Transforming accruals with Hyperbots Agentic AI co-pilot

Brian Kalish:
Well, good day. Good evening. Good morning. Wherever people may be in the world. Today I'm Brian Kalish. It's a pleasure to be with you here today. We're going to have a fantastic conversation. The topic for today is Hyperbots and accruals — the accruals processing copilot.

Quick background on myself: I’ve been involved in the FP&A, treasury, and investor relations space for the past 30 years. I’ve been partnering with Hyperbots over the last two years. And today, we have just a wonderful opportunity to dive a little deeper into accruals processing.

With us here today is Mike Vaishnav. If I may have mispronounced your name, I apologize. He’s one of the founding design partners and a seasoned CFO based in the Bay Area, here in the United States. He brings with him very deep expertise across global manufacturing, technology, and service sectors, having held senior finance leadership roles at Synex, eBay, Kativa, and Direct Line Global. His background spans FP&A, M&A, treasury, tax, and investor relations. We have very similar backgrounds. ERP systems — he currently advises startups and VCs, including Hyperbots.

Mike, a pleasure to be with you today.

Mike Vaishnav:
Happy to be here. Thanks for the introduction.

Brian Kalish:
Absolutely. I didn’t know if you wanted to add anything else to that.

Mike Vaishnav:
Enough.

Brian Kalish:
Great. So, you know, our audience really wants to understand how Hyperbots' invoice processing copilot is not just changing, but truly transforming the AP — the accounts payable — function.

Let’s jump right into the details. One of the challenges I’ve seen, and I’ve had the good fortune of traveling around the world, is this:
Here we are in 2025, Mike — why are month-end accruals still a challenge for finance teams?

Mike Vaishnav:
Yeah, so it all depends on how you set up the accrual process for the company. Quite a few companies still use manual, Excel-driven processes.

At month-end — let’s focus — there are multiple types of accruals, but I’ll focus on just two:
One is related to inventory or product, and the other is non-product or non-inventory, basically, services.

For product and inventory, the process can be relatively easier because you have a goods received note. When you receive goods, an inventory entry is made, and the system logs a “received but not invoiced” transaction. So that shouldn’t have a significant impact.

But the major challenges appear when you’ve received goods but don’t have a goods received note or any other documentation. That’s where creating the accrual becomes problematic.

Also, quite a few companies use Excel for this process. Some ERP systems handle it, but challenges remain, such as transaction volume and accrual timing. Many companies aim to close their books within the first 3 to 5 days of the month and finish accruals in 3 days. So those are tight cutoff dates.

You’ll see even more challenges with service or non-inventory items, like legal, accounting, or consulting services. You need to coordinate with the respective department to figure out how much service has actually been received. Sometimes there’s a blanket PO, or the PO won’t specify how much service has been delivered. And often, invoices haven’t yet been received.

Those are the more challenging types of accruals.

Another challenge is audit documentation. Auditors look at accruals in great detail. They don’t like broad, generalized accruals.

So these are some of the current challenges companies face — unless they have an automated process in place, whether through an ERP or an AI-based system.

Brian Kalish:
Absolutely. I mean, again, to me, it's a fascinating time period, right? Because — if I may be so bold as to think that we're contemporaries — you know, we've spent the last 30 years really dealing with this problem day in and day out. And I think we're truly in an environment where, with the advantages of AI, we’re kind of living — I don’t know if you would agree or not — we’re kind of living in the industrial revolution in real time. I mean, what we're talking about is truly transformational.

Brian Kalish:
With that said, I think one of the things the audience really wants to understand from you is: What is Hyperbots Accrual Copilot? I mean, really — what does it do for us?

Mike Vaishnav:
Hyperbots provides a pretty unique solution with the Accrual Copilot. Of course, there are other companies doing similar things, but if you look at Hyperbots focus, the Copilot calculates accruals automatically. It does this by analyzing outstanding invoices, purchase orders, and goods receipt notes. Based on that data, it performs automatic accrual calculations.

It also continuously learns from prior accruals or past forecasting experience, and it checks whether the accruals have been done correctly. It predicts accruals based on trends and seasonality, and raises questions if something looks off. For example, if there’s a significant variance from month to month, the Copilot highlights that and provides guidance through a chatbot or messaging system to ensure the accruals are handled properly.

Another unique feature is the agent workflow for non-inventory accruals. Traditionally, people would have to reach out to various departments — say, for consulting or legal services — asking, “Hey, how much service have we received?” Hyperbots automates this by sending messages or allowing users to interact via a chat interface, speeding up the process significantly compared to waiting for emails, phone calls, or documentation.

Brian Kalish:
Yeah, that makes sense.

Brian Kalish:
I think one of the ways people can better understand this is if you could map out the end-to-end flow of the Copilot.

Mike Vaishnav:
Absolutely. So, not in a strict order, but let me walk you through it.

First, it uses configurable policies for both goods and services, including recurring items and cutoff rules. Based on those policies, it guides what needs to be accrued.

Then there's the smart accrual engine. This handles daily, weekly, and monthly auto-adjustments. It considers timing differences, like 4-4-5 calendar types or weekend cutoffs, and adjusts accruals accordingly to make sure timing is accurate.

Next, it performs accrual discovery — looking at goods receipt notes, recurring non-PO scenarios, and pending invoice cases. It also handles dynamic GL coding, making suggestions based on historical data.

There are automated booking and exception workflows, with real-time notifications. Plus, it supports hashed and auditable records, which is critical for compliance and transparency.

These are just some of the unique features that make sure all accruals are processed properly and efficiently.

Brian Kalish:
No, thank you. That’s really helpful — I’m kind of a visual person, so it maps it out for me visually. I appreciate that.

Brian Kalish:
One of the things I also hear — and what’s fantastic about this — is that whether it's here in North America, in India, or in the Middle East, it’s not just proof of concept. It’s really proof of value. So it’s not PoC; it’s really PoV. With the Hyperbots Copilot, how would you describe what the real value proposition is here?

Mike Vaishnav:
I would define that in two parts — the intangible benefits and the tangible benefits. Of course, it achieves a lot of operational efficiency using the Hyperbot's Copilot, but we shouldn’t always just look at dollars and cents. Let’s first focus on the intangible benefits.

One of the key things is reducing human error. In the accounting field, accrual accuracy is critical. Manual processes are prone to errors, but with the Copilot, the error rate is significantly lower because it follows consistent trends and logic to identify potential issues early.

Auditability is another major benefit. Auditors are always looking for a clear audit trail, and the Copilot provides a solid, traceable history for each transaction. Then there’s timing — this process is much faster. What normally takes two to three days can be completed far more quickly, whether for inventory or non-inventory items.

Thanks to the agent workflow, the process is far more streamlined. And during month-end close, which usually feels like a fire drill, the automation means fewer last-minute scrambles. There are a lot more intangible benefits I could talk about, but these are some of the key highlights.

Now, when finance professionals evaluate any system, they look at the cost. No one wants to spend money on implementation unless there’s a clear return. What we've seen is that companies using the Hyperbots solution have experienced significant improvements in processing efficiency and a reduction in costs.

There are also control-related benefits, which are harder to quantify but definitely impactful. So if I were to sum it all up in a sentence or two, the Copilot delivers greater operational efficiency, and that includes both intangible benefits and measurable cost savings.

Brian Kalish:
So the Copilot identifies every single scenario with AI, wiping out over- and under-booking. It’s interesting — it’s almost like the old Pareto rule, right? It takes care of the straight-through processing for about 80%. And as we move into the next question, I think that’s one of the key considerations people need to have. This isn’t a replacement for everything. It’s about taking out 80% of the routine tasks so we can focus our people on the 20% that really matters.

With that in mind, how does that shift take place from an ownership perspective, from before to after implementation, in terms of what the AI worker agent is doing versus what humans were doing?

Mike Vaishnav:
At a high level, it’s AI automation. But step by step, let’s break down the processes — whether it’s accrual runs, matching goods receipts, or service POs — these were all manual and time-consuming tasks. AI automates all of this and makes the process more efficient.

The AI agent workflow also improves communication. Instead of endless back-and-forth emails confirming whether tasks are done, the Copilot automates that entire process, making it significantly faster.

Brian Kalish:
Excellent. So we’re talking about gains, whether tangible or intangible. What are the real key functional capabilities that are powering those gains through AI?

Mike Vaishnav:
There are several important functional capabilities. Let me highlight a few key ones:

• Configurable policies and cutoff cadence based on company-specific rules for accruals.

• Discovery of goods, services, and pending or recurring invoices.

• Dynamic GL coding.

• Automated booking and validation of postings.

• Real-time reversals — although many ERPs already handle that, so I see that as secondary.

• A unified dashboard with reach notifications replacing manual portals.

• A high-standard audit trail.

These capabilities are critical in driving operational efficiency through the solution.

Brian Kalish:
Great — that really clarifies things. One of the big concerns people always have with new tech is how it integrates into their current tech stack. So what’s the ERP integration experience been like in your work?

Mike Vaishnav:
Specifically for the Hyperbots AI implementation, several companies are already adopting it. For example, NetSuite can be up and running in 8 to 10 business days. For SAP and Dynamics 365, we offer connectors and drag-and-drop mapping.

This is an add-on solution — not a rip-and-replace approach. It can be easily bolted onto existing ERPs and customized to individual requirements. It’s not a one-size-fits-all product. Depending on the ERP system, implementation can take anywhere from a week to a couple of months.

Brian Kalish:
That’s really important, because people worry that they’ll have to tear out their existing systems. But just to confirm, this is an add-on, right?

Mike Vaishnav:
Yes, exactly. This is an add-on solution. It’s not an ERP replacement. You still need your ERP for core transactional processing. What Hyperbots provides is an efficiency layer that automates and enhances those processes.

Brian Kalish:
That leads into my next question: Why Hyperbots versus legacy methods that already exist?

Mike Vaishnav:
In one word: operational efficiency. Hyperbots offers greater efficiency, cost savings, faster processing, improved audit trails, and reduced workload for teams by eliminating manual tasks.

Another key difference is the pricing model. With many ERPs, you’re dealing with license-based costs — buying a certain number of seats. With Hyperbots, it’s a pay-per-transaction model, with no license or seat fees. That’s a major advantage for the Hyperbots Copilot.

Brian Kalish:
No, thank you, because some of the numbers that I've seen — again, it's about two weeks to replace the spreadsheets, two days shaved off the close calendar, variances now under 5%. I mean, that's really the Hyperbots delta. That’s what I'm seeing.

Mike Vaishnav:
I don't like to use specific numbers because results can vary — some companies have seen significant efficiencies, and others depend on timing and implementation. But overall, from what I’ve seen with Hyperbots, organizations are achieving much greater cost and operational efficiency, better timing in processes, and stronger audit trails.

Brian Kalish:
Well, Mike, I really appreciate your time today. I know the audience certainly appreciates it. You’ve shown us how agentic AI not only slashes costs but brings what’s really important to the finance office — precision and calm. We talked about the challenge of bringing calm to month-end, and this addresses that.

Thank you again for detailing how the Hyperbots Accruals Copilot is able to achieve this. Any final thoughts you’d like to share with the audience before we wrap up?

Mike Vaishnav:
Now we’re in an era where automation is essential. We need things done quicker and more efficiently. It's not about eliminating all manual processes but optimizing them. With the help of technology and AI, we can make that happen — instead of working like we did in the late ‘80s, ‘90s, or early 2000s. This solution makes month-end much smoother and more efficient.

Brian Kalish:
Great. Well, again, thank you so much for your time today. I know it was beneficial for me and certainly for the audience. With that, I wish you and our audience a good rest of your day, evening, morning — wherever you may be in the world. Take care, good day.

Mike Vaishnav:
Thank you. Glad to be here.