue, what's the real pain point, and what's the solution? AI won't solve every issue. Sometimes it's automation, or advanced technology that's deterministic, not a large language model sitting on top. And it's having the right team members, processes and checklist in place. This checklist is a good takeaway if you're listening in, to make sure you're assessing appropriately.
Lonnie Bloom, CPA: Let's go to the next section, key takeaways. We covered a lot today. I mentioned not letting AI buzzwords drive decisions, but I'd love to hear your takeaways. Jason, let's start with you.
Jason Berwanger: When I think about this, I think about great leaders and great accounting and finance practitioners, and especially the second bullet: constraint breeds innovation. You can really overhire a team. Without constraints, the team isn't very innovative, and the cost to run accounting scales proportionally with the business. As stewards of the company's finances and capital, great leaders don't want to do that or set that example.
But it swings the other way too. I've seen folks massively overbuild software internally: an internal ledger, an internal billing system. That helps the finance and accounting team, but weighed against the risk of maintaining controls around it, it's really expensive, including in opportunity cost. Why aren't your engineers innovating for your end customers instead of for your one-of-one back office? That innovation curve isn't in your favor.
On top of that, the most expensive and challenging part of the Root IPO wasn't building the tech to connect third-party systems into NetSuite. It was maintaining the controls. How do we prove we got everything? How do we retain everyone who worked on the projects? Like Wally said, when someone who built something leaves, they were often also the owner of that control and the control review. That's a huge tax when you build internally. So we're big believers that constraint drives innovation: you're not overhiring, overbuying software or overbuilding software. It keeps you grounded in first-principles problems, with a great team and the right tools. Lean too far in any direction and you end up overbuilding or overhiring, which is never a great place for a finance or accounting leader who's supposed to set the example for the org.
Wally Merkas: Well said. I'm the tech guy, the ERP guy, so I have to say the ERP is your system of record. But I'm not saying the ERP is where you start and build everything. Your architecture determines whether it stays clean or becomes a dumping ground, so be careful what you wish for. If it's your system of record, good hygiene matters.
On my side, we tend to use AI for discovery, to find where problems might be. I love where Jason and Lonnie went today: it can help you rescue fragmented, inconsistent data, so pay attention to that. And I'll leave you with this, as I pretend to be an accountant: a fast close might be a risky close, so be careful. The goal is a close you can explain, defend and trust. Jason touched on that earlier. Those are the things I'll leave you with.
Lonnie Bloom, CPA: Great. We'll go to Q&A in a second, so if you have questions, put them in the Q&A chat. Wally, Jason, any call to action for our listeners?
Jason Berwanger: There's a ton of great, innovative tech out there, but shiny object syndrome is tough. I follow the space every day, and a lot of what's marketed as ERP or AI is neither. It might still deliver real business and accounting value if that's the problem you have, and it may be a good solution. But don't fall in love with options based on how they're marketed or because they have AI natively built in. Most companies marketing that were founded before AI, as we think of LLMs and agents today, was known and commoditized. It's riskier than ever and harder than ever to shop for solutions. It's worth keeping the first-principles approach so you don't end up saying, "I chose this because it was AI-native," when it didn't actually solve your problems. If it didn't, the trade-off we discussed won't materialize, and you won't get a great outcome for your team or organization.
Wally Merkas: For me: transaction volume is real. That's probably the number one area where I see people make mistakes with ERP, regardless of vendor. Challenge your ERP vendor on what volumes are acceptable and when your rate goes up. Make sure you truly understand what it will cost you. We're seeing a lot of people move to subledgers, data warehouses and integrations. In healthcare, for example, we have clients with EHR and EMR systems, even health tech, who move that data into a data warehouse and then post an allocation journal entry back into the ERP for compliance. Think outside the box. Don't customize. Don't code first. Think it through, then move.
Lonnie Bloom, CPA: I'll add one from CPA.com, which put out guidance on five types of AI and how to evaluate software. The five include generative AI and agentic AI, but also plain advanced technology that isn't really AI. So when evaluating software and you meet that ".ai" company, ask: what's the AI? What technology actually drives your solution? Is it probabilistic or deterministic? Ask those questions and be skeptical as you evaluate. Really great conversation.
Lonnie Bloom, CPA: We have a couple of minutes for questions, so I'll read a few from the Q&A, rapid fire.
Question: What kind of close acceleration is actually realistic?
Jason Berwanger: I'll take that one. Realistically, you can accelerate almost any part of the close. The question is whether you should. Revenue and customer financials are great opportunities, because of their analytical value, their compliance value and the opportunity cost model. We often get folks to a daily automated close on the revenue side, with a broader monthly close that layers in cash reconciliations, commission amortization against revenue and so on, to tie in final gross profit. So there's still a monthly close, and you can realistically pull it to day zero or day one. To run the business you need financial information that ties back to the P&L, almost a mid-month P&L, and the revenue side is worth that trade-off.
What's not realistic is accelerating consolidations and multi-entity work with multiple prerequisites, where you need the final payroll number, third-party input from the business, or accruals that depend on other parts of the business. There's a lot of human input and judgment there, and I don't think you can get those to daily. You might take them from fifteen days down into a five-day close. That's realistic, but probably not much further, given the prerequisites, the judgment and the number of business partners involved.
Question: At what stage or maturity level should a company make the close a focal point?
Wally Merkas: Honestly, from well-funded pre-revenue all the way up. If you're not thinking through your close in some way, shape or form, you're missing something. I'm entrepreneurial, and Jason obviously is too. Lonnie, in our roles now, you and I have to be less entrepreneurial and more tactical, giving strategic advice on where clients are going. So everybody needs to think about it. If you want to go fast and you're not paying attention to your close, you may be missing valuable insights. Maybe your business model doesn't match your operating model, or things aren't connected. Some people see the month-end close as a statutory reporting mechanism, but it's also an operational reporting window where you can look back and catch things before they bleed on too long. In my opinion, you should start looking at the month-end close, at least as a process, very early.
Lonnie Bloom, CPA: Awesome. I think we're out of time. I really appreciate the conversation. Christine will come back with the final CPE housekeeping. Thanks, everyone, for your time and for listening in.
Christine Butchko: Thanks, Lonnie, Jason and Wally. That was really informative, and I certainly learned a lot. A bit of housekeeping before we wrap up: to receive CPE credit, you need to fill out the evaluation form. I've put a pop-up below where you can click in and complete it. We'll also email a link in about two hours, so if you miss it, it'll come to your inbox. You have until next Monday to complete the evaluation, and we'll send out CPE credits by February 17. If you have questions, reach out to me at christine@hubifi.com.
To reiterate: participation is tracked through the live polling questions, the session evaluation and Q&A. Once we've confirmed everything, you'll receive your one CPE credit, and certificates will be issued no later than a week after this event.
Thank you again for coming. We're hosting another CPE session with our head of product experience, Cody Leach, in March, on March 11, I believe, and you can sign up on our website. We hope to see you at the next one. Have a wonderful rest of your day.
.jpg)
.png)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
.jpg)
