Fixing Tax Risk with Hyperbots Sales Tax Verification Co-pilot

Ayo Fashina:
Good morning. Good afternoon. Good evening. My name is Ayo Fashina. I'm a strategic partner to Hyperbots and some other startups and scale-ups globally. My background is in internal finance, corporate finance, and project finance, with an engineering undertone that runs beneath everything else. I'm currently a director at the US Development Finance Corporation. Before that, I was a CFO at a logistics startup operating in seven African countries, and before that, I worked in development finance, which I'm still involved in today.

Today, I am speaking with Shalini. I would like to introduce Shalini, the CEO of Ikigai Advisory. She's providing strategic business and CFO advisory services, focusing on sustainable growth and AI-driven solutions for the office of the CFO. You can think of Ikigai as a trusted partner to CFOs and their function within organizations.

Previously, Shalini was a CFO at a biotech startup called Evolved by Nature. She has held several leadership roles at Iron Mountain and Procter & Gamble (P&G) and has over 15 years of experience across biotech, consumer goods, real estate, SaaS, and tech. She brings expertise in P&L management, fundraising, M&A, and business transformation. She’s known for driving cost innovation, process digitization, and building high-performing teams. Please join me in welcoming Shalini.

Shalini Urankar:
Thanks. It's a pleasure to be here. Looking forward to it.

Ayo Fashina:
Thank you, Shalini. Today, we’re examining Hyperbots’ Sales Tax Verification Co-pilot and how it neutralizes indirect tax risks. Hyperbots has several co-pilots, but today we’re focusing on this one and sharing how it eases the lives of finance teams and CFOs, addressing pain points we've all faced.

Shalini Urankar:
I’m looking forward to that and happy to answer any questions you have.

Ayo Fashina:
Great. Let’s start with this: why do you think indirect tax compliance is still such a headache for many companies, especially in the U.S., where tax regimes vary by state, and even by country?

Shalini Urankar:
Great question. Even in organizations that invest in process improvements, indirect tax remains a pain point. Despite the tools available, compliance is still challenging due to fragmented systems, human error, and constantly changing tax rules.

Ayo Fashina:
Can you give us a few examples?

Shalini Urankar:
Sure. One example is stale or wrong tax rates; vendors often use outdated rate tables, leading to over- or underpayment, which then causes credit memos, manual corrections, and downstream reconciliation. 

Economic nexus rules are another source of confusion, as thresholds vary by state and it's unclear when you've crossed them. Then there’s exemption certificate handling, often very manual, living in email threads or shared folders, causing cash leakage when exemptions are missed. 

Lastly, audits are still manual, with errors slipping through and being caught months later, sometimes during the audit itself. All this results in penalties, lost time, and wasted energy on avoidable problems.

Ayo Fashina:
It does sound like a big burden for finance teams, requiring significant manpower to minimize errors. That’s why the Hyperbots co-pilot is so relevant, and it can truly help CFOs and their teams.

Shalini Urankar:
Absolutely. This is one of those low-hanging fruits finance teams should adopt to avoid manual work and take advantage of available technology.

Ayo Fashina:
So, what is Hyperbots’ Sales Tax Verification Co-Pilot? From a CFO’s perspective, how would you describe it?

Shalini Urankar:
I’d call it a 24/7 tax inspector working in the background to prevent errors before they hit the books. It uses smart architecture combining computer vision and language models to process invoices like a human, but faster and more consistently. It extracts all tax-relevant fields, including origin and destination addresses, and validates the data against a tax dictionary covering 4,000+ jurisdictions. It flags issues, explains them, and suggests corrective actions, like prompting the AP team to request a missing exemption certificate. It’s proactive rather than reactive.

Ayo Fashina:
So it's a tool to prevent errors before they even start?

Shalini Urankar:
Exactly. Instead of relying on sample-based audits, you validate 100% of invoices in real-time, reducing exposure, improving accuracy, and freeing up time for strategic work. It’s a game changer.

Ayo Fashina:
Can you walk us through how the co-pilot works end-to-end?

Shalini Urankar:
Sure. It begins by reading the invoice, pulling line items, prices, quantities, and origin or destination zips. It classifies each item as taxable or exempt, finds the correct tax rate using real-time jurisdiction data, checks for nexus thresholds, and determines whether the tax charged is valid. Then it explains any discrepancies and can even issue credit memos or journal entries. All of this happens before the data reaches the ERP.

Ayo Fashina:
And before anything posts to the ERP?

Shalini Urankar:
There’s a final checkpoint. Only fully compliant vouchers go through; anything with a tax issue stays in a queue for AP. This also helps maintain a clean audit trail, using hash chaining and analytic dashboards by state, vendor, or product. So not only are errors fixed, but you're learning from them too. It closes the loop, detects, explains, resolves, and prevents issues.

Ayo Fashina:
That’s critical. You don’t want errors discovered by auditors. This co-pilot acts as an in-house auditor, working 24/7, making audits smoother and faster. I remember that at my previous company, our audit for the previous year was still open when I joined, mostly due to manual processes and errors. With this, we could’ve closed audits in a month and moved on to strategic tasks.

Shalini Urankar:
Exactly. Smaller companies especially suffer because of lean teams. My advisory services are often brought in to fix these very audit delays.

Ayo Fashina:
Because we’ve historically thrown bodies at the problem, but humans make errors. AI can do it better.

Shalini Urankar:
Exactly.

Ayo Fashina:
So, what’s the biggest value proposition of this co-pilot?

Shalini Urankar:
It transforms tax compliance from reactive to proactive. It automates invoice-by-invoice compliance instead of sample-based reviews. It flags vendor mischarges before payment, preventing cash leakage, which is especially critical in this economy. And it simplifies audits with traceable actions and nearly 100% compliance, giving finance teams confidence.

Ayo Fashina:
That’s a strong endorsement.

Shalini Urankar:
It reviews every line, boosts compliance, stops vendor overcharges, and offers auditors a 1-click trail. It doesn’t get better than that.

Ayo Fashina:
With the co-pilot in place, who owns tax verification in an organization now?

Shalini Urankar:
It shifts significantly. What was once manually handled by AP or tax teams is now automated. The co-pilot extracts data, validates rates, manages exemption certificates, and even issues credit memos. Teams move from being processors to reviewers and exception managers, focusing on oversight and value-added work.

Ayo Fashina:
Instead of manual tracing and matching, they just review and ensure everything’s right, which is much more strategic. Can you walk us through the co-pilot’s main functional capabilities? 

Shalini Urankar:
Exactly. Now, its main functional capabilities are:

• Line item & address validation: Accurately calculates tax jurisdiction.

• Tax category classification: Maps items correctly to taxable or exempt categories.

• Live tax dictionary access: No more static tables. Now data is current and real-time.

• Mismatch flagging with explanations: Helps both correction and team learning.

• Autonomous actions with oversight: Can issue credit memos, post use tax, alert teams via Slack, but only after finance approval.

• Audit trail: Everything is hash-chained for SOX-ready documentation.

• Pre-trained models: It works accurately from day one.

• Multi-ERP compatibility: Works with QuickBooks, NetSuite, SAP, and more. One dashboard across systems.

Ayo Fashina:
That’s comprehensive. And even if it doesn’t integrate directly today, Hyperbots can make it happen in a few weeks. You can just integrate it into your existing ERP and talk to it.

So, I'm thinking, we've been discussing what the co-pilot can do and how it works. I'm sure there are other tools built into some ERPs or accounting software that can do part of these things. So, how do you think this co-pilot is different? What sets it apart from the typical tools we have currently? How is it different from traditional tax engines inside ERPs and the RPA tools currently available, in your opinion?

Shalini Urankar:
The difference is huge. Most tax engines use template-based OCR batch processing, and they require a lot of human cleanup after posting. But the co-pilot is a much smarter, much more integrated tool. It uses Vision-Language Models with LLMs, so it understands invoices, even if the formats vary. That’s a common problem, right? You're never going to have similar formats across all your invoices.

That’s usually the first problem you hit from a data perspective. So, the co-pilot is very smart. It can understand all invoice formats, and the format doesn’t matter. It validates in-line before vouchers are posted, not after, which is a huge compliance advantage. You’re continuously monitoring your economic nexus because it's built-in, not using spreadsheets that are bolted on. This allows for continuous monitoring around that piece.

It can automatically resolve mismatches. It doesn’t just flag them, but it can also resolve them by showing credits or posting use-tax entries. And then there's the audit trail hash chain, which isn't buried in PDFs.

And, like you mentioned, it’s very easy to roll out. Hyperbots was able to get it up and running in four weeks. That’s a big plus point, especially since, when I was a CFO, my team’s biggest issue was not having the bandwidth and time to evaluate, work through sandbox environments, and implement a new tool, all while doing our regular day jobs, which are already crazy enough.

So, this is a big plus. Hyperbots works with you to get these tools running on time, in four to six weeks, which is huge.

Ayo Fashina:
Definitely. Speaking of Vision-Language Models, when Hyperbots was developing this tool, I was working with them. We provided some invoices - some handwritten, some typed, all in different formats, just to train the model so it could understand various types of input. And yes, I can confirm that Hyperbots' co-pilot can read even handwritten invoices, not just typed ones. The Vision-Language Model is sophisticated enough to differentiate.

So, thank you! And if you would kindly just provide one key takeaway or one line about this co-pilot we’ve been discussing today, what would it be?

Shalini Urankar:
I might do two lines! I'd summarize it this way: Hyperbots' Sales Tax Verification Co-Pilot catches 99% of tax errors before cash moves. It erases penalty risk and shrinks manual review by 90%, all while giving auditors an immutable, one-click trail.

And if I were to wrap it up: Tax compliance shouldn’t be a burden. With the right tech, it doesn’t have to be. Tools like the Sales Tax Verification Co-Pilot help us shift from firefighting mode to proactive control, and that’s a huge win for finance, audit, and the business as a whole.

Ayo Fashina:
Definitely. If you can go from firefighting mode, which was my mode for the first year and a half in my startup, to productive use of the finance team, where you’re talking about strategy, controlling costs, and improving margins, that’s incredibly helpful.

I think, in this day and age, finance teams should leverage more of the AI that’s available. There are a lot of capabilities in AI now. Yes, tools have been developed over time, but what AI can do today is far beyond what most finance teams realize. If they explore and look at the different co-pilots, some offered by Hyperbots, they could make their lives a lot easier and allow their teams to focus on strategic work rather than mundane, error-prone tasks.

So, thank you so much for your time, Shalini. And thank you for breaking down Hyperbot’s Tax Co-Pilot for us.

Shalini Urankar:
It’s my pleasure. I think it's a huge win for the office of the CFO!

Ayo Fashina:
Great, and there you have it.