RenewalDesk

Never miss an insurance renewal again.

RenewalDesk is a WhatsApp-first renewal tracker for individual insurance agents, POSPs, and small brokerages in India. It ingests insurer PDF policy lists, organizes renewals by insurer and month, and helps agents send approved customer reminders in local languages. The product is valuable because it protects commissions, reduces manual tracking, and fits the way agents already work day to day.

Business Goals

  • Reach 10 paying agents by the end of week 12 and 150 paying agents by month 12.
  • Achieve at least 25% conversion from activated trial users to paid subscriptions by month 6.
  • Keep monthly churn below 8% for the first 12 months, with a target of 5% by month 12.
  • Maintain gross margin above 70% on the Starter and Pro plans at typical usage.
  • Drive at least 30% of new paid signups from referrals and warm outbound by month 12.

User Goals

  • Let an agent upload insurer PDF lists and see all renewals organized within 10 minutes.
  • Reduce missed or forgotten renewals by surfacing daily and monthly reminders.
  • Create customer-ready WhatsApp drafts in the agent’s preferred language with one tap.
  • Make it easy to confirm extracted policy data before any reminder is sent.
  • Help small brokerage teams share renewal workload without duplicating spreadsheets.

Non-Goals

  • Do not scrape insurer portals or automate login-based data extraction in v1.
  • Do not auto-send customer messages without explicit agent approval in v1.
  • Do not provide policy-buying advice, insurer ranking advice, or commission-based recommendations.
  • Do not build a full CRM, claims-management suite, or lead marketplace in the initial release.

Independent Agent Ravi, 39 - Ravi sells health and general insurance to about 350 customers across multiple insurers. He currently tracks renewals in Excel and WhatsApp chats and loses time reconciling due dates each week.

Independent Agent Ravi, 39

  • As an independent agent, I want to upload insurer PDF lists in bulk, so that I can stop manually entering every policy.
  • As an independent agent, I want a daily digest of upcoming renewals, so that I can focus on the most urgent customers first.
  • As an independent agent, I want a WhatsApp draft in the customer’s language, so that I can send reminders quickly and consistently.

POSP Neha, 28 - Neha works mostly from her phone and handles a smaller book of business but is highly active on WhatsApp. She needs a simple workflow that takes very little setup and works with low-tech documents from multiple insurers.

POSP Neha, 28

  • As a POSP, I want the system to extract renewal dates from insurer PDFs, so that I do not have to type them manually.
  • As a POSP, I want to filter renewals by month and insurer, so that I can plan outreach around due dates.
  • As a POSP, I want to review every extracted field before sending, so that I can trust the reminders.

Small Brokerage Owner Suresh, 51 - Suresh manages a team of 6 agents and wants visibility into renewals across the brokerage. He needs shared access, consistent reminders, and accountability without adopting a complex enterprise system.

Small Brokerage Owner Suresh, 51

  • As a brokerage owner, I want team accounts and shared queues, so that renewals are not missed when one agent is busy.
  • As a brokerage owner, I want to see which renewals were contacted and by whom, so that I can monitor follow-through.
  • As a brokerage owner, I want role-based access, so that each agent only sees their own customer data unless I allow sharing.

PDF Intake and Extraction · High priority

  • Allow users to upload insurer policy PDFs and convert them into structured renewal rows grouped by insurer and due month.
  • Support bulk upload of multiple PDFs in one session.
  • Detect and map common columns such as customer name, mobile number, policy number, insurer, premium, and renewal date.
  • Show extraction confidence and flag low-confidence rows for review.
  • Handle scanned, blurry, password-protected, and mixed-language PDFs with OCR fallback.
  • Allow manual correction of any field before saving the record.

Renewal Dashboard · High priority

  • Provide a clear dashboard that organizes renewals by insurer, month, and search criteria so agents can work from one screen.
  • Show one section per insurer with counts of upcoming renewals.
  • Allow month filtering, date range filtering, and keyword search.
  • Highlight overdue, due-today, and upcoming renewals separately.
  • Persist uploaded files and extracted rows so agents can return later without reprocessing.
  • Support export to spreadsheet format for offline review.

WhatsApp Drafting and Digest · High priority

  • Generate reminder drafts and agent digests in WhatsApp-first workflows without sending anything automatically in v1.
  • Create a daily or monthly WhatsApp digest to the agent with priority renewals.
  • Generate one customer-specific WhatsApp draft per renewal with editable text.
  • Offer at least three language options for drafts, starting with English, Hindi, and one additional Indian language based on pilot demand.
  • Require the agent to approve each message before it is considered sent.
  • Include opt-out and compliance-friendly wording in default templates.

Review, Approval, and Audit Trail · High priority

  • Ensure every extracted renewal and every customer message is reviewed and traceable before action.
  • Require explicit confirmation of extracted renewal date and customer identity before scheduling reminders.
  • Keep an edit history for corrections to extracted fields.
  • Record who approved each draft and when.
  • Provide a pause switch for an agent or brokerage account.
  • Maintain an audit log of upload, edit, draft, approval, and send events.

Admin, Billing, and Access Control · Medium priority

  • Support paid subscriptions, plan limits, and user permissions for individual agents and small teams.
  • Enforce plan-based limits on customer count, number of insurer files, and number of users.
  • Support monthly and annual billing with GST added on top.
  • Allow a trial with no card for a limited customer count.
  • Support role-based access for brokerage owners, agents, and assistants.
  • Show plan usage and upgrade prompts when limits are approached.

Fast Onboarding for Busy Agents

  • Create account with phone number and WhatsApp verification.
  • Choose language and select whether you are an individual agent or brokerage.
  • Upload one insurer PDF to test extraction and confirm column mapping.
  • Review extracted rows, fix obvious errors, and approve the first renewal list.
  • Receive the first WhatsApp digest within 10 minutes of upload.
  • Invite teammates or upgrade the plan after the first successful renewal batch.

1. Sign Up and Set Up

  • The agent creates an account, verifies a mobile number, and chooses their preferred language and customer communication style.
  • Default to mobile-first onboarding with minimal typing.
  • Pre-fill plan limits and explain the trial clearly.
  • If the agent does not know their exact customer count, allow an estimate and update later.

2. Upload Insurer PDF

  • The agent uploads one or more insurer PDFs from LIC, HDFC, Star Health, or similar providers and waits for extraction.
  • Show file status immediately and queue processing in the background.
  • If a file is password-protected or unreadable, explain the reason and suggest OCR or manual entry.
  • Allow multiple PDFs to be uploaded in one session.

3. Confirm Column Mapping

  • The system proposes extracted fields and column mapping once per insurer format, then asks the user to confirm or edit.
  • Persist insurer-specific mapping so future uploads are faster.
  • Flag missing fields such as mobile number or renewal date.
  • Show confidence scores and clearly mark low-confidence rows.

4. Review Renewals by Month

  • The dashboard organizes all renewals by insurer, month, and priority so the agent can focus on what is due next.
  • Highlight urgent items with clear status labels.
  • Support search by customer name, policy number, or phone number.
  • If duplicates are detected, ask the user to merge or keep both records.

5. Draft and Approve Messages

  • The agent opens a customer record, sees a ready WhatsApp draft, edits if needed, and approves it.
  • Use local-language templates with the agent’s name and customer details inserted.
  • Do not send until the agent confirms the final text.
  • Show a five-minute undo window and a clear sent status once approved.

Power Features and Edge Cases

  • Team accounts for small brokerages with role-based permissions.
  • Per-insurer extraction quality dashboard to spot broken formats quickly.
  • Annual prepay billing and plan usage alerts.
  • Reminder scheduling rules by insurer, month, or customer segment.
  • Fallback manual entry for unreadable documents.
  • Pause mode for compliance review or seasonal campaign control.

Designed for WhatsApp-First Operators

  • Mobile-first layout with large tap targets and minimal form fields.
  • Clear status chips for extracted, reviewed, approved, sent, and failed states.
  • Fast loading on low-end Android phones and weak network connections.
  • Accessible color contrast and text sizes suitable for outdoor use.
  • Persistent save states so users never lose corrections during a bad connection.
  • Plain-language copy that avoids insurance jargon wherever possible.

Ravi runs a small insurance agency and spends each Monday opening spreadsheets, insurer PDFs, and WhatsApp chats to figure out which customers are due. He misses some renewals every month because dates are scattered across files and his reminders are written by hand.

With RenewalDesk, Ravi uploads insurer PDFs in bulk and sees all renewals organized by insurer and month. The app drafts WhatsApp reminders in Hindi and English, and Ravi approves each one from his phone before sending.

Within a few weeks, Ravi is saving multiple renewals per month and spending less time on admin. The business impact is direct: fewer lapses, protected commissions, and a simple workflow that fits how he already communicates with customers.

User-Centric Metrics

  • At least 80% of active users upload and review their first renewal list within 10 minutes of onboarding.
  • At least 70% of uploaded policies are extracted with no manual re-entry required.
  • At least 2 renewals are saved per active agent per month by week 12.
  • At least 60% of reminder drafts are approved and sent within 24 hours of creation.
  • At least 75% of users rate the product as easy to use in post-onboarding feedback.

Business Metrics

  • 10 paying agents by week 12 and 150 paying agents by month 12.
  • Trial-to-paid conversion above 25% by month 6.
  • Monthly churn below 8% for the first year.
  • At least 30% of new signups acquired through referrals, community posts, or warm outbound.
  • Gross margin above 70% on Starter and Pro plans at typical usage.

Technical Metrics

  • 99.5% monthly uptime for the core dashboard and upload pipeline.
  • Extraction job completion within 2 minutes for a standard PDF set of up to 500 rows.
  • Customer-facing data stored encrypted at rest and in transit.
  • No unauthorized message sends; 100% of outbound reminders require explicit approval in v1.

Tracking Plan

  • track_account_created with source, plan selected, and language.
  • track_pdf_uploaded with insurer name, file type, page count, and size.
  • track_extraction_completed with confidence score, row count, and error flags.
  • track_field_corrected with field name, old value, and new value.
  • track_digest_viewed with digest frequency and time to open.
  • track_draft_approved with language, channel, and send timing.
  • track_payment_started and track_payment_completed with plan, billing cycle, and GST amount.

Technical Needs

  • Frontend built with Next.js and TypeScript for a fast mobile-first web app.
  • Backend API with Node.js or NestJS handling uploads, extraction jobs, drafts, and billing.
  • Background job queue using BullMQ or a similar worker system for OCR and AI tasks.
  • Relational database such as PostgreSQL for customers, insurers, policies, and audit logs.
  • Object storage such as Amazon S3 or Cloudflare R2 for PDFs and generated artifacts.
  • Secure authentication with OTP-based login and session management.
  • Observability stack with structured logs, error tracking, and job retry monitoring.

Integration Points

  • WhatsApp Business API through an approved provider such as Meta Cloud API or a BSP.
  • OCR/extraction service such as Google Document AI, Azure Form Recognizer, or Tesseract fallback.
  • LLM provider for drafting and field normalization such as Claude, Gemini, or OpenAI.
  • Payment gateway such as Razorpay or Cashfree for Indian subscriptions and GST receipts.
  • Email or SMS fallback for operational notifications if WhatsApp delivery is unavailable.

Data Storage & Privacy

  • Store customer policy data only with explicit agent consent and a clear privacy notice.
  • Treat the agent as the controller and the platform as a processor under the data-processing agreement.
  • Encrypt data in transit with TLS and at rest with managed encryption keys.
  • Support deletion requests by customer, policy, or entire account with a documented retention policy.
  • Keep detailed audit logs for access, edits, approvals, and outbound messages, with limited admin visibility.

Scalability & Performance

  • Process uploads asynchronously so the UI stays responsive even for large PDFs.
  • Cache insurer-specific extraction mappings to reduce repeated processing cost.
  • Design the system for low-bandwidth mobile usage and intermittent connectivity.
  • Use queue-based rate limiting for AI and WhatsApp requests to avoid provider throttling.

Potential Challenges

  • Insurance PDFs vary widely by insurer and format; mitigate with insurer-specific mappings and a regression test set.
  • Extraction errors can damage trust; mitigate with confidence scoring, human confirmation, and visible source PDF review.
  • WhatsApp template rejection or pricing changes can hurt margins; mitigate with approved utility templates and cost monitoring.
  • Regulatory and privacy obligations are significant; mitigate with legal review, consent records, and India-based storage.
  • User acquisition may be slow in a fragmented market; mitigate with founder-led interviews, warm outreach, and referral loops.

Team & resourcing - Small bootstrap team - 1 founder PM, 1 full-stack engineer, part-time designer, and optional contractor for OCR/integration work.

Phase 1: Concierge Validation · Weeks 1-4

  • Manual service for the first 5 agents using uploaded PDFs and AI-assisted extraction.
  • Insurer-wise renewal sheets and a WhatsApp digest sent to the agent only.
  • Interview notes, pain-point validation, and a repeatable intake checklist.
  • Baseline extraction accuracy measurements across at least 20 real policy PDFs.

Phase 2: MVP Build · Weeks 5-12

  • Bulk PDF upload, insurer-aware column mapping, and editable renewal dashboard.
  • Month filters, search, and per-insurer sections.
  • WhatsApp draft generation in three languages with approval-first workflow.
  • Basic billing, onboarding, and audit logs.
  • Launchable trial for up to 100 customers per agent.

Phase 3: Paid Launch · Months 4-6

  • Plan limits, GST-aware pricing, and annual billing.
  • Daily and monthly WhatsApp digest scheduling.
  • Improved OCR fallback and unreadable-file handling.
  • Usage analytics, churn tracking, and onboarding nudges.
  • First small brokerage team accounts.

Phase 4: Scale and Automation · Months 7-12

  • WhatsApp API sending after approval with undo window.
  • Quote comparison inputs and renewal workflow enhancements.
  • Team permissions, shared queues, and role-based access.
  • Operational dashboards for extraction accuracy, send success, and margin monitoring.

Paste this into Cursor, Bolt, Lovable, or v0 to start building.

Build a mobile-first SaaS web app called RenewalDesk for insurance agents in India.

Product summary:
A WhatsApp-first renewal tracker where independent insurance agents, POSPs, and small brokerages upload insurer PDF customer lists, review extracted renewal data, see renewals organized by insurer and month, and generate approved WhatsApp reminder drafts in Indian languages. v1 must be approval-first: no automatic sending to customers.

Core features:
1. Auth with phone number OTP
2. Onboarding that captures role, language, customer count, and plan trial status
3. Upload multiple insurer PDFs
4. Background extraction pipeline for OCR plus structured field extraction
5. Insurer-specific column mapping that is saved per insurer format
6. Renewal dashboard with sections by insurer, filters by month, overdue/due-today/upcoming states, and search
7. Editable renewal records with confidence scores and audit history
8. WhatsApp draft generation for each customer in English, Hindi, and one additional Indian language
9. Daily or monthly digest to the agent only
10. Approval-first send flow with a 5-minute undo window and send status tracking
11. Billing plans with trial, monthly, annual, GST calculation, and usage limits
12. Team accounts for small brokerages as a later-ready but simple role model

Primary screens and flows:
Onboarding, upload screen, extraction review screen, renewal dashboard, customer detail screen, draft approval modal, digest settings, billing page, admin/audit log page.

Data model:
Users, Organizations, Roles, Plans, Subscriptions, InsurerProfiles, UploadedFiles, ExtractJobs, PolicyRecords, CustomerContacts, RenewalEvents, MessageDrafts, MessageApprovals, MessageSends, AuditLogs, LanguagePreferences, UsageLimits.
Each policy record should store insurer name, customer name, phone number, policy number, renewal date, premium, language, confidence score, source file, and approval status.

Tech stack:
Use Next.js 14, React, TypeScript, Tailwind CSS, PostgreSQL, Prisma, Redis/BullMQ for background jobs, S3-compatible storage for PDFs, and a serverless or containerized Node.js API. Use a WhatsApp Business API provider, an OCR service for PDFs, and an LLM API for message drafting. Include role-based access control, encrypted storage, structured logs, error tracking, and analytics event hooks.

Implementation notes:
Build for India-first mobile use, low bandwidth, and low-end Android devices. Make the UI simple, large-tap, and fast. Require explicit user confirmation before any customer message is marked ready to send. Add importer validation, duplicate detection, and readable error states for password-protected or blurry PDFs. Create seed data and simple admin tools so the team can test insurer-specific formats quickly.

Output a working product scaffold with reusable components, database schema, API routes, job queue workers, and polished responsive UI for the key flows above.

Business Idea

RenewalDesk: Business Plan for India Updated 16 Oct 2026 · Bootstrapped · Founders: 1 · Status: first agent in talks, no paying customers yet 1. Executive summary What we build: A WhatsApp-first AI copilot for individual insurance agents, POSPs and small brokers in India. The agent uploads the client PDF from each insurer portal (LIC, HDFC, Star Health and others). The software reads the rows, shows one dashboard section per insurer with renewals by month, sends the agent a daily or monthly WhatsApp digest, and prepares a one-tap WhatsApp draft for each customer in their language. The agent approves every customer message. Why it matters: Agents lose commission whenever a renewal slips. Most track renewals in Excel, diaries or memory. One saved renewal a month covers the subscription. Price: ₹299, ₹499 or ₹1,299 a month (150, 500 or 1,500 customers). Yearly is 10 months' price: ₹2,999, ₹4,999 or ₹12,999. GST of 18% is added on top. Margin is about 76% to 81% at typical use and 65% to 75% if an agent fills the plan. Goal for 12 months: 150 paying agents (base case), about ₹87,000 a month at a blended ₹579 per agent. Break-even is about 13 agents on a lean ₹6,000 monthly cost, and about 42 agents once we add ₹20,000 of growth spending. The one thing we must prove first: that agents will pay for bulk-upload renewal tracking. One agent has said he would pay ₹15,000 a year. Before building the full product, 8 of 20 agents must name renewals as a top pain and 5 must pay. 2. Problem • An agent with 300 to 500 customers has a renewal falling due almost every working day, each with a different insurer, date and product. • Tracking is manual. Missed or late renewals mean the customer buys elsewhere or lapses, and the agent loses the commission. • Comparing renewal options across insurers takes time, so many agents skip it. • Customers expect WhatsApp and local-language communication, which agents write by hand. 3. Solution and product scope Version What it does Human in the loop Concierge (weeks 1 to 4) Founders run it by hand for the first 5 agents. Each agent sends one PDF per insurer; we read it with an AI chat tool, check every row, and send back a renewal sheet by insurer and month plus the agent's WhatsApp digest Founder checks every row; we send nothing to customers v1 (weeks 5 to 12) Bulk PDF upload per insurer, column mapping confirmed once, one dashboard section per insurer, month filter and search, daily or monthly WhatsApp digest to the agent, click-to-WhatsApp draft for each customer in 3 languages Agent approves every customer message v2 (months 4 to 8) Automatic sending through the WhatsApp API after approval, quote comparison from agent-entered quotes, team accounts for small brokerages Agent approves templates; undo window Later Cross-sell prompts, customer self-service links, commission tracking Opt-in only Deliberate non-goals for v1: scraping insurer portals, auto-sending without approval, giving advice on which policy to buy. These carry the highest legal and reliability risk. 4. Market • IRDAI's 2024-25 annual report puts individual insurance agents at about 31.23 lakh, up 7.87% on the year. Health premiums reached ₹1,17,505 crore. • Older IRDAI data (FY2020-21) counted about 8.28 lakh point-of-sale persons (POSPs), a segment focused on standard motor, health and travel products. • Not all agents are reachable. My working assumption (to be tested): 20% are active, smartphone-fluent and willing to pay for software, about 6.2 lakh agents. Segment Agents Annual revenue at a blended ₹579 a month All individual agents (theoretical ceiling) 31.2 lakh ₹2,170 crore Serviceable (assumed 20%) 6.2 lakh ₹434 crore Realistic 3-year target (assumed 1% of serviceable) 6,246 ₹4.3 crore The 20% and 1% figures are planning assumptions, not sourced facts, and match the competitor analysis. The blended ₹579 assumes 40% Starter, 40% Pro and 20% Business on monthly billing, before GST. Agent interviews should replace these guesses. 5. Customer and competition Primary customer: an individual agent or POSP, general and health lines first, in tier-2 and tier-3 cities, using WhatsApp all day. Secondary customer: small brokerages with 3 to 20 agents (team plan, higher price). Alternatives agents use today: Excel, diaries, insurer portals and free tools. Named competitors (full detail in the competitor analysis): Agenex (about ₹999 to ₹3,499 a year, AI policy scanner and WhatsApp reminders in 12 languages), Droou ($67 to $267 a year plus a free tier), LIC's free Super Sales Saathi (April 2026), and large platforms such as PBPartners, Turtlemint and InsuranceDekho with RenewBuy, which recruit agents to sell their own products. Our edge: the agent's own workflow. Upload a bulk PDF from each insurer, get one dashboard section per insurer, a daily or monthly WhatsApp digest, and approval-first customer drafts in Indian languages. The honest risk: Agenex already does PDF scanning and WhatsApp reminders at lower yearly prices than our Pro and Business plans, so we win only by being better for agents with several insurers and hundreds of customers, and by founder-led support. 6. Go-to-market Following the four-step formula: 1. Industry: insurance distribution is dusty and fragmented. 2. Customer closeness: 20 agent interviews in the first 10 days, then 50 more. Shadow 5 agents for a day each. 3. Targeted content: 3 posts a week on LinkedIn, Instagram and Telegram/WhatsApp agent groups. Examples: "7 reasons customers don't renew", renewal-calendar templates, monthly IRDAI rule-change explainers. 4. Warm outbound: message everyone who engages with a post. Offer a free "renewal audit": upload 20 policies, get back a due-date sheet and drafted messages. Pricing: 14-day free trial for up to 100 customers, no card. Then Starter ₹299 a month (150 customers, 2 insurer files, 1 user), Pro ₹499 (500, 4, 1), Business ₹1,299 (1,500, 8, 3 users). Yearly is 10 months' price: ₹2,999, ₹4,999, ₹12,999. GST of 18% is added. Founding offer for the first agent: Business at ₹12,999 a year, price locked for 2 years, half (₹6,500) paid at the start. He named ₹15,000 a year, so do not discount below list. Channels to avoid early: paid ads, cold WhatsApp blasts to purchased lists (spam and legal risk). 7. Unit economics (per agent per month) Assumptions: 300 customers per agent, 100 WhatsApp messages a month, Claude Haiku 4.5 reading PDFs and drafting messages ($1 and $5 per million tokens), ₹96.67 to the dollar, hosting at ₹27 per agent at 200 agents, 18% GST on WhatsApp fees and a 2% payment gateway fee. Costs below are for a full plan. Typical use is assumed at 60% of the AI, WhatsApp and storage cost. Item (₹ per month, full plan) Starter ₹299 Pro ₹499 Business ₹1,299 AI (reading PDFs, drafting messages) 31 104 313 WhatsApp messages 9 29 86 File storage under 1 under 1 1 Share of hosting, WhatsApp fee, domain 27 27 27 Payment gateway 7 12 31 Total cost 74 172 457 Profit at a full plan 225 327 842 Margin at a full plan 75% 66% 65% Margin at typical use 81% 76% 77% WhatsApp is priced at the utility rate of ₹0.145 per message plus 18% GST. If Meta reclassifies our reminder templates as marketing (about ₹0.86 a message), WhatsApp cost rises about six times, to roughly ₹170 a month for a full Pro plan, and that plan's margin falls from 66% to about 37%. Re-check the live rate card before launch. AI prices are published list prices checked on 10 Oct 2026. Usage numbers are my assumptions until we measure 20 real policy PDFs. All figures are rounded. Fixed costs: about ₹6,000 a month in the lean stage (hosting and database ₹3,100, the Google AI plan we already hold ₹1,950, domain and small tools ₹950), founders unpaid. About ₹20,000 a month in the growth stage (adds the WhatsApp provider fee, accounting and legal, content and events). We move to the growth stage only after 40 agents. Break-even: each agent adds about ₹476 a month after AI, WhatsApp, storage and gateway costs (blended ₹579 less about ₹103 at typical use). That is 13 agents to cover ₹6,000 and 42 agents to cover ₹20,000. Lifetime value and acquisition cost: about ₹6,800 per agent at 7% monthly churn (₹9,500 at 5%, ₹5,900 at 8%). Target cost to win an agent is ₹1,500, so payback is about 3 months and lifetime value is about 4.5 times acquisition cost. This holds only if content and warm outreach work. 8. Twelve-month scenarios Scenario Month 3 Month 6 Month 9 Month 12 MRR at month 12 First profitable month (lean costs) Profit a month at month 12 Peak cash need Conservative 3 agents 12 30 60 ₹34,700 Month 7 ₹8,500 about ₹26,000 Base 5 30 80 150 ₹86,900 Month 5 ₹51,400 about ₹19,500 Optimistic 10 60 200 400 ₹2.32 lakh Month 4 ₹1.70 lakh about ₹16,500 Agent counts are planning targets, not forecasts, and revenue starts in month 3. The mix is 40% Starter, 40% Pro and 20% Business, so each agent adds about ₹476 a month. Profit at month 12 is agents × ₹476 less ₹20,000 fixed, before the ₹1,500 cost of winning each new agent. Peak cash need covers the lean stage only (₹6,000 fixed, ₹300 per new agent). Plan to find about ₹25,000. Your ₹5,000 and the first agent's ₹6,500 upfront cover less than half; a few yearly prepayments (a Pro year is ₹4,999) close the gap. Spend ₹20,000 a month on growth only when profit pays for it. 9. Failure modes and contingencies This is the error-handling section: what can go wrong, how we notice, and what we do. 9a. Product and technical Failure Early warning Built-in handling Fallback AI misreads a policy PDF (wrong renewal date, premium or name) Extraction confidence below threshold, agent edits the field Show extracted fields next to the PDF, require one-tap confirmation before any reminder is scheduled; never send on unconfirmed data Manual entry form; log every correction to improve prompts Scanned, blurry, password-protected or vernacular PDFs Extraction returns empty or low confidence OCR step, password prompt, clear "could not read" message with the reason Agent types 4 fields (customer, insurer, policy number, renewal date) Wrong or duplicated WhatsApp message sent Agent rejects drafts, customer complaints Approval-first sending, per-customer daily cap, idempotent send queue, 5-minute undo window Pause switch per agent and globally; apology template WhatsApp template rejected or reclassified as marketing Rejection email in provider dashboard, cost spike Submit templates as utility with transactional wording; monitor template status daily; alert if cost per message jumps Pre-approved backup templates; SMS or email fallback WhatsApp number flagged or banned Quality rating drops, delivery failures Send only to customers the agent already serves; honour STOP replies; ramp volume slowly Second number per agent; move to approved provider AI vendor outage or price rise API errors, latency, bill growth Retry with backoff, queue jobs, cache results, budget alerts Secondary model provider; reminders still run from stored dates without AI Data loss or breach Failed backup checks, access anomalies Encrypted storage, daily backups with restore test monthly, role-based access, audit log Incident runbook: contain, notify users and regulator as the law requires Insurer formats change Extraction accuracy drops for one insurer Per-insurer accuracy dashboard, regression set of sample policies Insurer-specific prompt fixes within a week 9b. Business Failure Early warning Response Agents say renewals are not their top pain Fewer than 8 of 20 name the same problem Stop. Re-interview for the next-ranked pain (claims chasing, quotes, cross-sell) or switch to the backup ideas (hospital claims, logistics reconciliation). Interested but unwilling to pay Pilots finish with under 30% converting Test per-renewal pricing or free-tier plus paid add-ons; talk to the non-payers within 48 hours High churn (above 8% a month) Agents stop uploading after week 3 Weekly "renewals saved" summary, onboarding call for every new agent, annual prepay Insurers or platforms build the same feature Competitor launches in the agent's own app Stay multi-insurer and local-language; move up to brokerage teams Acquisition too slow Under 10 new pilots a month after month 3 Shift effort to agent communities, brokerage owners and referral credit (one free month per referral) Founder conflict or burnout Missed weekly goals Written roles, weekly review, vesting schedule from day one 9c. Legal and regulatory (confirm with a lawyer before launch) • Personal data: We hold customers' names, phone numbers and policy details on the agent's behalf. India's Digital Personal Data Protection Act, 2023 and its Rules apply. Plan for consent records, a clear privacy notice, a data-processing agreement with each agent, deletion on request, and storing data in India. • Insurance rules: IRDAI regulates how agents solicit and what they say. Product messages must come from the licensed agent and stay factual. We provide drafts and comparisons, and the agent is responsible for what they send. No product advice or ranking of insurers by commission. • WhatsApp policy: Only message people who opted in or are existing customers, include an opt-out, and keep templates accurate. • Terms of service: the agent is the data controller, we are the processor; limit liability for AI extraction errors with the confirm-before-send design. • First 5 agents on a consumer AI plan: the Gemini app's terms differ from the paid API. Turn activity off, remove phone numbers before upload, get written consent, and move to the paid API from the sixth agent. Never use a free API key with customer data. • "Insurance" in the name: a company name that suggests an insurer may draw an objection from IRDAI or the company registrar. Use it as a brand name only and confirm with a CA before registering a company. 10. Ninety-day plan and kill criteria Weeks Goal Pass mark 1 to 2 20 agent interviews, 5 shadowing days 8 or more name the same pain 3 to 4 Concierge pilot with 5 to 10 agents (manual plus AI) 5 agents use it weekly 5 to 8 Build v1 (PDF upload, renewal list, WhatsApp drafts in 3 languages) 15 agents onboarded in one evening each 9 to 12 Start charging, 3 posts a week, warm outreach to everyone who engages 10 paying agents, under 8% monthly churn, at least 2 renewals saved per agent a month Kill or pivot rule: if by week 12 fewer than 5 agents pay and fewer than 2 renewals per agent are saved, we stop building and go back to interviews or switch idea. Review each Friday. 11. Immediate next actions 1. Register the domain (renewaldesk.in is free; confirm the final name first) and open a WhatsApp Business number. 2. Ask the first agent for the column headings of each insurer PDF (is the phone number included?) and his real customer count: about 300 in total, or 300 to 400 per insurer. 3. Offer him the founding deal: Business plan, ₹12,999 a year, half paid now. 4. Find 4 more agents for the concierge pilot. For each: written consent, phone numbers removed, one real PDF tested first. 5. Send the 5-question interview message to 40 agents and POSPs. 6. Talk to a lawyer or CA about DPDP, IRDAI and the "insurance" name before taking real customer data. Sources • IRDAI Annual Report 2024-25 figures via Outlook Money: https://www.outlookmoney.com/insurance/irdai-report-202425-growth-in-savings-health-and-digital-adoption • POSP count (IRDAI FY2020-21) via Cyril Amarchand Mangaldas: https://corporate.cyrilamarchandblogs.com/2022/09/point-of-sales-persons-an-alternate-distribution-channel-for-insurers/ • WhatsApp India per-message rates and service-message change: https://news.aisensy.com/news/123053-whatsapp-per-message-pricing-update and https://m.aisensy.com/blog/whatsapp-pricing-update-october-2026/ • Marketing rate and 18% GST: https://montymobile.com/blogs/whatsapp-business-api-pricing-in-india-inr-rates-gst-and-the-2026-currency-migration-deadline • Competitor names, prices and partner counts: Competitor analysis, with its own source list. • AI model prices ($ per million tokens), checked 10 Oct 2026, and our usage assumptions: the final pricing workbook (Inputs and Cost and price sheets).

Make My PRD

Design by The Resonance | Powered by GPC – The AI Transformation Company

    PRD: RenewalDesk: Business Plan