As first‑time entrepreneurs coming from a pure engineering background, compliance was never our strongest muscle. GST had some visible initiation points - registrations, filings, portals, returns. We stumbled, learned and eventually self-taught our way through it.

But TDS? Not even close.

If you haven’t figured it out already, we are a B2B SaaS company with all our customers in India. Most of them are SMBs. For a long time, not one customer mentioned TDS. So we assumed that this wasn’t something we needed to worry about. That illusion broke the day one of our newly onboarded customers (one of the larger, more process driven), asked us whether they should be deducting TDS from our payments.

That single question sent us down a rabbit hole. We levelled up on TDS knowledge, and suddenly the world looked very different. It wasn’t just about our revenue anymore. We realised we were also required to deduct TDS on several of our own vendor payments. Many of these had slipped through simply because of the nature of the transactions. Even our consultant had overlooked a few, largely because they weren’t deeply exposed to digital‑first vendors like CPaaS providers.

One such vendor was Kaleyra - our SMS provider. Kaleyra explicitly asks about TDS applicability before accepting payments. Back in the early days, when our spends were small, we had consulted our CA and marked TDS as “not applicable”. Over time, our usage and payments grew significantly. We kept paying without deducting TDS.

When we finally raised this with our consultant, the issue became obvious in hindsight. We had missed TDS deductions. The outcome? Pay the TDS along with penalties and interest.

Tax Deducted at Source (TDS) is fundamentally an expense-side obligation.

When you incur certain expenses and make payments to vendors, the responsibility to deduct tax does not lie with the vendor, it lies with you, the buyer. You deduct the tax at the time of payment or credit, pay that amount directly to the government, and remit only the net amount to the vendor.

In practical terms:

Common expense categories where TDS becomes relevant include:

The tricky part is that TDS applicability is contextual. It depends on:

Unlike GST, invoices rarely state “TDS applicable”. The law assumes the payer will evaluate and comply - an assumption that breaks easily in modern, digital-first expense stacks.

Missing TDS isn’t just a bookkeeping issue. It has very real consequences.

If TDS is applicable and you don’t deduct it:

Worse, these issues usually surface late during audits, assessments, or when a counterparty explicitly asks uncomfortable questions.

Resolving missed TDS typically involves:

None of this is technically hard but all of it is painful, time‑consuming, and completely avoidable with early detection. In our case, a simple contextual check at the time of invoice ingestion could have flagged the risk.

The core problem with TDS compliance is not intent - it’s context.

TDS lives at the intersection of invoices, vendor history, business behaviour, and tax law. Humans usually look at these pieces in isolation. An invoice is reviewed for amounts and GST. Vendor history sits quietly in accounting software. Applicable sections and thresholds live in the Income Tax Act. There’s another dimension that makes this even harder for us: the law itself keeps changing. TDS thresholds, rates, and applicability rules are periodically revised through budget announcements, notifications, and circulars. A setup that was correct a year ago may silently become non-compliant today.

This is exactly where AI fits naturally.

A context-aware AI system doesn’t just read an invoice - it understands intent, context and patterns across time. Instead of treating each invoice in isolation, AI evaluates it the way an experienced practitioner would:

Instead of asking “Is TDS applicable on this invoice today?”, AI reframes the question to: “Given what this business does, how this vendor is used, and how payments are trending, is TDS something we should already be handling?” That shift from static checks to contextual reasoning is what prevents surprises like penalties, interest, and retroactive clean-ups. And that’s the lens through which modern compliance needs to operate.

This belief is central to how we think about Taxor. We see it as an AI-powered compliance layer that continuously reviews business transactions, applies current tax logic, and connects invoices with vendor behaviour and business context - so risks surface early, not months later.

For growing businesses, compliance shouldn’t be a reactive clean-up exercise. It should be something that runs quietly in the background, keeping you confident as you scale.