The coverage
On 25 June 2026, News18 Hindi published a report from its Delhi desk about an AI tool that reads and verifies bills sent over WhatsApp. The tool is Zoffer. The interview was recorded at the inauguration of the Atal Incubation Centre at Netaji Subhas University of Technology (NSUT), New Delhi, where Zoffer is incubated.
| Publication | News18 Hindi (Delhi / Local18) |
| Published | 25 June 2026 |
| Reporter | Anjali Singh |
| Language | Hindi (with video) |
| Headline | बिल की हेराफेरी पकड़ेगा यह AI टूल, WhatsApp पर फोटो भेजते ही सामने आ जाएगी ओवरचार्जिंग और टैक्स की गलती |
| Headline, translated | "This AI tool will catch bill fraud — send a photo on WhatsApp and overcharging and tax errors surface instantly" |
| Read it | Full report on hindi.news18.com |
The core of the report, in the reporter's own framing, is that checking a bill no longer needs hours of work — whether it is a restaurant bill or an invoice worth crores. Send the photo, and the AI returns the GST, tax and arithmetic problems it found.
“आप बस अपने बिल या रसीद की फोटो व्हाट्सएप पर भेजते हैं. ज़ॉफर आर्टिफिशियल इंटेलिजेंस की मदद से उस फोटो को पढ़कर अपने आप अकाउंटिंग और बही-खाते की एंट्री बना लेता है.”
— News18 Hindi, 25 June 2026. “You simply send a photo of your bill or receipt on WhatsApp. Zoffer reads that photo using AI and automatically creates the accounting and ledger entry.”
The claims, checked one by one
News coverage compresses. We have taken every substantive claim in the report and put our own assessment next to it — including the two we would phrase differently.
| What the report said | Verdict | The fuller picture |
|---|---|---|
| You send a bill photo on WhatsApp and the AI creates the accounting entry automatically | Accurate | This is exactly the product. Photos, PDFs, voice notes and plain text all work. The drafted entry then waits for human approval before it posts. |
| It surfaces GST, tax and calculation errors, and possible overcharging | Accurate, with limits | Zoffer checks a bill against itself — line items versus total, GST rate versus line value, duplicated rows, tax applied where it should not be. It cannot tell you whether a vendor's price is fair. Judging that means knowing the market, which is a different question from reading the document. |
| The merchant does not have to type anything manually | Accurate | Capture is typing-free. Review still takes a person: someone reads and approves each entry, which is the point. |
| NSUT's Atal Incubation Centre built the tool | Needs context | Zoffer is built by Zoffer Innovations Private Limited, an independent DPIIT-recognised company. NSUT's incubation centre hosts and supports us — which is why the interview happened there — but the product is ours, not the university's. |
| The tool gives around 90% correct results | Needs context | That was a founder's field estimate at the time, for field-level extraction across typical Indian bills — crumpled thermal receipts, phone photos, mixed-language invoices. It is not an audited benchmark, and accuracy moves with image quality and invoice format. Treat it as an indication, not a guarantee. |
| Traders are using it in large numbers | Directionally true | We had real paying users at the time and have more now, but no verified user count was published and we are not going to invent one here. If you want the current number, ask us on WhatsApp and we will tell you honestly. |
| The tool was designed to save merchants time and money without technical knowledge | Accurate | This is the design brief, stated plainly. If a shop owner can forward a message, they can use Zoffer. |
How the bill check actually runs
The report describes the outcome. The mechanism underneath it runs in four steps.
- You forward the bill on WhatsApp. A photo of a paper bill, a PDF invoice from email, a voice note describing a cash expense, or a line of text. No app to install, no login at the moment of capture.
- The AI reads the document. It extracts vendor name, invoice date, line items, taxable value, GST rate and amount, and the total — and while doing so, checks those numbers against each other.
- Problems get flagged. If the line items do not sum to the stated total, if GST is charged at a rate the line value does not support, or if a row appears twice, Zoffer marks it rather than quietly writing a clean-looking entry over a dirty bill.
- A human approves before anything posts. The drafted entry sits on the dashboard at bills.zoffer.in until you or your accountant approves it. Only then does it reach Zoho Books. Tally support is in early access.
What Zoffer checks, and where your CA takes over
The headline frames this as catching fraud or manipulation. The tool does catch bills that do not add up. Here is exactly what it reads, and where your accountant picks up:
- Zoffer checks the document itself: line items against the stated total, GST against line value, duplicate rows, and tax applied where it should not be. If a restaurant charges ₹400 for a dish worth ₹150, the arithmetic is sound and the bill passes, because price fairness is a market call and that one stays yours.
- A flag tells you exactly where to look. Your CA makes the compliance call from there.
- Zoffer extracts and drafts the entry. GST filing happens in your books, by you or your accountant.
- You get a clear account of what is wrong with the document. How you take it up with the vendor is your call.
We would rather set this out plainly on our own site, so you know what you are getting before you send your first bill.
Why a Hindi local-news desk picked this up
The coverage came from Local18, News18's local reporting network, in Hindi. That is the audience this product was built for: shop owners, distributors, restaurant operators and small manufacturers who run real books, deal with GST every month, and have never opened accounting software in their lives.
For that reader, what matters is the interface: WhatsApp, an app they already use every day, in a language they already use it in. There is no new software to learn and nothing to type. Everything else about Zoffer follows from that one decision.
Frequently asked questions
Yes. The tool is called Zoffer, built by Zoffer Innovations Private Limited, a DPIIT-recognised startup incubated at NSUT in New Delhi. It runs on WhatsApp: you forward a photo of a bill or receipt and it reads the document, extracts vendor, date, amount and GST, and drafts the accounting entry for you to approve.
Not quite. Zoffer is built by Zoffer Innovations Private Limited, an independent company. NSUT's incubation centre hosts and supports the startup, which is where the News18 interview took place during the centre's inauguration. The incubator provides support and infrastructure; the product is the company's own.
The ~90% figure quoted in the News18 report was the founder's field estimate for field-level extraction accuracy at that time, across typical Indian bills including crumpled thermal receipts and phone photos. It is not an audited benchmark. Accuracy varies with image quality and invoice format, which is exactly why every entry goes to a human for approval on the dashboard before it is posted to the books.
Zoffer flags internal inconsistencies in a bill — line items that do not add up to the stated total, GST computed at a rate that does not match the line total, duplicate line items, and tax that has been applied incorrectly. It cannot tell you whether a vendor's price is fair compared to the market, and it is not a legal audit or a consumer complaint tool.
No. Capture happens entirely inside WhatsApp — you forward a bill photo, PDF, or voice note to Zoffer's number. You only open the web dashboard when you want to review and approve the drafted entries before they post to your books.
Free for 20 bills per month. Pro is ₹999 per month for 200 bills, or ₹899 per month on auto-renewal. Tally support is in early access.
Source
Anjali Singh, बिल की हेराफेरी पकड़ेगा यह AI टूल, WhatsApp पर फोटो भेजते ही सामने आ जाएगी ओवरचार्जिंग और टैक्स की गलती, News18 Hindi (Local18, Delhi), 25 June 2026.