What we build09 of 12Reports23 things we build
You learn what went wrong at month end.
AI watches every number as it moves, explains it in plain words and warns you early, and your managers decide what to do.
Tested on twenty of your own past reports before it goes live, and kept working as AI changes.
Eashan Gupta, founder. You meet him on the first call.
All 23
Everything we build for reports and forecasts.
Each one opens with what we build, three examples and what we don't promise.
A sample stock forecast: what runs out where, with a transfer and a reorder waiting for a yes.
-
What will sell, forecast by item and place
Forecasts each item's sales by location from your history, festivals and promotions, and drafts the reorder for your yes.
What we build
- A forecast for every item and location, refreshed each week from your own sales.
- Festivals, promotions, price changes and weather fed in as known future events.
- A range for every number, so you see the likely, low and high case.
- Draft reorders and transfers between locations, waiting for your buyer's yes.
- A monthly check of forecast against what actually sold.
For example, in your trade
- Store chains. A footwear chain gets next month's demand by store and size, with Diwali and the end of season sale built in, and a draft order for its buyer.
- Agriculture. A bakery chain forecasts tomorrow's bread and cakes for each outlet, so less goes in the bin each night.
- Diagnostic labs. A pharmacy chain sees cold and fever medicines rising by store before the monsoon and moves stock between branches.
What we don't promise. We don't promise a forecast will be right; it's a range, it needs some sales history, and it can't see a shock nobody saw coming.
-
One live view of the whole business
Sales, cash, stock and jobs from every system in one view, updated on its own, on your phone.
What we build
- Connections to your accounts, ERP, CRM, bank and sheets that refresh on their own.
- One clean data store where each customer, product and supplier is recorded once.
- A view for each role on phone and desktop: owner, sales, operations, finance.
- Alerts when a number crosses a line you set.
- A morning check that every source arrived and the totals add up.
For example, in your trade
- Manufacturing. A pump maker's owner sees orders, dispatches, receivables and stock for both plants on one phone screen, updated every hour.
- Diagnostic labs. A chain of clinics sees patients seen, revenue per doctor and pending insurance claims for every branch side by side.
- NGOs. An NGO sees children reached, funds spent and field visits by district, ready for the donor's quarterly report.
What we don't promise. We don't promise a number is right when the system it comes from is wrong; every figure shows its source, and gaps are flagged.
-
Ask your numbers in plain words
Ask the company's data a question in plain words and get the answer, a chart and its source.
What we build
- A map of your business words to your data, so 'billing' means what your accountant means by it.
- Questions in English or Hindi on phone, chat or desktop, answered with the number, a chart and the query behind it.
- Answers limited to what each person is allowed to see.
- A test set of known questions that every change must pass before it goes live.
- Good questions saved as weekly reports.
For example, in your trade
- Store chains. A store manager asks which sizes sold out in Indore last week and gets the list and a chart.
- Moving. An operations head asks which customers paid late more than twice this quarter and gets names and amounts.
- Education. A school group's director asks in Hindi how many students are behind on fees at each branch.
What we don't promise. We don't promise every question gets an answer, and a wrong query can give a confident wrong number. Each answer shows its query and source, says so when the data can't answer, and figures for the board or the bank are checked by your finance person.
-
Reports that say why the numbers moved
Your MIS and board pack, written for you, with what changed, what drove it and what to look at.
What we build
- Your MIS and board pack built from live data, in your format.
- A written note on what moved and which customer, product or branch drove it, with the figures.
- A five minute Monday brief on the owner's phone: cash, sales, delays and customers going quiet.
- Follow up questions on any line, answered from the same data.
- Your finance head signs off before anything goes to the board or the bank.
For example, in your trade
- Manufacturing. A component maker's owner reads on Monday that margin fell because two large customers moved to cheaper items, with the invoices behind it.
- Law firms. A CA firm's partners get one page a month on billing, write-offs and unpaid fees by partner.
What we don't promise. We never let the note claim a cause the data doesn't show; when a reason is a guess, it says so.
-
Your cash, thirteen weeks ahead
A daily forecast of cash in and out for thirteen weeks, including which customers will pay late.
What we build
- A thirteen week cash view built from your bank, receivables, payables, payroll and tax dates.
- A prediction of when each customer will really pay, from how they paid before.
- A warning when any week dips below the cash you need to keep.
- What if views: a big client pays a month late, or you hire five people.
- A weekly note to the owner and the accountant.
For example, in your trade
- Construction. A contractor sees in week one that two delayed running bills will leave salaries short in week six, with time to chase them.
- Manufacturing. An auto parts supplier sees which buyers pay past the agreed date and how much cash that ties up each month.
- Agencies. A design agency sees retainers and milestone payments against payroll and knows three weeks ahead when to chase.
What we don't promise. We don't promise exact cash dates or give financial advice; the forecast is a range that gets better as the books stay current.
-
Spend checked, and every leak flagged
Shows what you buy from whom, and flags duplicate bills, price creep and odd discounts as they appear.
What we build
- A spend view: every item, supplier and price paid, across sites and over time.
- Daily checks over purchases, payments, discounts, refunds and sales.
- Your own rules, plus pattern checks that learn what normal looks like for you.
- Each flag with its evidence and the person who should look at it.
- Nothing blocked or reversed without a person's yes, and a monthly count of what was caught.
For example, in your trade
- Moving. A freight company is told when a diesel claim doesn't match the route driven, or a vendor's rate creeps above the contract.
- Store chains. A store chain sees which counters give the most manual discounts, and which stores' returns jump after month end.
- Hospitality. A restaurant group sees one supplier's price for paneer creeping up across its outlets over three months.
What we don't promise. Each flag is a lead for a person to check; we don't promise to catch every leak, and the system accuses nobody.
-
Which customers, jobs and products really make money
Shows the real margin on every customer, job and product, after every cost is counted.
What we build
- Costs traced to each job: people's time, materials, freight, discounts, returns and late payment.
- Margin by customer, product, branch and salesperson.
- The customers and products that cost more than they bring, named.
- What happens to profit if you reprice, change terms or drop one.
- A monthly refresh tied to your books.
For example, in your trade
- Law firms. A law firm sees which clients get the most write-offs and which partners' matters run over budget.
- Moving. A transporter sees margin per route after tolls, waiting time and empty return trips.
- Recruitment. A staffing firm sees margin by client after replacements, slow payment and recruiter hours.
What we don't promise. Where your records don't track a cost we estimate it and mark it; we never present an estimate as exact.
-
What your marketing really brought back
Traces each enquiry and sale to the ad, post, event or referral behind it, marking what's estimated.
What we build
- Source capture on every form, call and WhatsApp enquiry, with tracking links.
- Each sale tied to where the buyer first found you and what made them get in touch.
- A model that estimates what each channel adds where tracking can't see, like word of mouth or hoardings.
- One page a month: spend, enquiries, sales and return by channel.
- Where to move next month's budget, as a suggestion for your yes.
For example, in your trade
- Education. A coaching institute sees which channels bring admissions and which only bring enquiries.
- Brokerages. A developer sees the cost per site visit and per booking for each portal, ad and broker.
- Diagnostic labs. A dental chain sees which campaigns bring patients who come back for treatment after the first visit.
What we don't promise. Credit for offline channels can only be estimated and we say so; we don't promise exact credit for every rupee.
-
Customers going quiet, spotted early
Spots the customers ordering less, paying slower or complaining more, weeks before they leave.
What we build
- A health score for every account from orders, payments, complaints, usage and messages.
- The reason behind each drop, in plain words.
- A weekly list for each account owner, ranked by what's at stake.
- A suggested next step, a call, a visit or an offer, for a person to choose.
- A record of which saves worked, so the score keeps improving.
For example, in your trade
- Manufacturing. A paint maker sees which dealers' orders have fallen three months running, and the likely reason.
- Software companies. A software company sees which customers stopped logging in and which raised three tickets this month.
- Education. A college spots students slipping on attendance, marks or fees weeks early, so a teacher can call.
What we don't promise. A score is an early warning for a person to act on; we don't promise to stop every customer from leaving.
-
Your competitors and market, watched every week
Tracks competitors' prices, launches, hiring and reviews, and new rules, with a short sourced note every Monday.
What we build
- A watch list of competitors, key customers, suppliers and regulators, with the pages and feeds to check.
- Daily checks of prices, product pages, job posts, reviews, tenders, filings and news.
- A Monday note on what changed and why it matters to you, with a source for every line.
- An alert for every big move, like a price cut or a new plant.
- Search across everything collected, by company or topic.
For example, in your trade
- Store chains. An electronics retailer sees each morning which online sellers cut prices on its top 200 products.
- Manufacturing. A cables maker hears when a rival posts jobs for a new plant in Gujarat, and when copper prices jump.
- Finance. A lending company gets a note when the regulator issues a circular that changes its collection rules.
What we don't promise. We watch public sources and data you're allowed to use, and don't promise to see a competitor's private plans.
-
A suggested price, every day
Suggests prices from your demand, costs and competitors' public rates, and a person sets the final number.
What we build
- Your sales, costs, stock and won and lost quotes in one place.
- Competitor prices collected where they're public.
- A suggested price or band for each product, room, route or job.
- Floors, ceilings and minimum margins it can never cross.
- Each suggestion logged against what you chose and what happened.
For example, in your trade
- Hospitality. A 40 room hotel gets tomorrow's suggested rates from its bookings, a wedding season in town and nearby hotels' prices, and the manager decides.
- Moving. A freight broker sees a suggested rate for each route from recent loads, diesel prices and the odds of a return load.
- Manufacturing. A fabrication shop sees the price range it usually wins at for a job type before it sends the quote.
What we don't promise. Prices never change on their own beyond the limits you set, and we don't promise a revenue lift. Each client's prices and data stay separate, rivals never share one engine's view of each other, and the final price is always your own decision.
-
Your next branch or big bet, tested in numbers first
Scores sites for a new branch and shows what a big decision does to cash and profit first.
What we build
- Your own branches' results tied to the neighbourhoods around them.
- Candidate sites scored on population, incomes, footfall, competitors and transport.
- Overlap with your existing branches, so a new one doesn't eat an old one.
- Side by side plans for a branch, a second shift, a hire or a price change, with every assumption written down.
- Projections in the format your bank or investors ask for.
For example, in your trade
- Agriculture. A cafe chain compares five shortlisted sites in Hyderabad on office crowds, rivals nearby and its own sales in similar areas.
- Diagnostic labs. A diagnostics chain finds the areas its patients travel farthest from, a sign a new centre would fill.
- Education. A preschool brand ranks localities by young families, incomes and nearby schools before it scouts.
What we don't promise. Location data in India is patchy and a plan is only as good as its assumptions; we don't promise a site's sales.
-
Your cameras, counting and checking for you
The cameras you already have count visitors and vehicles, time queues and flag unsafe work as it happens.
What we build
- A link to your existing recorder, with new cameras only where the view is poor, faces blurred and footage kept on your premises by default.
- Counts of people, vehicles and goods by hour and zone, and number plates logged at the entrance.
- Your checklist turned into what a camera can see: helmets, safety belts, hairnets, no-go zones, smoke.
- An alert to the person in charge with the clip, and a daily score by site and shift.
- Search of recorded footage by a plain question, with the clip returned.
For example, in your trade
- Moving. A warehouse logs every truck's number plate in and out and sees how long each one waited at the dock.
- Construction. A site in-charge gets a phone alert when someone works at height without a safety belt.
- Agriculture. A central kitchen sees which shifts skip hand washing and hairnets, from cameras it already had.
What we don't promise. We don't promise to catch everything: cameras need light and a clear view, they support a supervisor without replacing one, and we follow privacy law on recording people. Smoke spotted on camera is an early warning and never replaces your fire alarms.
-
Defects and damage spotted by camera
A camera on the line spots scratches, wrong labels and missing parts, and keeps a photo of every reject.
What we build
- A camera and lighting at the right point on your line or dock.
- A model taught on photos of good parts, so anything different gets flagged.
- A reject signal to the line, or an alert to the operator.
- A photo and reason for every reject, by shift and machine.
- A weekly view of what's causing scrap and returns.
For example, in your trade
- Manufacturing. A bottle cap maker rejects cracked rims and wrong colours before they're packed.
- Agriculture. A snacks maker checks every pouch's seal, label and date print as it leaves the packing machine.
- Moving. A warehouse photographs every inbound pallet and flags crushed cartons before the receipt is signed.
What we don't promise. We don't promise zero defects; accuracy depends on light, line speed and examples, and we measure it on your own parts before go-live.
-
Machines that warn before they break
Reads vibration, power and temperature from your machines and warns maintenance days before a likely breakdown.
What we build
- Low cost sensors on the machines that matter, or the signals your machines already give.
- A normal pattern learned for each machine.
- Early warnings with the likely fault and what to check.
- A maintenance job raised in your system, with the parts it needs.
- Power use by machine, so waste shows up too.
For example, in your trade
- Manufacturing. A textile mill is warned days ahead that a spinning frame's motor bearing is wearing out.
- Brokerages. A mall sees which chillers and pumps use the most power and which are drifting toward a fault.
- Power contractors. A solar park sees which inverters produce less than the sunlight says they should, and which panels need cleaning.
What we don't promise. We don't promise to predict every failure; a new machine needs weeks of readings before its warnings can be trusted.
-
Measure and count from a phone video or photo
Turns a phone video or a few photos into a counted, measured list for a quote or audit.
What we build
- A guided capture on the customer's, surveyor's or field rep's phone.
- Items, rooms, shelves or work done listed from the video, with sizes and counts.
- Volume, weight or quantity estimated as a range.
- A draft quote, audit or progress note for a person to check.
- Every photo kept as proof, with time and place.
For example, in your trade
- Relocation. A family films a walk through their home and the mover gets a room by room list and a volume range before quoting.
- Store chains. A field rep photographs a shelf and gets share of shelf, missing products and price labels checked from one photo.
- Construction. A site's weekly 360 degree walk is compared with the plan, and work running behind is flagged.
What we don't promise. We don't promise a final quote, claim or progress figure from video alone; it's a first pass that a person checks.
-
Orders likely to come back, caught before you ship
Scores each cash on delivery order for the risk it comes back, and confirms risky ones before they ship.
What we build
- A risk score on every cash on delivery order from the address, pin code, past orders and how the order was placed.
- A confirmation on WhatsApp or a short call for the risky ones, before the parcel leaves.
- A prepaid offer, like a small discount, where it's worth more than the risk.
- Addresses checked and corrected with the customer before dispatch.
- A monthly read of returns by pin code, product and courier, so the rules get better.
For example, in your trade
- Store chains. A footwear brand confirms every order over ₹3,000 from a high return pin code on WhatsApp, and cancelled orders never reach the courier.
- Agriculture. A dry fruits seller offers a small prepaid discount on festive orders the score marks as risky.
- Diagnostic labs. An online pharmacy confirms the address on first time cash orders before the medicines are packed.
What we don't promise. A score only decides when to confirm or ask for prepayment; you set the rules and review them every month, and no area is blocked on a score alone.
-
Carbon numbers and ESG answers your buyers ask for
Turns power, fuel, freight and material records into carbon figures and the ESG answers buyers and regulators ask for.
What we build
- Power, fuel, freight and material data pulled from your bills, meters and accounts, plant by plant.
- Carbon per product and per shipment, worked out by the method your buyer or regulator names.
- The emissions data your EU importers must report under the carbon border rules, in their format.
- BRSR answers and buyer questionnaires drafted from the same numbers.
- A record of every figure's source and method, ready for a verifier.
For example, in your trade
- Manufacturing. A steel parts maker sends its Dutch importer the emissions for each shipment, worked out from its furnace power and the steel it bought.
- Moving. A transporter gives a listed customer the carbon for every lane it runs, from fuel bills and kilometres driven.
- Agriculture. A tea exporter answers a European retailer's sustainability questionnaire from its own power, fuel and fertiliser records.
What we don't promise. Numbers are estimates until an accredited verifier checks them, and no report claims more than the data shows.
-
What people say about you online, watched every day
Watches social media, forums, news and videos for your brand, and raises a crisis while it is small.
What we build
- A daily watch on public posts, forums, news, review sites and videos for your names and products.
- Mentions sorted into praise, complaint, question and risk, in the language they're written in.
- An alert for anything spreading, with the post and who's sharing it.
- Replies drafted for your team to approve, and complaints sent to service with the details.
- A weekly read of what people say about you and your competitors.
For example, in your trade
- Agriculture. A snacks brand hears that a video of a foreign object in a pack is spreading, with the batch number from the video for quality to check.
- Education. A university sees a student's thread about hostel food gaining replies, and gets the warden and the student talking before it spreads.
- Finance. A lender sees complaints about a collection agent's calls on a forum and raises them with the collections head.
What we don't promise. We read only public posts, never track private individuals, and a mention isn't treated as fact until someone checks it.
-
Fuel, idling and driving watched across your fleet
Reads GPS, fuel sensors and fuel cards, and flags fuel loss, idling, harsh driving and missed servicing per truck.
What we build
- GPS, fuel sensor, fuel card and trip data brought together for every vehicle.
- Fuel drops that don't match the distance driven flagged with the time and place.
- Idling, harsh braking, speeding and night driving scored by driver.
- Servicing, tyres and papers tracked by kilometre and date.
- A monthly cost per kilometre for each truck, route and driver.
For example, in your trade
- Moving. A transporter with 120 trucks sees three that lose fuel at the same stop on every trip, with the times to check.
- Construction. A contractor sees which excavators idle most on site and how much diesel that costs each month.
- Agriculture. A dairy's milk tankers are scored on harsh braking and speed, and the drivers who need it get coaching.
What we don't promise. A flag is a lead to check and never proof of theft, drivers are told what's tracked, and fuel sensors must be calibrated before their numbers are trusted.
-
Farms, sites and rooftops checked from satellite and drone pictures
Reads satellite and drone pictures to check crop health, site progress, stockpiles and solar panels, and flags what changed.
What we build
- Satellite pictures of each farm, site or plot, read every few days for change.
- Crop health, sowing and harvest dates estimated for each field.
- Drone surveys turned into maps, counts and volumes for sites, stockpiles and solar parks.
- A list of what changed, with before and after pictures, for a person to check.
- Field visits planned around the places that most need one.
For example, in your trade
- Finance. A rural lender sees which of its financed farms show no sowing a month into the season and sends field officers there first.
- Power contractors. A solar park operator gets a drone map of panels running hot after each survey, with the rows listed for cleaning or repair.
- Construction. A road contractor measures its sand and aggregate stockpiles from a monthly drone flight.
What we don't promise. Clouds, resolution and season limit what a satellite sees, so every flag is checked on the ground before anyone acts on it.
-
A credit limit for every dealer, set from how they pay
Suggests each dealer's credit limit and terms from how they pay and order, and warns when one starts slipping.
What we build
- A payment and order history for every dealer and customer, from your books and bank.
- A suggested limit and terms for each, with the reasons, for your finance head to approve.
- Warnings when an account pays slower, orders oddly or crosses its limit.
- Public GST filing records and court cases checked for warning signs.
- Orders over the limit held for a person, and never blocked on a score alone.
For example, in your trade
- Manufacturing. A cables maker's finance head sees a suggested limit for each of 600 dealers, with the three paying slower every month flagged.
- Diagnostic labs. A pharma distributor holds new orders from a chemist whose cheques bounced twice, and the area manager decides.
- Agriculture. A biscuit maker raises limits for its fastest paying distributors before the festive season, with each change approved by finance.
What we don't promise. It suggests and never decides, and for a regulated lender the credit rules and fair lending duties stay with your credit officer.
-
A hundred customers interviewed, in their own words
An AI interviewer talks with each customer or lost buyer in their language, and returns themes in their words.
What we build
- An interview guide written with you, with follow up questions the AI asks when an answer is vague.
- Interviews by voice call or chat in the customer's language, at a time they choose.
- Themes counted across every interview, each with the customer's own words.
- A short report on what to fix first, with every quote linked to its interview.
- Invitations only to people who agreed to be contacted.
For example, in your trade
- Software companies. A software company interviews 80 customers who cancelled this quarter and learns most left over one missing report.
- Education. A college talks with 150 students who accepted a place and didn't join, and hears why in Hindi and English.
- Store chains. A mattress brand asks 100 recent buyers how they first heard of it and what nearly stopped them buying.
What we don't promise. People are told it's an AI and take part by choice, and no quote is used outside the company without consent.
Most owners don't know
What AI can now see.
A camera still at the dock: 18:42, the last truck out.
Ask your camera footage a question
Ask your footage when the last truck left, and get the clip.
A camera that learns from good parts alone
A camera learns good parts from photos, and flags anything different.
A walk-through video becomes a quote
A customer films a walk-through, and AI lists and sizes the job.
It works today. We keep it working tomorrow.
How it works- Work it out, most consultants stop here
- Build it, most software firms stop here
- Run it
- Answer for it
- We stay
Tell us which number you always see too late.
In 45 minutes with Eashan you leave knowing what we'd build first, and why. Nothing to sign.
What stays yours
- Your accountsin your company's name
- Your codeevery line of it
- Personal dataabout your customers and staff, on your own machine unless you approve otherwise in writing
- Plain files and a how-to guideif you stop
Eashan Gupta, founder. You meet him on the first call.