SB StartupBasket
All ideas
72 /100 GO Medium complexity

Bakaya — arrears ledger for Ayushman-empanelled hospitals

Computes what the State Health Agency owes your hospital under Ayushman Bharat, interest included, and drafts every deduction appeal.

HealthTech SaaS India Hospital revenue cycleGovernment payerCompliance-drivenSMBSolo-builderAssociation-distributedHindi+English
— views
Evaluation Scores
72/100

GO

Overall Score

17
Problem
12
Demand
12
Build
11
Distrib.
9
Revenue
7
Time
4
Defense

Bakaya

1. One-liner

Computes what the State Health Agency owes your hospital under Ayushman Bharat, interest included, and drafts every deduction appeal.

2. Trend signal — why now?

Ayushman Bharat PM-JAY is the world’s largest government-funded health cover: 32,000-plus empanelled hospitals, roughly half private, 8.7 crore admissions authorised. The private half is paid by State Health Agencies (SHAs) that contractually owe payment within 15 days of claim submission and, under the hospital MoU, 1% interest per week on anything later. In 2026 that promise has collapsed in state after state, and for the first time a High Court has turned the interest clause into an enforceable number.

The 2026 sequence:

  • 24 March 2026 — Jammu & Kashmir: ₹176 crore pending, no payment to private hospitals since August 2025. Republic World quotes an official confirming the MoU “mandates payment within 15 days (accruing interest after the 16th day)”.
  • 9 June 2026 — 135 J&K hospitals and dialysis centres announce a scheme exit from 1 July over ₹275 crore in dues, citing “Section 6 of the Memorandum of Understanding” under which they are “entitled to one per cent interest per week on delayed payments, which has not been paid”.
  • 23 July 2026 — Himachal Pradesh High Court, M/S Maatri Medicity & Orthocare Hospital v. State of HP: orders release of ₹25.22 crore of approved AB-PMJAY bills within three weeks “with interest at 1% of the claim amount per week after 15 days of delay”, and directs the same interest on all future delayed payments. The state “cannot keep on evading its financial obligations under the pretext of… a vigilance inquiry”.
  • 29 August – 1 September 2026 — Haryana: IMA says ~650 hospitals are owed ~₹1,200 crore pending more than six months against the 15-day term. Services suspended for 24 hours on 1 September; indefinite withdrawal threatened from 16 September. Demands include “proper justification for deductions” and “a fair and transparent claim-settlement process”.
  • 28 August 2026 — IRDAI’s Insurance Advisory Committee calls for mandating universal integration with the National Health Claims Exchange (NHCX), which already reports 83 payers, 42,687 provider facilities and 23.4 million claims. Hospital claims data is becoming structured, machine-readable and portable.

This is not new pain. AHPI, representing ~15,000 private hospitals, wrote to the government in May 2024 demanding interest on late payment and reporting average delays of “6-8 months”. A Rajya Sabha reply in April 2025 admitted 609 private hospitals had exited the scheme since 2018. Punjab’s IMA nursing-home cell suspended cashless services over ₹600 crore. What changed in 2026 is that the interest entitlement went from an association talking point to a court order, and the deduction dispute went from grumbling to a headline demand. Every one of those 650 Haryana hospitals now needs a document it does not have: a per-claim statement of what is owed, since when, at what interest, and which deductions it is contesting inside the 15-day appeal window.

The gap in the tooling. The NHA’s Transaction Management System (TMS) shows a hospital each claim’s stage: submitted, under review, approved, settled. It does not show aging against the 15-day clock, does not compute interest, does not reconcile the approved amount against the bank credit that finally arrives (often net of a 10% TDS deduction in several states), and does not aggregate deduction reasons. Terra Insight, the one vendor that writes about PM-JAY reconciliation, says so directly: hospitals must manually match “TMS-approved claim amount against the actual bank credit received” and separately investigate why “settlement amounts may differ from approved package rates”. Their product is an enterprise reconciliation engine with a 2–4 week configuration project, pitched at hospitals doing 500-plus claims a month. Every HMS vendor with a PM-JAY module (MocDoc, Zospital, CureNearMe, Healthray) sells the front end — eligibility, pre-auth, claim submission via NHCX. None sells the back end: aging, interest, deduction appeals, grievance filing.

Provenance:

3. The opportunity

The incumbent here is not a software company. It is the hospital’s own accounts clerk, an Excel sheet, and the IMA district secretary’s WhatsApp group. When the SHA pays, it pays a batch: one bank credit covering dozens of claims, net of deductions the hospital learns about only by opening each claim in TMS, net of TDS, months after approval. When the SHA deducts, the reason sits in a system-generated rejection letter per claim, and the hospital has 15 days from rejection to appeal to the District Grievance Nodal Officer or the money is gone. When the SHA is late, the hospital is owed 1% per week under its MoU, but nobody at the hospital has ever computed that number, because the officials “deny interest for delayed payments” (Tribune, July 2025) and the hospital has no document to argue with.

The result is three leaks that nobody is plugging:

  1. Appeal-window leak. Rejections and partial payments arrive piecemeal. A hospital doing 200 claims a month gets rejection letters daily. The 15-day appeal clock starts on each one. Most lapse. The NHA’s own manual says partial payments can be reconsidered through the TMS erroneous-claim section once documents are produced — a route almost no small hospital works systematically.
  2. Interest leak. The 15-day term and the 1%-per-week clause exist in the MoU. The HP High Court has now enforced them. But interest is a computed entitlement: it needs the submission date, the approval date, the credit date and the claim amount for every claim, rolled up into a demand. No hospital in Haryana can produce that sheet today. That is why the IMA argues in crores and the SHA answers in press releases.
  3. Reconciliation leak. Approved ≠ credited. Terra Insight’s own framing: at 500 claims a month, “spreadsheet tracking produces systematic errors — missed claims, unreconciled bank credits, and package rate variances.” Below 500 claims, hospitals don’t even have the spreadsheet.

Bakaya is the ledger that sits between TMS and the bank statement and turns each leak into a dated, drafted action: appeal this rejection by Thursday, reconcile this batch credit against these 43 claims, and here is the interest statement — claim by claim, week by week — to hand your IMA district secretary or your lawyer.

Why HMS vendors haven’t done it: their module is sold on getting the claim in, and their buyer is the hospital IT lead at empanelment time. The recovery side is bought by the owner at the moment the cheque doesn’t come, which is a different moment, a different buyer, and a document rather than a workflow. Why Terra Insight hasn’t done it for the small hospital: their engine is configuration-heavy and priced for the volumes where a 2–4 week project makes sense. The 30–150 bed hospital doing 50–300 Ayushman claims a month is exactly the band both incumbents step over — and it is the band that IMA Haryana says is taking loans to pay salaries.

4. Target market

  • Primary customer: Owner-director or administrator of a private hospital with 30–150 beds, empanelled under AB-PMJAY in a trust-model state, doing 50–300 PM-JAY claims a month, where Ayushman is 20–60% of inpatient revenue. In practice this is the nursing-home tier that dominates IMA Haryana’s 650, Punjab’s nursing-home cell, the 135 J&K institutions, and the Kangra association. The user is the hospital’s accounts officer (the TMS “ACO” role) and the medical coordinator (MEDCO) who handles queries; the payer is the owner.
  • Why they buy: in their words — “Small and medium-sized hospitals are on the brink of collapse. Many have been forced to take loans to pay salaries” (Dr Naresh Vermani, Kangra association president, Tribune, 24 Mar 2025). “Payments are only released after repeated warnings of service withdrawal, and that too every 3-4 months” (Dr Mahaveer P Jain, IMA Haryana president, Tribune, 10 Jul 2025). “Clearing the pending amount is not enough because this is happening after every few months” (Dr Dhirendar Soni, IMA Haryana secretary, Medical Dialogues, 7 Aug 2025). They want the money, they want the deductions explained, and they want a paper trail that survives the next round of “payments will be made on time”.
  • Rough TAM reasoning: ~16,000 private empanelled hospitals nationally. Strip out the tertiary chains (who have RCM teams) and the tiny single-specialty centres with under 20 claims a month, and roughly 5,000–7,000 hospitals sit in the target band. In the acute states alone — Haryana (~650–700), Punjab, J&K (135), HP, Gujarat (233 exits since 2018 suggests thousands still in), Telangana/Andhra (AHPI names them for chronic delay) — the first 1,000 are reachable through six associations.
  • Why now for them: the September 2026 Haryana standoff, the June 2026 J&K exit notice and the July 2026 HP judgment have made “what exactly are we owed, with interest” the question every hospital owner is being asked by their association and their bank. The answer today is a guess.

5. Product sketch (MVP)

  • Claim ledger import — the hospital uploads its TMS claim reports and bank statements (Excel/PDF); Bakaya builds one row per claim with submission, approval, credit dates and amounts, and matches batch credits to claims, TDS included.
  • 15-day aging board — every approved-but-unpaid claim shown against the contractual clock, sorted by days late and rupees at risk; weekly summary to the owner on WhatsApp.
  • Interest statement — per-claim interest computed under the hospital’s MoU term (1% per week after day 15 by default, editable per state; the 0.1%-per-day KPI variant for insurance-model states), rolled into a dated demand letter addressed to the SHA CEO with the claim schedule annexed.
  • Deduction and rejection inbox — every rejection letter and partial payment parsed for the stated reason, bucketed (documentation, package mismatch, pre-auth timing, clinical justification), with the 15-day DGNO appeal deadline on each.
  • Appeal drafter — for each contestable deduction, a draft grievance in the CGRMS format citing the package code, the pre-auth approval, and the manual’s reconsideration clause, with the document checklist the hospital must attach.
  • Association pack — one-click export of the hospital’s pending schedule in the format IMA district committees and High Court petitions actually annex: claim ID, package, approved amount, days pending, interest accrued.
  • Query watch — open TMS queries surfaced with their response deadline, because “Delayed or no query reply” is a named rejection ground in the NHA manual.

6. AI angle — what’s load-bearing

Two things are impossible without a language model at this price point.

First, reading the payer’s paper. Rejection letters, deduction remarks, query text and partial-payment notes come as free text, in English and Hindi, inconsistent across SHAs and claim-processing doctors. Turning “documents not legible / discharge summary lacks investigation reports / package 1.2.3 not payable with 4.5.6” into a structured reason, a category, a recoverability verdict and a checklist of what to attach is a reading-comprehension job over thousands of near-duplicate but not identical notes. A rules engine breaks on the third state.

Second, drafting the reply. The appeal to the DGNO and the reconsideration request through TMS are semi-formal documents that must cite the right package code, the pre-auth number, the manual clause and the clinical facts from the discharge summary. Today the MEDCO writes them from scratch or doesn’t write them. Bakaya drafts them from the claim record, the rejection text and the hospital’s own documents, for the MEDCO to sign.

Remove the AI and you have a spreadsheet template with a date-diff formula. That is what hospitals have now. It isn’t working.

7. Localization angle (if any)

This is India-only by construction and the local quirks are the product.

  • Language: TMS and SHA correspondence run in English; the owner and accounts clerk in Haryana, Punjab and J&K work in Hindi/Punjabi/Urdu. Weekly WhatsApp summaries and appeal drafts in Hindi with English annexures.
  • Pricing: ₹4,999–14,999 per month by claim volume. Benchmark: HMS subscriptions at ₹10,000 a year for small hospitals; empanelment consultants at ₹25,000–1 lakh one-time; a High Court writ at several lakh. A hospital with ₹40 lakh stuck for six months is owed roughly ₹9–10 lakh in contractual interest — the subscription is noise against that.
  • Distribution: IMA state and district chapters, AHPI, state nursing-home associations, and the WhatsApp groups where the 650 Haryana hospitals coordinated a suspension. These are the most organised small-business networks in India right now, and they are actively looking for the document Bakaya produces.
  • State variance: trust vs insurance model, portability 30-day rule, state-specific packages, TDS treatment, and the MoU clause wording differ by SHA. Encoding that per state is the moat and the reason a generic RCM tool won’t bother.

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

  • Pricing: ₹4,999/month (≤100 claims), ₹9,999/month (≤300 claims), ₹14,999/month (unlimited, multi-unit). Optional 10% success fee on deductions recovered through appeals drafted in Bakaya, capped per claim. No success fee on interest — that fight is the association’s and the court’s, and pricing on it would make the product look like a lawyer.
  • ACV: ₹1.1 lakh ($1,300) blended, before success fees.
  • Rough math to $1M ARR: 800 hospitals × ₹9,000/month × 12 = ₹8.6 crore ≈ $1.0M. That is fewer hospitals than are currently suspending services in Haryana alone.
  • Rough math to $5M ARR: 3,000 hospitals on subscription (~half the target band) plus success-fee revenue on recovered deductions, plus the natural adjacency — the same ledger for private-insurer/TPA claims moving onto NHCX, where the deduction-and-delay problem is identical and the hospitals are the same.
  • Expansion path: per-unit pricing for multi-hospital owners; the NHCX private-insurer ledger as a second module; state-scheme variants (Chiranjeevi/MAA in Rajasthan, Aarogyasri in Telangana/AP) where AHPI says the merger into PM-JAY caused the delays in the first place.

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

  • Haryana first, this month. The IMA Haryana chapter has publicly named the problem, the number (₹1,200 crore), the count (~650 hospitals), and the demand (justified deductions, transparent settlement). Offer the state president and district secretaries a free “pending-dues-with-interest statement” for any member hospital that uploads its TMS export — the document the chapter needs for its 16 September deadline and its next meeting with the SHA. Convert 10% of participating hospitals to paid within 60 days when the statement becomes a living ledger rather than a one-off PDF. 650 hospitals × 30% participation × 10% = ~20 paying customers from one chapter.
  • The 135 J&K institutions and the HP petitioners. Both groups are in active litigation or exit posture and have already cited the MoU interest clause in public. Their lawyers need the annexure Bakaya produces. Approach the J&K Private Hospitals and Dialysis Centres Association spokesperson and the Kangra association directly with a worked example built from public package rates. Target 15 customers across the two.
  • Punjab nursing-home cell. ₹600 crore pending, services suspended, a High Court date already in the record. Same play as Haryana, run by the same kind of secretary. Target 25.
  • Empanelment consultants as resellers. The consultancies that get hospitals empanelled (Hospertz, MantraCare, hospitalempanelment.com and dozens of regional operators) have the hospital’s trust and no post-empanelment product. Offer 20% recurring on referred hospitals. Target 20 customers in 90 days through five consultants.
  • Medical Dialogues and Tribune readers. Every article in this file has a comment thread of hospital owners. A monthly “Ayushman dues tracker” — state-wise pending days computed from public statements — published as a free page earns inbound from the remaining states before we have a chapter relationship there. Target 20 inbound.

10. Build complexity — justification

Medium. The ingestion problem is real but bounded: TMS provider-side MIS reports and Excel exports, bank statements from a dozen Indian banks, and free-text rejection letters. Reasoning over the rejection text and drafting appeals is a standard LLM job over a small, growing corpus. The state-rule layer (trust vs insurance model, portability, TDS, MoU clause) is configuration, not code. No TMS API is required for v1 — the hospital uploads what it can already download — which removes platform-approval risk but means a manual step until NHCX exports mature. Two people, 10–12 weeks to a v1 that produces an accurate aging ledger, interest statement and appeal drafts for one state; a further month per additional state model.

11. Gating checklist

GatePass?Note
Legal in target market✅Hospital processes its own claim and bank data; appeals and grievances are rights spelled out in the NHA manual (§5.1.12, §5.1.14). No patient clinical data needs to leave the hospital beyond what is already in the claim record.
Ethical — no harm / dark patterns✅Helps hospitals recover money they are contractually owed; keeps Ayushman beneficiaries’ hospitals solvent. Guardrail: refuse to draft appeals that misstate clinical facts.
Market exists (evidence above)✅₹1,200 crore (Haryana), ₹600 crore (Punjab), ₹275 crore (J&K) in dispute; HC precedent on interest; AHPI’s 15,000-hospital demand.
1–5 person team can build this✅Two builders, one with hospital-billing exposure.
Launchable with <$50K / ₹40L✅₹8–12 lakh to first paying chapter.

12. Feasibility score

AxisWeightScoreNotes
Problem intensity2017/20Hospitals shutting, borrowing to pay salaries, suspending services. Hair-on-fire. Docked because the product arms the claim rather than making the state pay.
Demand evidence1512/15Multiple independent signals across four states and two years, a court order, an association demand. No one is yet paying for this exact tool, which is the point and the risk.
Build feasibility1512/15Excel/PDF ingestion plus LLM reading and drafting; no API dependency in v1. State-model configuration is grind, not research.
Distribution clarity1511/15Named chapters, named counts, an active deadline. Uncertainty: converting an association relationship into individual hospital subscriptions.
Revenue mechanics159/15Pricing is cheap against the money at stake, but small-hospital software spend is historically low (₹10K/year HMS benchmarks) and the interest may never actually be paid outside a courtroom.
Time to first revenue107/10Free statement to paid ledger in 60 days is plausible given the September deadline; hospitals in cash crisis are also slow to add costs.
Defensibility104/10Execution moat plus an accumulating corpus of deduction reasons and per-state rules. HMS vendors could bolt on aging; they are unlikely to touch appeals and interest, but nothing stops them.
Total10072/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · sales-heavy · domain-expertise-required — someone who has sat in a hospital accounts office watching an SHA batch credit land, or can hire that person for the first month.

Key assumptions to validate (3–5)

  1. Assumption: A hospital can download from TMS (or already keeps) claim-level data with submission, approval and settlement dates sufficient to compute aging without portal automation. How to test: Get TMS exports from five Haryana hospitals via the IMA chapter and build the ledger by hand in week one.
  2. Assumption: Owners will pay a subscription for a recovery ledger rather than expecting the association or their HMS to provide it free. How to test: After delivering 30 free statements, offer paid conversion at ₹4,999 and count.
  3. Assumption: A meaningful share of deductions are recoverable through the 15-day DGNO appeal or the TMS reconsideration path when properly documented. How to test: Run 50 appeals across three hospitals and measure rupees restored within 60 days.
  4. Assumption: Interest statements, even if not paid by the SHA voluntarily, are valued because they anchor negotiation and litigation. How to test: Ask IMA Haryana and the J&K association whether they would annex Bakaya’s schedule to their next representation; a yes from either is the signal.
  5. Assumption: The per-state rule variance is configuration, not a rewrite. How to test: Build Haryana, then attempt Punjab in under two weeks.

Risk flags

  1. Political risk: a state clearing its backlog in one tranche removes the urgency for that state for 3–4 months — Haryana’s own pattern per IMA. Mitigation: aging and deduction appeals are recurring even when the backlog is cleared; sell the ledger, not the crisis.
  2. Platform dependency: TMS export formats and the shift to NHCX-based flows can change without notice. Mitigation: ingest what the hospital can download today; treat NHCX as an upgrade, not a dependency.
  3. Regulatory optics: a tool that helps hospitals demand interest from the state could draw SHA hostility toward hospitals seen using it. Mitigation: position as reconciliation and appeals; the interest statement is one export among several.
  4. Willingness-to-pay floor: small hospitals in a cash crunch are the worst-timed buyers. Mitigation: the association-sponsored free statement gets the data in; the subscription is sold when the first batch credit reconciles.

14. Structured verdict

Score:                  72/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder paired with an ex-hospital accounts/TPA-desk operator; Hindi-speaking; comfortable in IMA meeting rooms
Time to revenue:        8–10 weeks (free statement to first paid conversions inside the Haryana deadline cycle)
Capital to launch:      ₹8–12 lakh ($10–14K)
Top 3 assumptions to validate first:
  1. TMS exports from 5 hospitals contain the dates needed to compute aging and interest — build by hand in week one
  2. 10% of hospitals receiving a free dues-with-interest statement convert to a paid ledger within 60 days — run through IMA Haryana
  3. Properly drafted appeals recover a measurable share of deductions — 50 appeals across 3 hospitals, rupees restored in 60 days
Kill criteria:
  - Abandon if <5 of 30 hospitals receiving the free statement convert to paid within 60 days
  - Abandon if TMS data available to hospitals lacks approval/settlement dates and no workaround yields per-claim aging
  - Abandon if 50 appeals recover <5% of contested deduction value in 60 days and associations decline to annex the interest schedule

15. Next step — 1-week validation sprint

  • Day 1–2: Call the IMA Haryana state secretary and two district secretaries (Karnal is on record). Offer a free pending-dues-with-interest statement for any member hospital before the 16 September deadline. Collect TMS exports and bank statements from five volunteers.
  • Day 3–4: Build the five ledgers by hand: aging against 15 days, interest at 1%/week, batch-credit reconciliation, deduction buckets with appeal deadlines. Hand each hospital its statement and one drafted appeal. Note every field that TMS did not provide.
  • Day 5: Decide go / no-go on three numbers: did all five exports contain the dates needed; did at least three of five owners ask to keep the ledger updated monthly at a stated price; did the chapter agree to annex the schedule to its SHA representation. Two of three is a go.

Interested in a detailed proposal?

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

Contact us

info@startupbasket.ai