Apka issue, our tech
We find the one job in your business worth fixing — the orders, the follow-ups, the chasing — fix it, and leave it running where your team can use it without us.
Case studies
"Built" means it runs today inside Benir Group. "In progress" is scoped and being built. "Modelled" means a fix designed from documented results at other Indian businesses, with sources.
231 corporate clients across five divisions, and nobody could say which client bought what, or where the gaps were.
One dashboard that maps every client to every division, scores the cross-sell opportunity, and exports target lists per city.
20-litre corporate water orders arrived by phone and WhatsApp, got retyped by the desk, and dispenser complaints lived in someone's head.
A client app and a staff portal: per-site order links, driver schedules, and service tickets, all in one place.
Hundreds of open invoices chased by hand, in whatever order the accounts desk remembered.
A day-30 to day-120 reminder sequence over WhatsApp and email, with a human gate before any key account is contacted.
Telecallers work the same list every week; most calls are to people who were never going to buy.
An Indian-language voice agent does the first pass and books meetings only with people who said yes.
Orders come in as messages and voice notes; someone decodes them by hand every afternoon.
An assistant that reads the message, confirms quantity and price, and routes the order to the right person.
Three delivery-app tablets, a counter POS and a printer, and a kitchen that re-reads the same ticket twice.
Every order from every source lands on one screen, already grouped by status.
How it works
Four steps, the same every time, whether the fix takes a week or a quarter. Pricing is per project, agreed up front.
We talk about one annoying thing. Not a transformation plan: one part of your day that wastes time or loses money.
A short written brief: what gets fixed, what stays untouched, and what "done" looks like. You approve it before anything is built.
Built in phases on your real data, scoped narrow on purpose so it ships in weeks, not quarters. You see it working before you commit to more.
It goes live and gets watched. If it isn't saving time, it gets tuned until it does. Then we talk about the next annoying thing.
Get in touch
Tell me what it is. If there's a fast, narrow fix for it, I'll say so. If there isn't, I'll say that too. Usually a reply the same day.
Prefer email? nayan@benir.in
Fastest way to reach me. Write a line or two about what's slowing you down, or just send a voice note.
Opens WhatsApp with your message typed out. Nothing is sent until you press send there.
Benir sells office supplies, packaged water, e-waste disposal, corporate gifting and PPE to the same companies. The question "which of our water clients don't buy supplies yet?" used to take a week of spreadsheets.
1,200+ clients, five divisions, and no single view of who buys what. Sales pitched water to cities where water couldn't be delivered.
One dashboard: every client mapped to every division, an opportunity score, and exportable target lists per division and city.
Sat with the KAMs. The real question was "which of our water clients don't buy supplies yet?"
A written brief: which spreadsheets, what colour-coded rows to skip, what never to change.
Import → scoring → targeting, each phase checked against the real client file before the next.
Live on Benir's server. Serviceability rule added after the first week of real use.
| Client | Managed by | Divisions | Score |
|---|
Bailley supplies 20-litre water to corporate offices across Bengaluru and Chennai. Dozens of sites, each with its own contract, minimum order and dispenser that occasionally breaks.
Water orders came by phone and WhatsApp and were retyped by the desk. Dispenser complaints lived in chat threads.
A client app with a per-site order link, a staff portal with the day's orders and driver routes, and service tickets per site.
Shadowed the desk for a week of orders. Every site had rules nobody had written down.
A spec with the contract rules in it: bottle sizes, minimums, who may order.
Backend → client app → staff portal, three review rounds on real sites.
Live internally. Bulk login upload and site messaging added from team feedback.
Each open invoice gets the right reminder at the right time, in a tone that fits the customer. A person only steps in when a human call is actually needed.
Hundreds of open invoices chased in whatever order the accounts desk remembered. Key accounts sometimes got reminders nobody approved.
A collections agent runs a day-30 → day-120 sequence over WhatsApp and email, with a human gate on key accounts.
Pulled the ageing report. Most of the money sat in a few accounts nobody wanted to call.
A brief with the sequence, the tone per customer type, and who must approve what.
Ledger read → message drafting → send & track, tested on a handful of accounts first.
Rolling out. Tone and timing tuned as replies come in.
| Customer | Invoice | Late | Amount | Status |
|---|
Demo names and amounts are illustrative.
It works the list, speaks the customer's language, asks the qualifying questions, and only hands over the people who said yes to a meeting.
Telecallers work the same few hundred numbers each week. Most calls go to people who were never going to buy.
An Indian-language voice agent makes the first call, asks the qualifying questions, and books meetings only with people who said yes.
Listen to a week of real calls. Find the three questions that separate buyers from browsers.
A script brief: languages, questions, what counts as "qualified", when to hand over.
Voice agent → calendar booking → CRM notes, tuned on 50 real calls before scaling.
Live on the full list. Script adjusted weekly from call outcomes.
Sources: Mahindra Finance & Sarvam: 12 languages, 1 crore+ calls · Caller.digital (vendor case study): 30-min speed-to-lead for a Kanpur solar firm · Edesy (one vendor's published price): ₹4–6 per minute
Customers type the way they already do. The assistant reads it, confirms quantity and price, and routes the order to whoever needs to act on it.
Orders arrive as messages, voice notes and photos of lists. Someone decodes them by hand every afternoon.
A WhatsApp assistant reads the message, confirms items, quantity, price and slot with a one-word approval, and routes the order to the right person.
Read a month of order chats. Note every way customers say "same as last time".
A brief: catalogue, prices, delivery slots, what to confirm, what to escalate.
Text → voice notes → photos, one input type at a time, live with a few trusted customers.
Live for everyone. Unclear messages go to the owner, never guessed.
Source: Sushant Kuttalam Chandrasekar, LinkedIn (one founder's public write-up, a single example rather than a benchmark)
Swiggy, Zomato, the counter and WhatsApp all land in one place. The kitchen taps once to move an order along. Nobody retypes a ticket.
Three delivery-app tablets, a counter POS and a printer. The kitchen re-reads tickets and a wrong item goes out once or twice a day.
One kitchen display on any TV or tablet. Every order from every source lands in a lane: new, preparing, ready.
Stand in the kitchen at peak. Count how many times a ticket is read aloud.
A brief: sources, item names to normalise, what "ready" means for this kitchen.
One source first (usually the busiest app), then the rest. Live on a spare TV.
Live at peak. Lane timings show the manager where the kitchen slows down.
No email required, nothing sent anywhere. Answer these and you'll get a straight read on whether there's a fix worth doing in your business right now — including "not yet", if that's the truth.
Pick yours to see the problems that come up most, what each one is typically worth, and where we have done the work ourselves. Whatever your trade, we start with a diagnosis, not a product.
Sit with your business, work out where the time and money are leaking, and say which one thing is worth fixing first — sometimes the answer is nothing yet.
The trades we can help, the problems that recur in each, and the arithmetic behind them. Calibration, not a price list.
Some trades we have worked in ourselves. Others are new ground. We say which is which, so you know how much is proven.
If none of these is quite your problem, that is the normal outcome — and the reason the first conversation is a diagnosis rather than a demo.
Check it yourself
Five of these, one per kind of fix. They follow the industry you picked above. Defaults are deliberately conservative, and more than one of them will tell you a fix isn't worth doing.