What we build08 of 12Paperwork21 things we build
Documents are still read and typed by hand.
AI reads and checks every document that comes in, drafts the reply in your format, and a person signs off before it moves.
Tested on twenty of your own past documents before it goes live, and kept working as AI changes.
Eashan Gupta, founder. You meet him on the first call.
All 21
Everything we build for paperwork.
Each one opens with what we build, three examples and what we don't promise.
A sample hiring desk: applications read, a shortlist with reasons, interviews booked.
-
Every CV read against the real job
Reads every application against what the role really needs and gives recruiters a ranked shortlist with reasons.
What we build
- The role written down with the hiring manager: must haves, good to haves and deal breakers, in plain words.
- Every CV and application form read, whatever the format or language, and matched to that list.
- A ranked shortlist with the reason for each place, and nobody turned away without a recruiter seeing why.
- Job posts drafted from the same list, and every applicant told where they stand.
- A monthly check that the shortlist isn't skewed by gender, age or college, so a bias is caught early.
For example, in your trade
- Recruitment. A staffing firm reads all 3,000 applications for a warehouse client and sends a shortlist of 60, each with its reason.
- Diagnostic labs. A hospital group matches nurse applicants to each ward's needs, from registration and experience to the languages they speak.
- IT firms. An IT services firm reads every CV for a cloud role against the skills the project really uses, so buzzwords don't beat experience.
What we don't promise. It never turns anyone away on its own, never scores age, gender, caste, religion or a photo, and a recruiter reads the reasons before anyone is rejected.
-
Supplier invoices read and matched before you pay
Reads every supplier invoice, matches it to the order and receipt, and flags duplicates and wrong amounts.
What we build
- Invoice capture from mail, WhatsApp photos and scans.
- Fields read and checked: GSTIN, tax, totals and due dates.
- Matching to the purchase order and goods receipt, line by line.
- Duplicate and odd amount flags, each with the reason.
- Approved entries posted to your accounting system.
For example, in your trade
- Construction. A contractor's site bills are matched to the purchase orders, and a cement bill at the wrong rate is held for a person.
- Hospitality. A restaurant group sees that its vegetable supplier billed the same delivery twice.
What we don't promise. Nothing is paid or posted without your accounts team's approval, and blurred or handwritten lines always go to a person.
-
Staff expenses from a photo of the receipt
Staff send a photo of the bill, and it's read, coded, checked against your policy and ready to approve.
What we build
- Bills sent as photos on WhatsApp or the app, or forwarded from mail, read in any language.
- Each one coded to the right head, project or client, with the GST details captured.
- Your policy checked line by line: limits, duplicates, weekend claims and missing bills flagged.
- Approval in one tap for the manager, and entries ready for your accounts system.
- Input credit marked only where the bill carries your GSTIN and the rules allow it.
For example, in your trade
- Telecom. A service company's 80 technicians send fuel and parts bills from the road, and finance sees them coded by job each morning.
- Diagnostic labs. A pharma company's medical reps photograph their travel bills, and claims above the city limit go to the area manager with the reason.
What we don't promise. It flags and prepares but never approves a claim or pays anyone, and your accountant decides what input credit you claim.
-
Customer orders entered without anyone typing
Reads purchase orders from mail, PDF, Excel and WhatsApp and enters them, flagging anything that doesn't match.
What we build
- One intake for mail, PDF, Excel and photo orders.
- Lines matched to your item codes, prices and each customer's terms.
- Stock and credit limit checked before the order goes in.
- Only the unsure lines sent to a person.
- Orders entered in your ERP or accounts system, with the confirmation drafted.
For example, in your trade
- Manufacturing. A nuts and bolts distributor's daily order emails go straight into its system, and only the ones with unknown part codes reach a person.
- Agriculture. A dairy's retailers send order photos on WhatsApp that land in the system before the morning dispatch.
What we don't promise. No order goes in on a guess, and anything unclear waits for someone on your team.
-
Onboarding papers checked before day one
Reads IDs, company papers and certificates, checks them against each other, and flags anything missing or expired.
What we build
- Reading of PAN, Aadhaar with the number masked, passports, GST and company certificates.
- Cross checks between documents: names, dates, numbers and addresses.
- Missing and expired papers listed for the customer in plain words.
- Checks against official sources where you have access.
- A complete file for your officer to approve.
For example, in your trade
- Finance. An NBFC branch gets a borrower file where the names on the PAN and the bank statement don't match, flagged before credit sees it.
- Recruitment. A staffing firm collects and checks two hundred new joiners' papers before their first day.
What we don't promise. Your officer always makes the final call, Aadhaar numbers are hidden, and papers are kept only as long as the law allows. Reading a paper doesn't prove it's genuine, so anything that matters is checked at its source, like the issuing registry or DigiLocker, before your officer signs off.
-
Contracts checked against your rules
Reads every contract against your standard terms, marks risky clauses, and suggests your fallback wording.
What we build
- A playbook of your positions: liability, payment terms, termination and jurisdiction.
- Clause by clause review with each risk rated and explained.
- Suggested changes as tracked edits in Word.
- A plain words summary for the business owner.
- Key dates and obligations saved to your contract register.
For example, in your trade
- Software companies. A software company's sales team sees which clauses in a bank's contract break its liability cap before legal opens it.
- Construction. A contractor gets every clause in a subcontract compared with the matching clause in the main contract.
What we don't promise. It is a first read for your lawyer and never legal advice, and nothing is agreed without your lawyer's sign off.
-
Every renewal and notice date in your contracts, tracked
Pulls renewal dates, notice periods, price rises and promises from signed contracts and reminds you in time.
What we build
- Every signed contract, lease and licence read, old scans included.
- A register of dates, notice periods, prices and obligations, each linked to its clause.
- Reminders to the owner well before each date.
- Answers to questions like which suppliers can raise prices this year.
- New contracts added as they are signed.
For example, in your trade
- Store chains. A retail chain sees every store lease with a rent increase due next quarter.
- IT firms. An IT firm is reminded ninety days before a software licence renews itself.
What we don't promise. We can only track contracts we are given, and the clause itself always wins over the summary.
-
Tender and RFP responses drafted from your past bids
Reads the tender, lists every requirement, and drafts answers from your past bids and approved facts.
What we build
- A requirements list with every must, should and document the tender asks for.
- An eligibility check against your turnover, experience and certificates.
- Answers drafted from your past bids and approved company facts.
- A missing items list with an owner and a date for each.
- A final check that every page, form and annexure is in before submission.
For example, in your trade
- IT firms. A systems integrator gets most answers to a 300 question bank RFP drafted from last year's bids, for its team to check.
- NGOs. An NGO's grant proposal lists every funder requirement and answers each with evidence from its annual reports.
What we don't promise. We never submit a bid, and every figure, certificate and commitment in it is confirmed by your people.
-
Proposals, reports and letters drafted in your format
Drafts the documents you send every week on your letterhead, from your records, with approvals built in.
What we build
- Your templates built once: proposals, reports, letters, certificates and work orders.
- Figures pulled from your CRM, ERP and spreadsheets into the right places.
- A wording check against your style and approved claims, and a brand check on logo and layout.
- Comments, approvals and version history in one place.
- Word and PDF outputs, sent for e-signature when needed.
For example, in your trade
- Law firms. An engineering consultancy's monthly progress reports for twelve clients come out of its project tracker on the first of each month.
- Education. A school sends every parent a term report written from the marks sheet and the teacher's two line note.
What we don't promise. Every number comes from your records, and nothing goes out until someone on your team has checked it.
-
Site and inspection reports written from photos and voice notes
Turns photos, videos and voice notes from a site visit into a dated, signed report.
What we build
- A phone app that works offline on site.
- Photos and video stamped with time and place.
- Voice notes turned into written findings, room by room or item by item.
- A report in your format, signed on the phone.
- A sealed PDF that holds up in a dispute.
For example, in your trade
- Brokerages. A property manager's move out inspection becomes a signed report with every photo stamped before the tenant leaves.
- Construction. A site engineer's walk round becomes a daily progress report with photos tagged to each floor.
What we don't promise. The report records what the camera and the inspector saw, and never states a cause or a value the inspector didn't.
-
A hundred documents asked the same questions at once
Asks every document in a pile the same questions and returns one table, each answer linked to its page.
What we build
- Bulk upload, or a link to the folder.
- Your questions as columns: dates, amounts, parties, clauses and risks.
- One row per document with the answer and the page it came from.
- Cells the AI isn't sure of marked for a person.
- Export to Excel or straight into a report.
For example, in your trade
- Finance. A lender has three hundred loan agreements read for change of control clauses, and gets one table back.
- Brokerages. A buyer of an office building gets every tenant's lease terms in one table, with the page for each.
What we don't promise. The table is a starting point for your experts, and any answer that matters is checked against its page.
-
Old paper, scans and handwriting turned into searchable records
Turns scans, stamped papers, handwritten registers and Indian language documents into clean records you can search.
What we build
- Reading of scans, photos, tables, stamps and handwriting.
- Indian scripts as well as English.
- Fields pulled into a spreadsheet or your system, each with a confidence mark.
- A spot check by your people on a sample before the rest runs.
- Everything searchable, with the original image beside the text.
For example, in your trade
- Brokerages. A land developer turns thirty years of handwritten sale deeds and registers into a searchable list of plots and owners.
- Education. A college reads its old mark registers so a former student's transcript is issued without a search through the archive.
What we don't promise. Handwriting and faded scans are read with some errors, so we test on a sample of your own pages first and a person checks every name and number that matters. Handwritten pages in Indian scripts are the hardest, so some may need a person to type them.
-
Export papers checked before the bank sees them
Checks invoices, packing lists, bills of lading and certificates against the letter of credit and each other.
What we build
- The letter of credit and every shipping document read.
- Cross checks of names, quantities, weights, dates, ports and descriptions.
- A list of mismatches with the term each one breaks.
- Corrections drafted for the documents you prepare.
- A record of every check to share with your banker.
For example, in your trade
- Agriculture. A rice exporter's document set is checked against the LC, and a weight mismatch between the packing list and the bill of lading is caught.
- Moving. A freight forwarder checks every client's document set the night before it goes to the bank.
What we don't promise. The bank's examiner has the final word, so we don't promise a document set will be accepted.
-
Drawings read and quantities countedNew
Reads building drawings and specs, counts the fittings, measures the areas, and lists the work for your estimator.
What we build
- Whole drawing sets and specifications read, sheet by sheet.
- Doors, fixtures and symbols counted, and floor and wall areas measured, for your estimator to check.
- Scope summaries and open questions drafted from the specifications.
- Quantities laid into your rate sheet.
- Changes between drawing revisions highlighted.
For example, in your trade
- Construction. An interiors contractor gets door, flooring and ceiling quantities for a twelve floor office set to check before pricing.
- Power contractors. A rooftop solar installer gets roof areas and a first panel layout from an architect's drawings.
What we don't promise. Measured quantities are a draft for your estimator, and no bid is priced from them without a person checking.
-
Claim files checked against the policy
Reads a claim's bills, reports and forms, checks them against the policy, and lists what's covered or missing.
What we build
- Bills, reports, photos and forms in the file read.
- A summary of dates, amounts and what each document shows.
- A check against the policy's cover, limits and exclusions, with the clause linked.
- Missing documents listed for the customer.
- Doubtful items flagged for a claims officer.
For example, in your trade
- Finance. A health insurer's claims team gets a one page summary of a hospital bill set against the policy's room rent limit.
- Moving. A transporter's damage claim file is checked for the survey report and photos before it goes to the insurer.
What we don't promise. The AI never approves, rejects or values a claim, and every decision stays with a licensed officer.
-
Bank statements, GST returns and accounts read into one view
Reads bank statements, GST returns and accounts into one clear view of cash flow, dues and red flags.
What we build
- Bank statements read from bank PDFs, scanned ones included, or from consented data feeds.
- GST returns and income tax returns matched against the statements.
- Income regularity, bounced cheques, EMIs and large transactions found.
- Mismatches and signs of tampering flagged as questions.
- A one page summary in your credit format.
For example, in your trade
- Finance. An NBFC's credit team gets twelve months of a trader's statements read, with three bounced cheques and an unexplained cash deposit flagged.
- Law firms. A CA firm builds a client's loan file with statements, GST returns and accounts reconciled in one sheet.
What we don't promise. We never make a lending decision, and a tampering flag is a question for your officer and never an accusation.
-
Documents translated with the layout intact
Translates contracts, manuals and notices into Indian and foreign languages, keeping tables and layout in place.
What we build
- Translation of Word, PDF and slides with layout, tables and numbering kept.
- Your glossary, so product and legal terms come out right.
- Indian languages and the languages of your export markets.
- A fluent reviewer's check before anything binding goes out.
- Source and translation side by side for checking.
For example, in your trade
- Manufacturing. A machine maker's operating manual comes out in Hindi, Tamil and Arabic with every diagram label in place.
- Government. A district office publishes each scheme notice in English, Hindi and Urdu together.
What we don't promise. Machine translation can change legal meaning, so a fluent person checks every binding document.
-
Forms and applications filled from your records
Fills government, bank, insurance and visa forms from papers you already hold, ready for your team to submit.
What we build
- One store of your checked details and documents.
- Forms filled field by field from that store, each field traced to its source.
- Checks for missing attachments and mismatched details.
- Submission ready packs that the AI never submits itself.
- Reminders for renewals and expiry dates.
For example, in your trade
- Relocation. A relocation firm fills a family's visa and registration forms from the passports already on file, ready for the consultant to check.
- Moving. A customs broker drafts shipping bills from the invoice and packing list, with every field traced to its page.
What we don't promise. We prepare forms and never file or submit them, and a person checks every one before it goes.
-
Decks made from your notes and numbers, in your brand
Turns a brief, a report or this quarter's numbers into a finished deck in your template, ready to edit.
What we build
- Your slide template built once: layouts, colours, fonts and the logo in the right places.
- A first deck from a brief, a report or a voice note, with the story laid out slide by slide.
- Charts and tables drawn straight from your accounts, CRM or sheets, each with its source.
- Files that open and edit in the presentation software your team already uses.
- A check that every number on a slide matches the data before anyone presents it.
For example, in your trade
- Finance. A wealth manager's quarterly review deck for each family is built from the portfolio report, with every chart sourced, for the adviser to check.
- Manufacturing. A component maker's board deck is ready on the third working day of the month, with sales, margin and cash charts drawn from the books.
- Agencies. An agency turns a client's brief into a 20 slide pitch in its own template, ready for the team to check.
What we don't promise. Every number on a slide comes from your data with its source, and a person checks the deck before anyone sees it.
-
Legal research and first drafts, with every citation checked
Finds judgments and rules behind a matter, builds a timeline from the file and drafts notices for your lawyer.
What we build
- Research across Indian judgments, statutes and circulars, each answer quoting the paragraph it rests on.
- Every citation checked against the official text before a lawyer sees it.
- A timeline of events and a list of documents built from the case file.
- First drafts of notices, replies and briefs in your firm's format.
- A note of what the research didn't find, so nobody takes it as complete.
For example, in your trade
- Law firms. A litigation firm gets the recent High Court judgments on a cheque bounce defence, each with the paragraph quoted and checked.
- Brokerages. A developer's legal team builds a timeline of a buyer's complaint from five years of letters before drafting its reply to the consumer commission.
- Finance. A lender's legal desk drafts a recovery notice from the loan file, with each clause it relies on cited.
What we don't promise. AI can invent cases, so every citation is checked against the official text before a lawyer sees the draft; it is never legal advice and a lawyer signs everything.
-
Tax notices read and replies drafted for your CA
Reads a GST or income tax notice, checks it against your returns and books, and drafts your CA's reply.
What we build
- Each notice read in plain words: what's asked, the amount, the section and the deadline.
- The figures in question matched against your returns, your books and the GST portal's statements.
- A draft reply with the supporting invoices, ledgers and reconciliations attached.
- Every deadline tracked, with reminders to your CA before it falls due.
- A file of past notices and replies, so the next one starts from what worked.
For example, in your trade
- Manufacturing. A pipe maker's notice about a mismatch in input credit is matched to the three suppliers who filed late, with their invoices attached for the CA.
- Law firms. A CA firm sees every client's open notices and deadlines on one screen, each with a first draft reply ready to review.
- Store chains. A D2C brand's notice about cash deposits is matched to its marketplace payouts and bank statements before the CA writes back.
What we don't promise. It drafts and never files, it isn't tax advice, and your CA decides what goes to the department.
Most owners don't know
What AI can now read.
A floor plan with its doors found and counted for your estimator.
AI can count what's on a drawing
It counts the doors and fittings on your drawings for your estimator.
Export papers checked before the bank checks them
A wrong name or weight is fixed before it holds up your money.
Old handwritten registers can be read
Old handwritten registers turned into records you can search.
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 documents your team still types by hand.
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.