SB StartupBasket
All ideas
70 /100 GO Medium complexity

Farq — wage-revision claim drafter for security agencies

Turns every state minimum-wage notification into per-client differential bills, revision letters and arrears claims before the window closes.

HRTech SaaS India Compliance-drivenSMBWhatsApp-firstMultilingualAI-agentSolo-builder
— views
Evaluation Scores
70/100

GO

Overall Score

14
Problem
10
Demand
12
Build
11
Distrib.
11
Revenue
7
Time
5
Defense

Farq — the differential a security agency funds until the client agrees to pay it

1. One-liner

Turns every state minimum-wage notification into per-client differential bills, revision letters and arrears claims before the window closes.

2. Trend signal — why now?

Indian security agencies live on a structural mismatch. The guard’s wage floor is set by the state and moves the day a notification is published. The agency’s price is set by a contract and moves only when the client accepts a revised rate. Between those two dates the agency funds the difference — the farq — out of its own thin margin, and if it lets the contract renew without objecting, the courts say it has forfeited the claim. That mismatch has always existed. What changed in 2026 is that every large state moved at once, by amounts no one had priced, and several moved retroactively.

  • Haryana, +35% across every category, effective 1 April 2026. The Labour Department’s notice “Revised Minimum Wages w.e.f. 01/04/2026” was issued on 9 April 2026 — nine days after the rate took effect. Unskilled went to ₹15,220.71 a month. Industry bodies asked for a phased rollout instead of “a one-time 35% jump”; Deepak Maini of the Progressive Federation of Trade and Industry said “a sudden hike could hit MSMEs particularly hard” (Storyboard18, 21 April 2026). Haryana had 1,116 licensed security agencies at the last parliamentary count.
  • Karnataka, up to ~60%, notified 22 May 2026 with immediate effect, collapsing 81 scheduled employments (security agencies included) into one skill-and-zone grid. Bengaluru unskilled is now ₹23,376 a month per Deccan Herald’s report of the Labour Minister’s announcement. The Karnataka Employers’ Association is in the High Court calling it “an astronomical increase of up to 60%” that is “completely unaffordable and confiscatory,” and arguing the notification is “non-est in the eyes of law” because the Minimum Wages Act was repealed on 21 November 2025 (LiveLaw, 2 June 2026). The court issued notice, refused to consider interim relief until the Union responded, and as of the last reporting I can find the notification remains in force.
  • Maharashtra notified on 5 August 2026 with effect from 1 July 2026 — five weeks of arrears for 60 scheduled employments the day it landed (factoHR, SGCMS). Telangana notified 30 May 2026 under the Code on Wages, effective 1 June. Tamil Nadu and Uttar Pradesh both revised from 1 April 2026. The central sphere revised its VDA from 1 April 2026.
  • Delhi has skipped two revisions. The 1 April 2025 notification “was still the notification in force through August 2026, which means Delhi has passed two revision dates, October 2025 and April 2026, without issuing new rates” (hrengage.ai Delhi page, 22 August 2026). The same page warns: “states that notify late usually backdate … the difference falls due from a date that has already gone by, so it is worth holding a provision against the gap.” When Delhi does notify, every agency with guards in the capital will owe arrears for up to eighteen months on the same day.
  • The courts have just re-stated both edges of the trap. On 18 May 2026 the Delhi High Court held in Moolchand Khairati Ram Hospital v. Vijender Singh that the obligation to pay minimum wages “cannot be rendered illusory by payment of a static amount for years together despite upward revision of minimum wages,” and ordered arrears of differential wages (Mondaq summary). The agency must pay the guard from the effective date, no matter what the client has agreed. And in Mi2C Security Facilities v. North Delhi Municipal Corporation (Delhi HC, 23 October 2018) an agency that claimed three years of revised-wage differential for ~400 guards lost, because it had renewed the contract at the old rate without objection: “Those who contract with open eyes must accept the burdens of the contract along with its benefits.” Pay now, claim on time, or eat it.
  • The industry’s own body describes the cash squeeze. CAPSI chairman Kunwar Vikram Singh, writing on 28 June 2026, lists “delayed payments to security agencies and guards” among the issues raised with the Home Ministry in March 2026, says “private security agencies operate on thin margins, employing large numbers of low-wage workers,” and records CAPSI’s demand “that payments from clients must be released before the 7th of the month … and that non-release of invoices by clients must be penalised.” A 2013 Delhi High Court record (Tops Security v. Tops Detective) put the older, uglier version on paper: agencies are “sometimes constrained to pay less than the minimum wages to the Security Guards … clients do not pay more than what has already been agreed … cannot press its clients to pay more, for fear of losing its business.”
  • Government clients are contractually bound to pay the difference — if the agency claims it. The GeM Manpower Outsourcing SLA, clause 11 of the Special Terms: “In case of any changes in the minimum wages or any statutory wage component as per the Applicable Laws during the Contract period, the Buyer shall pay the Service Provider the differential amount in wage … such increase in the wages will not have any impact on the service charges.” The clause sets no timeline and no documentation standard. Over 1 million manpower resources were hired through GeM in FY2024-25 from 33,000+ service providers (AIR, 2 April 2025).
  • The vendors stop at the invoice. Shivaizer (the most complete Indian guard-ops platform I found) lists “auto-calculated invoices,” “client-specific rate handling” and “rate escalation tracking,” and claims a “60% reduction in billing disputes.” Its page has no minimum-wage revision workflow, no revision letter, no arrears computation, no statutory pass-through. SecurityForce’s site has no billing features at all. Guard-management software is sold on ghost-guard leakage and GPS attendance, not on the one event that moves 25–60% of the agency’s cost base overnight.
  • Reading the notifications just got cheap. Gemini 3.8 Flash shipped on 2 September 2026 at $0.75 per million input tokens (introductory to 31 December 2026), with PDF and image input and a 1M context. A state notification is a scanned table in Kannada, Marathi or Hindi with zone × skill × employment cells; extracting it per state, per revision, is now a rupees-per-document job rather than a data-entry vendor’s product.

Honest gap: I could not get inside CiteHR (the HR forum where these threads live — the site is JavaScript-rendered and the proxy returned only the landing page), and LinkedIn is closed. The operator voice above comes from court records, the industry association, and employer federations, not from agency owners posting in their own words. That costs this idea points on demand evidence and holds confidence at Medium. It does not change the mechanics: the courts, the GeM clause and the 2026 notification calendar are primary sources.

Provenance:

3. The opportunity

The gap is the interval between “the state moved the floor” and “the client signed the new rate.” Nobody owns it.

  • The guard-ops vendors (Shivaizer, SecurityForce, Canticle, eBlackDog) sell the monthly run: attendance, GPS, muster, invoice. Their invoice engine takes a rate card as input. When the rate card should change, they wait for the agency to type in the new number. They do not watch the gazette, they do not rebuild the statutory stack, they do not draft the letter, and they do not compute what the agency has been funding since the effective date.
  • The compliance publishers (Simpliance, TeamLease RegTech, factoHR, SGC, United Consultancy) sell the notification: a PDF, a rate table, a blog post. They stop at “here are the new rates.” The agency still has to turn ₹23,376 basic into a per-guard, per-site, per-client billing rate with PF, ESI, bonus, leave reserve, gratuity, uniform, service charge and GST layered on — for every client, in every zone, and then argue it.
  • The principal-employer side is already productised — including by this catalog (VendorProof, Chaukas: verify the contractor paid the revised wage before releasing the bill). The client’s compliance team can check that the agency paid the guard the new rate. The agency has nothing that helps it get that new rate paid back.
  • The courts have set the rule, not the tool. Moolchand says pay from the effective date. Mi2C says claim before you renew or lose it. GeM clause 11 says the buyer pays the differential but names no form, no deadline, no proof standard. That is a claim with no claim form — the shape this catalog keeps finding unbuilt.

Farq is the claim form. On notification day it reads the gazette, rebuilds every client’s rate card, drafts the revision letter with the notification attached, computes the arrears from the effective date, issues the supplementary invoice, and keeps a running tally of what the agency is funding per client until each one accepts. When a client goes quiet, it produces the objection that keeps the Mi2C door open.

4. Target market

  • Primary customer: The owner or operations head of a PSARA-licensed security or manpower agency with 200–3,000 guards deployed across 20–300 client sites, typically in two to four states. Revenue ₹4–60 crore a year at ₹18–30K billed per guard per month. Billing is done by one accounts person in Excel plus Tally, or in a guard-ops platform whose rate card that person maintains by hand.
  • Why they buy: Because a 35% or 60% floor increase on a 5–10% service charge is a solvency event, not an admin task. In Haryana, an agency with 500 guards at the old ₹11,275 floor saw its statutory wage bill rise by roughly ₹20 lakh a month on 1 April 2026; every month a client sat on the revision letter, the agency funded that. In Karnataka the same 500 guards in Bengaluru cost ₹45–55 lakh a month more from 22 May. The agency must pay the guard from day one (Moolchand), the client pays when it agrees, and if the agency lets the contract roll it forfeits the gap (Mi2C). CAPSI’s own words: “thin margins,” “delayed payments.”
  • Rough TAM reasoning: 23,000+ licensed agencies nationally (July 2024), against 16,427 in January 2022 — the register grows ~15% a year. Most are tiny. If 15% run 200+ guards, that is ~3,500 agencies in the primary band; add housekeeping and facility-management contractors on the same minimum-wage-plus-service-charge model and the count roughly doubles. 33,000+ manpower service providers sit on GeM alone, every one of them with clause 11 in their contract. At ₹1.4 lakh a year, 600 agencies is ₹8.4 crore — about $1M ARR — from under 3% of the primary band.
  • Why now for them: 2026 is the first year every large state repriced in the same six months, three of them retroactively, one by 60%. Delhi’s overdue notification will land with up to eighteen months of arrears attached. The Code on Wages (effective 21 November 2025) changed the definition of wages under the PF and bonus stack, so even the statutory layering has to be redone. And the 1 October 2026 VDA round is three weeks away.

5. Product sketch (MVP)

  • Notification watch. Every state labour department, the central CLC, and the Maharashtra Security Guard Boards, checked daily; WhatsApp alert within hours of a notification with the effective date, backdating period, and which of the agency’s sites it touches.
  • Gazette reader. Extracts the zone × skill × employment table from the scanned PDF (Kannada, Marathi, Hindi, Tamil, Telugu, English) into the agency’s rate structure, with the source cell shown beside every number.
  • Rate-card rebuild. For every client and site: new basic + VDA, then employer PF, ESI, statutory bonus, leave reserve (the 1/6th reliever), gratuity provision, uniform and kit, service charge at the contracted percentage, GST. Old rate, new rate, per guard, per month, per site.
  • Revision letter. Per client, bilingual, citing the notification number, date and effective date, the contract’s statutory pass-through clause (read from the uploaded agreement), the new rate card, and the notification as an attachment. Sent from the agency’s letterhead by email and WhatsApp; delivery and read receipts logged.
  • Arrears and supplementary invoice. Differential from the effective date to today for each client, updated daily; supplementary invoice generated the day the client accepts, or on demand.
  • Exposure ledger. One screen: rupees the agency has funded since each effective date, by client, by state; which clients accepted, which are silent, which refused; the renewal date on each contract and a warning when a contract is about to roll with a claim still open.
  • Objection keeper. If a client goes silent past 30 days, a formal reminder that puts the claim on record; if a contract renews, a covering note reserving the revised-wage claim so Mi2C cannot be cited against the agency.
  • GeM mode. For minimum-wage-based GeM contracts, the clause-11 differential claim with the notification, the revised wage register, and proof of payment to workers, formatted for the buyer’s bill-passing desk.
  • Proof pack for the client’s compliance team. Revised wage register, PF/ESI challan alignment and a one-page summary the principal employer’s VendorProof-style check will accept without a phone call.

6. AI angle — what’s load-bearing

Three jobs that a rate table and an Excel macro cannot do:

  1. Reading the gazette. Each state publishes a different layout — a scanned PDF, zone tables, skill grades, employment schedules, sometimes a “special allowance” that replaces VDA, sometimes a uniform grid that replaces 81 employment-specific tables. The model reads the document as an image, extracts the cells, and maps them to the agency’s sites and grades. The 2 September 2026 Flash pricing makes this a few rupees per notification instead of a subscription to a data vendor plus a clerk.
  2. Reading the contract. The pass-through clause is worded differently in every agreement — “notified wages plus service charge,” “incremental wages will be provided,” a fixed rate with no clause at all (the Mi2C trap). The model finds the clause, classifies the exposure, and drafts the letter to fit it — or flags that the contract has no pass-through and the agency needs a different conversation.
  3. Working the reply. Clients answer by email and WhatsApp: “send the notification,” “our board meets next month,” “we’ll pay from the notification date, not the effective date,” “renew at the current rate.” The model classifies each reply, updates the exposure ledger, and drafts the next move.

Remove the AI and you have Simpliance’s PDF plus the accountant’s spreadsheet — which is what exists today and which nobody uses at the speed the 2026 calendar demands.

7. Localization angle (if any)

This is an India-only play, and the localization is the product.

  • Language: Notifications arrive in the state language; the letter goes out in English with a Hindi, Kannada, Marathi or Tamil version for the client’s site admin.
  • Channel: WhatsApp is where the agency owner lives and where the client’s admin manager reads. Letters, reminders and the exposure ledger all deliver there.
  • Rails: Not relevant to the product; ₹ pricing on UPI autopay for the subscription.
  • Regulation as moat: The statutory stack (PF ceiling, ESI threshold, bonus eligibility, leave reserve norms, Security Guard Board levies in Maharashtra) is India-specific and changes with the Code on Wages rules. Every state’s cadence differs — April/October VDA in most, July/January in Maharashtra, yearly in Karnataka, none in Delhi lately. A global staffing tool will never carry this.
  • Adjacent geographies later: The Gulf’s WPS-plus-labour-supply model and the US Service Contract Act (FAR 52.222-43, which gives contractors 30 days to claim a wage-determination price adjustment) have the same shape. Not for v1.

8. Business model — path to $1M–$5M ARR

  • Pricing: ₹4,999 a month for agencies under 300 guards; ₹12,999 for 300–1,000; ₹24,999 for 1,000–3,000. All states, all clients, unlimited letters. GeM claim packs included. Annual plans two months free.
  • ACV: Blended ~₹1.4 lakh (about $1,650), assuming a mix weighted to the middle tier.
  • Rough math to $1M ARR: 600 agencies × ₹1.4 lakh = ₹8.4 crore ≈ $1M. That is ~2.6% of the 23,000 licensed agencies, or ~17% of the 200+-guard band.
  • Rough math to $5M ARR: 1,800 security agencies plus 700 housekeeping and facility-management contractors on the same model, plus a principal-employer tier (₹9,999 a month for a corporate or IT-park facilities head who wants every vendor’s revision letter checked against the notification — the buy side of the same event). ₹42 crore ≈ $5M.
  • Expansion path: Guard-count tiers grow with the agency. Per-claim packs for GeM buyers. A “renewal desk” upsell that tracks every contract’s roll date and drafts the reservation-of-claim note. Later, a rate-benchmark report (what agencies in your city are billing per skill after the revision) sold to both sides.

9. Go-to-market wedge — first 100 customers

  • Trigger-timed WhatsApp on notification day. The Delhi notification is overdue and will backdate; the central and several state VDA rounds land 1 October 2026. On the day, message every Haryana, Delhi and Karnataka agency on the state controlling-authority lists (Goa Police publishes its list as a PDF; psara.gov.in is the national register; data.gov.in has the state-wise counts — Gujarat 2,203, Punjab 1,228, Haryana 1,116) with a free, finished rate card and revision letter for one client. Converting the other 40 clients is the paid plan. Target 1,500 agencies messaged, 8% open a conversation, 25% of those pay: ~30 customers per round.
  • GeM awardee lists. GeM bid pages for “Manpower Outsourcing Services — Minimum Wage Based” list the awarded service provider. Pull the last twelve months of awards in the four big states, email the differential-claim pack for their live contract. Government buyers are contractually bound by clause 11; the agency just needs the form. 400 awardees, 5% convert: 20 customers.
  • Compliance consultants as resellers. United Consultancy, SGC Services, Prakash Consultancy and their peers already publish every notification and already bill these agencies for PF/ESI returns. Offer 25% recurring for every agency they sign; they get a product to sell on the day their own newsletter goes out. Five consultancies × 8 agencies each: 40 customers.
  • CAPSI state chapters. CAPSI claims 22,000 member agencies and its chairman has put “delayed payments” on the Home Ministry’s agenda. A chapter webinar in Gurugram and Bengaluru titled “What Haryana’s 35% and Karnataka’s 60% cost you per month per client, and how to get it back” — with the exposure ledger demoed live on a volunteer’s contract list. 10 sign-ups per session.
  • Employer federations’ own noise. FKCCI and the Karnataka Employers’ Association are on record that the hike will “cripple” enterprises. Their member companies are the clients who will push back on revision letters. Offer the principal-employer tier to five IT-park and industrial-estate facility heads as a checker, and let them tell their vendors where the letter came from.

10. Build complexity — justification

Medium. The hard part is not the model; it is the corpus and the rules. Roughly 30 notification sources to watch (states, the CLC, the Maharashtra guard boards), each with its own layout and cadence; a statutory stack engine that must be right for PF ceilings, ESI thresholds, bonus eligibility and the leave-reserve convention, and re-checked as Code on Wages rules land; contract-clause extraction with a human-review step; a letter, invoice and ledger layer; WhatsApp and email delivery. A pair ships v1 for Haryana, Karnataka, Delhi and the central sphere in 10–12 weeks, then adds states one notification at a time.

11. Gating checklist

GatePass?Note
Legal in target market✅Drafting a contractual claim under a statutory pass-through is ordinary billing work. The agency signs and sends; Farq does not practise law.
Ethical — no harm / dark patterns✅The product makes the agency pay the guard on time and recover it from the client — it is the mechanism that keeps Tops Security’s “constrained to pay less” from happening.
Market exists (evidence above)✅Three courts, one association, one national procurement clause, five state notifications in six months. Missing: agency-owner forum quotes.
1–5 person team can build this✅Pair, 10–12 weeks, off-the-shelf vision-language model plus a rules engine.
Launchable with <$50K / ₹40L✅₹8–12 lakh: two people for a quarter, model spend under ₹20K a month at launch volumes.

All five pass.

12. Feasibility score

AxisWeightScoreNotes
Problem intensity2014Solvency-grade in revision months (35%, 60%, backdated arrears), quiet in between. Multi-state agencies see six to ten events a year; single-state ones two. The Mi2C forfeiture rule turns a slow leak into a deadline, which is what lifts this above “annoying admin.”
Demand evidence1510Courts (2013, 2018, 2026), CAPSI’s Home Ministry agenda, GeM clause 11, employer federations fighting the hikes, and a vendor category that visibly stops at the invoice. No verbatim agency-owner complaints — CiteHR and LinkedIn were closed to me.
Build feasibility1512Vision-language extraction of gazette tables is the new cheap part. The statutory stack and the 30-source watcher are grind, not research.
Distribution clarity1511Named lists (state PSARA registers, GeM awardees), a named trigger (notification day), named resellers (compliance consultancies) and a named association. Conversion at trigger time is unproven.
Revenue mechanics1511₹5–25K a month against a monthly exposure of ₹20–50 lakh is an easy sentence. The doubt is collection discipline in a segment that itself gets paid late; annual-upfront on UPI autopay is the hedge.
Time to first revenue107Delhi’s overdue notification and the 1 October 2026 VDA round are the launch events. First paid agencies inside 6–8 weeks if v1 is ready for either.
Defensibility105Shivaizer could add a “revision letter” button in a quarter. The moat is the per-state notification corpus, the statutory rule set, and the client-acceptance history that accrues per agency — real by month 12, thin at month 3.
Total10070

13. Qualitative modifiers

Founder-fit tags

technical-heavy · sales-heavy · domain-expertise-required

Needs someone who can build a document-extraction pipeline and someone who has sat with a security agency’s accounts head and knows what “1/6th reliever” and “Zone I VDA” mean. A labour-law compliance consultant as a co-founder or first hire is the fastest route.

Key assumptions to validate (3–5)

  1. Assumption: Agencies are funding the differential for months, not days, after a revision. How to test: Call 30 agencies in Haryana and Karnataka; ask how many clients had accepted the April/May 2026 rate by 31 July and who paid the guards in the meantime.
  2. Assumption: A better letter changes client behaviour. How to test: Draft the Karnataka revision letter for five agencies by hand, have them send it to their ten slowest clients, and measure acceptance at 30 days against their own earlier letters.
  3. Assumption: The 30 notification sources can be watched and read reliably enough that the agency trusts the number. How to test: Back-test the extractor on the last 24 months of Haryana, Karnataka, Maharashtra, Delhi and CLC notifications against Simpliance’s tables; require zero cell errors before launch.
  4. Assumption: Agencies will pay monthly for an event-driven product. How to test: Offer annual-only at launch; if fewer than 1 in 4 interested agencies pay upfront, reprice as per-revision packs.
  5. Assumption: The Karnataka notification survives the High Court. How to test: Track the KEA petition; if it is quashed, measure whether agencies still want the reversal computed (they will owe refunds or hold credits) — a quash is more recompute work, not less.

Risk flags

  1. Regulatory risk: The Karnataka notification is under challenge as issued under a repealed Act, and the Code on Wages rules could reset how every state notifies. A quash creates a reverse-differential problem (credit notes to clients, recovery from guards is impossible), which the product must handle or it looks foolish on the biggest 2026 event.
  2. Platform dependency: None on a private platform. The dependency is on ~30 government sites that change URLs, publish scans, and sometimes post nothing for a year (Delhi). Manual fallback and a human-verified feed are mandatory.
  3. Market timing: Event-driven demand. If the next two quarters are quiet — no Delhi notification, modest October VDA — trial-to-paid stalls. Mitigation: the exposure ledger and renewal-desk features are useful between events; the Code on Wages state rules will keep generating change.
  4. Incumbent response: Shivaizer already tracks “rate escalation.” A revision-letter module is a natural add for them. Farq’s answer is to be the notification-to-claim layer that sits above any ops platform, and to move first on the October round.
  5. Collection risk: Selling to a segment whose defining complaint is that it gets paid late. Annual upfront, UPI autopay, and a price that is a rounding error against one month’s differential.

14. Structured verdict

Score:                  70/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder paired with a labour-law compliance
                        consultant who already services security agencies
Time to revenue:        6–8 weeks if v1 is live before the 1 October 2026 VDA round
                        or Delhi's overdue notification, whichever lands first
Capital to launch:      ₹8–12 lakh ($10–15K)
Top 3 assumptions to validate first:
  1. Agencies funded the April/May 2026 differential for 2+ months on 3+ clients —
     30 calls in Haryana and Karnataka, ask for client-by-client acceptance dates
  2. A notification-cited letter with the rate card attached gets accepted faster —
     five agencies, ten slow clients each, 30-day acceptance vs their own baseline
  3. Gazette extraction is trusted — back-test on 24 months of five sources, zero
     cell errors against Simpliance/TeamLease tables
Kill criteria:
  - Abandon if fewer than 8 of 30 agencies can name a client that delayed or refused
    the 2026 revised rate for more than 30 days
  - Abandon if Shivaizer or SecurityForce ship a notification-driven revision letter
    and arrears module before v1 is live
  - Abandon if fewer than 1 in 4 agencies who ask for the free single-client letter
    convert to a paid annual plan within the same revision round

15. Next step — 1-week validation sprint

  • Day 1–2: Pull the Haryana and Karnataka agency lists from the state controlling-authority registers and the GeM manpower awardees for both states. Call 30 agency owners or accounts heads. Two questions only: “When your state’s 2026 rate took effect, on what date did each of your ten largest clients accept the revised bill?” and “Who paid the guards the difference in the meantime?” Log per-client acceptance lag and rupees funded.
  • Day 3–4: For the five agencies with the worst lag, hand-build the Karnataka 22 May rate card and the revision letter with the notification attached, plus the arrears figure from 22 May to today. Have them send it to their ten slowest clients. Watch replies for 72 hours.
  • Day 5: Go if at least 10 of 30 agencies funded two or more months of differential for three or more clients, and at least 3 of 5 agencies sent the letter and got at least one written acceptance or a request for the notification within 72 hours. No-go if agencies say clients accepted within the month on their own, or if the letters went unsent because the owner would rather not “disturb” the client — that is Tops Security in 2026, and it means the pain is real but the buyer will not act on it.

Interested in a detailed proposal?

Get a deep-dive with market research, competitive analysis, and implementation roadmap.

Contact us

info@startupbasket.ai