Auto, education, and SME loans are structurally different products with different income verification requirements, collateral frameworks, and regulatory overlays — but they share a common qualification challenge: the gap between a customer expressing interest and a complete, credit-ready application reaching the underwriting team. Kallix closes this gap through a structured voice qualification flow tailored to each loan type.

For vehicle loans: the agent qualifies based on vehicle type (new car, used car, two-wheeler, commercial vehicle, EV), LTV eligibility (85–90% for new cars, 75–80% for used vehicles), borrower income (salaried via Form 16/salary slip or self-employed via ITR), CIBIL score threshold (typically 700+ for new vehicles, 650+ for used), and vehicle age for used vehicles (maximum 10–15 years from manufacture depending on lender policy).

For education loans: the agent qualifies the student (age, admission status, institution type), screens the co-applicant (parent/guardian income for loans above Rs 4 lakh without collateral), checks whether the loan amount requires collateral (above Rs 10 lakh typically requires property, FD, or insurance), and screens for CSIS (Central Sector Interest Subsidy Scheme) eligibility for economically weaker section students.

For SME loans: the agent verifies Udyam registration (mandatory since 2020), collects business vintage (minimum 2–3 years for most products), screens for CGTMSE eligibility (collateral-free up to Rs 5 crore for eligible MSMEs), and initiates Account Aggregator consent for bank statement analysis — replacing the 3-month physical bank statement submission process with a 60-second digital pull.

Across all three categories, the agent delivers the KFS (Key Fact Statement) as required by RBI Digital Lending Guidelines 2022 and routes qualified applicants to the LOS (Loan Origination System) with pre-filled application data.

Vehicle loan qualification via voice primarily serves two audiences: customers inquiring on dealer showroom floors who want an instant in-principle eligibility check, and inbound leads from auto OEM or bank digital campaigns who need pre-qualification before visiting a dealership. Kallix handles both with a 6–8 minute structured qualification call.

For new car loans: the agent collects the vehicle model (or on-road price range), the customer's net monthly income (salaried or self-employed), existing EMI obligations (for FOIR calculation), and residence type (owned/rented — affects risk scoring at some lenders). It then calls the bureau API for a soft pull CIBIL score, calculates the eligible loan amount at the lender's LTV policy (typically 85% on on-road price for new cars), computes the EMI, and assesses FOIR including the proposed EMI. If FOIR is within the lender's cap (typically 40–55% for salaried), the customer is pre-qualified.

Manufacturer subvention schemes (zero-percent EMI or reduced rate offers from Maruti, Hyundai, Tata, Honda, Mahindra, Kia) are checked in real time against the lender's subvention tie-up list — if the customer's chosen vehicle qualifies, the agent advises on the subvention rate and tenor.

For used cars: the agent additionally collects the vehicle registration number, which it cross-references with the lender's vehicle valuation database (tied to platforms like OBV, CarDekho, Cars24 valuator APIs) to estimate the current market value and determine the eligible loan amount. The vehicle's manufacture year is checked against the lender's maximum age policy (10 years from manufacture for most banks, 15 years for some NBFCs). RC transfer completion is flagged as a disbursement condition — the agent sets this expectation upfront.

Two-wheeler loans are among India's highest-volume retail lending products — over 2 crore two-wheelers sold annually, with 60–70% financed. Qualification is simpler than car loans but requires different income thresholds given the smaller ticket size.

For two-wheelers: the agent collects the vehicle model and on-road price, income source (salaried, self-employed, or agricultural — some lenders serve farmers for two-wheelers without ITR), and checks CIBIL with a 650+ threshold (some NBFCs like Bajaj Finance, TVS Credit, Shriram Finance go to 600 or no-score for first-time borrowers with good income). LTV is 85–90% on the on-road price. For NTC (New-To-Credit) two-wheeler applicants without a CIBIL score, the agent checks for alternative data — Aadhaar-linked income records, utility bill payment history, or a salaried account salary credit pattern — and routes to the lender's NTC underwriting desk.

For EV loans: the qualification adds FAME II subsidy screening. FAME II (Faster Adoption and Manufacturing of Hybrid and EV) Phase II provides purchase subsidies for electric two-wheelers (up to Rs 15,000 per vehicle), three-wheelers, and four-wheelers below Rs 15 lakh (four-wheelers; this threshold was revised periodically). The agent confirms: (1) whether the specific vehicle model is on the FAME II eligible model list (maintained by HEAVEY Industries); (2) whether the dealer is a certified FAME II beneficiary — the subsidy is applied at the point of sale by the dealer, not post-purchase.

State-level EV subsidies (Delhi Rs 5,000–30,000 per two-wheeler, Maharashtra Rs 10,000, Gujarat Rs 20,000, Rajasthan Rs 10,000) are identified based on the customer's registration state and added to the effective subsidy amount in the eligibility summary. The final loan amount is on-road price minus FAME II subsidy minus state subsidy.

Commercial vehicle (CV) loans are structurally different from car loans — the borrower's income is often derived from the vehicle itself (a truck owner-operator running on a specific route), and traditional income documents (ITR, salary slip) may not capture the actual cash flow. Kallix's CV qualification flow is adapted for this segment.

For owner-operators (the most common CV loan borrower): the agent collects the number of existing vehicles owned (fleet size), the type of load carried (goods, passengers), the route operated (national highway or state road — affects permit type and income stability), and the monthly net income as declared. Bank statement analysis via Account Aggregator provides 12 months of cash flow data — the agent identifies regular freight payment credits, operating expense debits, and calculates a sustainable net income figure for FOIR purposes.

For first-time operators (buying their first commercial vehicle): the income assessment is more challenging — no existing vehicle cash flow to reference. The agent collects the employment/business history (e.g., driver working for a fleet owner, now buying own vehicle), existing assets (land, residential property, savings), and any letter of intent from freight aggregators (Mahindra Logistics, Rivigo, TCI, Delhivery) confirming contracted load. This documentation set is routed to the CV underwriting desk for manual review.

Key documents for CV loans: vehicle fitness certificate (issued by RTO, valid 2 years for new vehicles and 1 year thereafter), route permit (State Carriage or National Permit for inter-state operations), commercial driving licence (HMV licence for HCVs), and motor insurance. The agent confirms all are available or in-process before flagging the application as complete-ready.

For fleet operators financing additional vehicles: the agent checks the existing loan-to-fleet ratio (most lenders cap at 3:1 — Rs 3 outstanding per Rs 1 repaid per month), NPA status on existing CV loans, and the fleet's NACH debit record.

Education loans have a unique qualification structure — the primary borrower (student) has no income at application time, the co-applicant (parent or guardian) bears the income qualification, and repayment is deferred for the duration of the course plus 12 months (or 6 months after employment, whichever is earlier). Kallix's education loan qualification flow handles all three dimensions.

Institution check: the agent confirms whether the institution is on the IBA-approved list (the Indian Banks' Association Model Education Loan Scheme covers over 3,000 institutions). For overseas education, the agent checks whether the destination country and institution are eligible — most banks cover USA, UK, Canada, Australia, Singapore, New Zealand, Germany, and Ireland; some extend to Japan, France, and other OECD countries. Institution-level NAAC accreditation grade (A or A+ preferred) and NIRF ranking are additional signals at some lenders.

Course check: the agent confirms the course type — engineering (B.Tech, M.Tech), medicine (MBBS, MD), management (MBA), law (LLB, LLM), or other professional courses. Pure arts or humanities courses at non-premier institutions may not qualify at all banks. For vocational courses and skill development programmes, the agent routes to the specific lender offering PMKVY-linked education loans.

Loan amount and collateral matrix: under the IBA Model Scheme, loans up to Rs 4 lakh require no security; loans Rs 4–10 lakh require a co-applicant guarantee (no collateral); loans above Rs 10 lakh require tangible collateral (residential property, FD, LIC policy with surrender value, or CGFSEL coverage). The agent screens which band the requested loan falls into and advises accordingly.

For overseas education: the agent additionally checks FOREX disbursement requirements (fees are paid in foreign currency directly to the institution), and screens for the ED-CG scheme (overseas education loan guarantee for SC/ST students under NSFDC).

The co-applicant in an education loan is the primary income qualifier — the student typically has no income, making the co-applicant's credit profile the critical risk variable. Kallix's education loan qualification flow therefore spends the majority of its assessment time on the co-applicant.

Co-applicant income assessment: for salaried co-applicants, the agent collects net monthly take-home salary, tenure with current employer, and outstanding EMI obligations. Account Aggregator consent is requested to pull 6 months of bank statement — the salary credit pattern confirms the stated income and identifies any hidden EMI debits not captured by the co-applicant's verbal declaration. FOIR is calculated including the projected education loan EMI post-moratorium — if FOIR exceeds 50% (the typical bank cap for education loans), the agent advises on a higher down-payment or co-borrower addition.

For self-employed co-applicants (business owners, professionals): the agent collects 2-year ITR data (average of net taxable income across 2 years, with a 70–80% haircut at most lenders), profession type, and business vintage. Account Aggregator pulls GST GSTR-3B data and business bank statement simultaneously.

Collateral assessment for loans above Rs 10 lakh: the agent confirms the type of collateral being offered — residential property (LTV 75%, value estimate from CBS or recent valuation report), FD (100% of FD value as collateral), or LIC policy (surrender value minus margin). For property collateral, the agent confirms the property is free of existing mortgage (CERSAI check), the title is clear (legal heir chain confirmed verbally), and the property is in the name of the co-applicant or an immediate family member.

The moratorium period (EMI-free during course + 12 months) is confirmed to the co-applicant — simple interest accrues during this period, and the agent advises whether the bank's policy is to add it to the principal or allow it to be paid separately. Pre-EMI interest payment during moratorium improves the co-applicant's debt position post-moratorium.

CSIS is administered by the Ministry of Education through Canara Bank (the nodal bank) and is available to all Indian students who take an education loan from any scheduled commercial bank under the IBA Model Scheme. The subsidy covers 100% of the interest accruing during the moratorium period — for an Rs 8 lakh loan at 9% per annum over a 4-year course, this can amount to Rs 2.88 lakh in interest subsidy.

Eligibility screening: the AI asks two qualification questions: (1) Is the applicant's total parental/family income from all sources below Rs 4.5 lakh per annum? (2) Does the student hold an income certificate from a competent authority (SDM, Tehsildar, Revenue Officer, or BDO) confirming income below Rs 4.5 lakh? The income certificate is mandatory for claiming the subsidy — the agent explains this document requirement and, if the certificate is not yet obtained, guides the student to the correct government office.

For technical courses at premier institutions (IITs, IIMs, NITs, central universities), CSIS is available regardless of institution type as long as the income criterion is met. For private institutions, the institute must be NAAC-accredited or on the AICTE/UGC/MCI/BCI approved list.

Additional interest subsidy schemes: Dr. Ambedkar Central Sector Scheme (for OBC and Economically Backward Classes students pursuing overseas education); Padho Pardesh Scheme (for minority students pursuing overseas education — this scheme was in limbo post-2022; the agent advises the student to confirm current status with the Ministry of Minority Affairs). Some states run parallel interest subsidy schemes (Maharashtra Annasaheb Patil Corporation scheme for OBC students; UP Dr. Ambedkar Scholarship) — the agent identifies the student's state and advises on applicable state-level subsidies.

SME and MSME loan qualification is one of the most complex pre-screening challenges in retail lending — these borrowers often lack formal financial statements, have irregular cash flows, and may not have audited books below a certain turnover threshold. Kallix's SME qualification flow addresses this through Account Aggregator-based data extraction rather than physical document collection.

Udyam registration verification: since July 2020, Udyam registration (at udyamregistration.gov.in) replaced the older Udyog Aadhaar registration and is mandatory for accessing MSME benefits including CGTMSE guarantee. The agent collects the Udyam Registration Number (URN) and confirms the enterprise category (micro, small, or medium) — this determines CGTMSE fee structure and eligible loan amount.

Business financial assessment: the agent requests Account Aggregator consent to pull 12 months of the business current account bank statement and GST GSTR-3B returns. From the bank statement, the AI calculates: average monthly turnover (gross credits), average cash balance, NACH/cheque return frequency (bounce rate — lenders typically require under 5%), and major outflow patterns. From GSTR-3B, it extracts monthly output tax liability as a proxy for revenue (since GSTR-3B is harder to manipulate than self-declared turnover).

For borrowers below GST threshold (annual turnover below Rs 40 lakh for goods, Rs 20 lakh for services): GSTR data is unavailable. The agent relies on bank statement analysis and ITR (Schedule BP for business income) for income verification.

Property-backed SME loans: for loans requiring mortgage collateral, the agent collects property type, estimated value, existing CERSAI registration status, and municipal khata/property tax receipt status — routing to the bank's property valuation team for formal assessment.

For working capital: the agent assesses the borrower's trade cycle — 'How many days of credit do you extend to your customers? How many days credit do you receive from suppliers?' — to calibrate the working capital requirement and determine whether a CC (cash credit) or overdraft structure is appropriate.

CGTMSE was established by the Government of India and SIDBI in 2000 to provide credit guarantee coverage to banks and NBFCs extending collateral-free loans to MSMEs. The guarantee covers 75–85% of the loan amount (85% for micro enterprises and women-owned businesses) in case of default, enabling lenders to extend credit without requiring the borrower to mortgage property.

Screening criteria: the AI confirms (1) the enterprise is a micro or small enterprise per the new MSME definition (investment up to Rs 10 crore, turnover up to Rs 50 crore for small; up to Rs 1 crore investment and Rs 5 crore turnover for micro); (2) the enterprise is in manufacturing or services (trading enterprises are typically excluded or covered at a lower guarantee percentage); (3) the loan purpose is legitimate business use — working capital, term loan for equipment, or business expansion; and (4) the borrower has not been previously classified as NPA with any CGTMSE-member institution (searchable via CIBIL MSME rank).

Lender CGTMSE membership: the guarantee is only available if the lending bank is a CGTMSE member institution (all major PSBs, leading private banks, and many NBFCs are members). The agent confirms the lender's membership before screening the borrower — routing to an alternative lender if the primary lender is not a member.

CGTMSE annual guarantee fee: the borrower pays a guarantee fee (currently 1.5% per annum for loans above Rs 1 crore, 1% for below Rs 1 crore — paid by the lender and typically passed to the borrower) in addition to the loan interest rate. The agent factors this into the effective cost of credit disclosed in the KFS.

For hybrid collateral situations (borrower has some collateral but not sufficient for the full loan amount): some lenders use partial CGTMSE coverage — guaranteeing the uncollateralised portion while taking available collateral for the rest.

Pradhan Mantri Mudra Yojana (PMMY), launched in 2015, provides refinancing support to MLIs (Member Lending Institutions) that extend loans to micro-enterprises and non-farm income-generating activities in the informal sector. Over Rs 23 lakh crore has been disbursed under Mudra since inception (as of FY24), making it one of the largest MSME credit programmes in the world.

Shishu tier (up to Rs 50,000): primarily for first-time borrowers in micro-enterprises — street vendors, small manufacturers, artisans, and petty traders. No collateral required. Income verification relies on cash flow declarations and Aadhaar-linked business identity. The AI qualifies Shishu applicants in under 5 minutes: confirming the business activity, GST or trade licence (if any), and bank account details. For the Pradhan Mantri Street Vendor's Atmanirbhar Nidhi (PM SVANidhi) sub-scheme (up to Rs 10,000 for street vendors), the agent additionally confirms the vendor's certificate issued by the Urban Local Body.

Kishore tier (Rs 50,001 to Rs 5 lakh): requires 6 months of bank statement, Udyam registration, and ITR for the last 1–2 years. The agent pulls bank statement via Account Aggregator and evaluates the average monthly turnover — a business with Rs 2 lakh/month turnover typically qualifies for a Rs 2–3 lakh Kishore loan. CIBIL check is required at most MLIs.

Tarun tier (Rs 5 lakh to Rs 10 lakh): the most rigorous Mudra qualification — requiring 2-year ITR, 12-month bank statement, balance sheet if available, and a business plan or project report for the additional loan amount. The agent qualifies the basic financial eligibility and routes to the credit manager for project report review, since the business plan assessment cannot be fully automated.

For women borrowers: Mudra loans carry a 25 bps interest rate concession at most PSBs (priority under PM Mudra Mahila scheme). The agent identifies women applicants and flags the concession in the qualification summary.

Credit bureau integration is the anchor of pre-qualification — without a live bureau score, any income-based qualification is incomplete and the lender's risk team will re-run it anyway. Kallix integrates bureau checks within the call, making the qualification outcome meaningful rather than a provisional estimate.

Bureau API call: the agent collects PAN (the universal credit identity in India) and date of birth, sends a soft inquiry to the primary bureau (CIBIL for most banks; Experian and CRIF High Mark as secondary bureaus at some lenders). The soft pull returns the score in 8–12 seconds without creating a hard inquiry on the borrower's credit file — important for applicants comparing lenders, since multiple hard inquiries within 30 days lower the CIBIL score.

For vehicle loans: CIBIL threshold is typically 700+ for new vehicles and 650+ for used. Below 650 routes to the NBFC channel (Bajaj Finance, Mahindra Finance, Shriram Finance) where sub-prime vehicle lending is more developed. The bureau pull also returns existing active loan count and any 90+ DPD (days past due) flags — a single 90+ DPD in the last 24 months typically results in decline at PSBs.

For education loans: the co-applicant's CIBIL score (650+) is checked, and existing loan obligations are used for FOIR. The student's bureau file is also pulled — a thin file (no history) is acceptable; an active negative file (defaults or settlements) may affect the application at some lenders.

For SME loans: the agent pulls both the proprietor's personal CIBIL (for sole proprietorships) and the business CIBIL MSME Rank (a separate score for businesses based on their credit history, tax compliance, and banking behaviour). The CIBIL MSME Rank range is 1–10; Rank 1–3 represents the lowest risk. Lenders typically require Rank 4 or better for unsecured SME loans.

FOIR calculation: the agent totals all existing EMIs from the bureau report, adds the proposed loan EMI, and divides by net monthly income. Results above the lender's FOIR cap trigger a loan amount reduction suggestion — the agent recalculates the maximum eligible amount at the FOIR ceiling.

PSB Loans in 59 Minutes is SIDBI's fintech-powered platform that provides SME borrowers with an in-principle approval from participating public sector banks (SBI, Bank of Baroda, Punjab National Bank, Bank of India, Union Bank, HDFC, Axis, and others) within 59 minutes of application submission, based on automated analysis of GST returns, IT returns, and bank statements.

Kallix's integration with the PSB Loans in 59 Minutes platform: when an SME applicant has provided Account Aggregator consent for GST and bank statement data during the qualification call, Kallix can submit the pre-formatted data to the PSB Loans API, receive the in-principle approval amount and the list of partner banks willing to lend, and communicate the result to the borrower within the call — without requiring the borrower to visit the portal separately.

The portal's algorithm evaluates: GSTR-3B monthly turnover for the last 12 months, ITR income for the last 2 years, bank account statement for 6 months (average balance, credit/debit ratio, NACH returns), CIBIL score for the borrower and the business, and Udyam registration details. Loan amounts from Rs 1 lakh to Rs 5 crore are processed; above Rs 5 crore requires direct bank application.

For GST-registered businesses with turnover above Rs 40 lakh: the PSB Loans portal offers the most streamlined path — in-principle approval in 59 minutes, followed by formal application and physical verification, with sanction typically in 7–10 business days. The AI positions this as the recommended path for eligible SME borrowers, highlighting the speed advantage versus traditional bank application (4–8 weeks in many PSBs).

For Mudra and CGTMSE loans via the portal: the PSB Loans in 59 Minutes platform also handles Mudra Kishore and Tarun loans, and flags CGTMSE guarantee eligibility in the in-principle approval — making it a one-stop screen for government scheme-linked SME lending.

RBI's Digital Lending Guidelines (September 2022) fundamentally changed the compliance requirements for AI and digital channels used in loan origination. Kallix's compliance architecture is built around these guidelines as mandatory defaults.

LSP disclosure: the agent opens every loan qualification call with: 'I'm calling on behalf of [Lender Name]. This call is powered by Kallix, a Lending Service Provider.' This disclosure satisfies RBI's requirement that the LSP identity not be concealed from the borrower.

KFS delivery: the Key Fact Statement must include 8 mandatory data points — Annual Percentage Rate (APR), loan amount, tenor, EMI, total interest payable, all fees and charges (processing fee, insurance, GST on fees), penal interest rate (maximum 2% per month above the contracted rate per RBI's October 2023 circular), and the grievance redressal officer contact. Kallix generates the KFS from the lender's product configuration and dispatches it to the borrower's registered WhatsApp and email within 60 seconds of qualification completion — before the borrower submits the formal application.

3-day cooling-off period: for first-time borrowers accessing a digital loan product for the first time with a given lender, RBI mandates a 3-day window during which the borrower can exit the loan without any penalty. The agent confirms the cooling-off right verbally and captures acknowledgement, which is logged in the LOS.

No automatic enhancement: the agent cannot recommend a loan amount higher than what the borrower's eligibility supports — the guideline prohibits LSPs from nudging borrowers to accept higher-than-required loan amounts. Kallix's script enforces the eligible maximum as the upper bound, with the borrower free to request less.

For NBFC-originated products: FLDG (First Loss Default Guarantee) arrangements between the NBFC and the LSP are capped at 5% of the loan portfolio per RBI's 2022 guidelines — this affects Kallix's contractual structure with NBFC partners but is transparent to the borrower.

A below-threshold CIBIL score doesn't necessarily mean a definitive decline — it means the primary product on the primary lender is likely unavailable, and the agent's role is to identify the best alternative path rather than simply ending the qualification call.

Explaining the bureau result: the CIBIL report includes reason codes that explain why the score is at its current level. The agent translates these into plain language: 'Your score is at 620, primarily because of one missed payment on a personal loan in September 2023 and a credit card outstanding that's at 82% of your limit. Both these can be addressed within 3–6 months.'

Score improvement advice: (1) regularise any overdue EMIs — a single settled overdue EMI can improve the score by 40–60 points within 2 credit cycles (approximately 60 days); (2) reduce credit card utilisation below 30% — high utilisation is the second-most-common score drag; (3) avoid new credit applications for 90 days — hard inquiries lower the score marginally; (4) dispute any erroneous entries directly with the credit bureau or the reporting lender.

Alternative lender routing: for vehicle loans where the bank requires 700+ and the customer is at 650, the agent identifies NBFC partners (Bajaj Finance, Mahindra Finance, Shriram Finance, HDB Financial Services, IIFL Finance) that operate at 600+ thresholds and routes the application accordingly. For SME loans, the agent explores the CGTMSE collateral-free route, where some lenders apply modified risk assessment rather than pure CIBIL scoring.

Secured alternative: if the customer has an FD, gold, or property as collateral, the agent offers a secured loan option — secured loans typically have no minimum CIBIL requirement since the collateral mitigates the credit risk.

Document collection is the step where most loan applications stall — customers receive a generic checklist, don't know which version of a document is needed, upload wrong formats, or simply delay. Kallix addresses this with a loan-type-specific, context-aware checklist delivered immediately after qualification, with guided WhatsApp upload and real-time OCR validation.

For vehicle loans (new car): the checklist includes income proof (last 3 salary slips + latest Form 16 for salaried, or last 2-year ITR for self-employed), identity and address proof (Aadhaar, PAN), bank statement for 6 months (or AA pull consent), and vehicle proforma invoice from the dealer. For used cars: add RC copy, existing insurance policy, and the seller's NOC from the financier (if the vehicle is under an active loan).

For commercial vehicles: additional documents — commercial driving licence (HMV or LMV Transport), vehicle fitness certificate (Form 38 from RTO, valid for 2 years on new vehicles and 1 year thereafter), route permit (National or State Carriage), and motor insurance certificate with commercial purpose endorsement.

For education loans: the loan-specific documents are the offer/admission letter from the institution, the fee structure issued by the institution (broken by academic year), the co-applicant's last 3 salary slips or 2-year ITR, and the institution's NAAC/AICTE/UGC accreditation certificate. For overseas education: additionally, the I-20 (USA), CAS (UK), COE (Australia), or equivalent admission confirmation, and FOREX fee payment schedule.

For SME loans: the checklist is Udyam registration certificate, last 2-year ITR with computation, 12-month bank statement (or AA consent), last 12 months GSTR-3B (or AA consent to pull from GSTN), business constitution documents (partnership deed, MOA/AOA for companies), and — for property-backed loans — property documents (title deed, khata, latest property tax receipt, encumbrance certificate for the last 15 years).

OCR validation: WhatsApp-uploaded documents are processed by Kallix's OCR engine — the PAN number is extracted and matched against the CBS application; the ITR income figure is extracted and compared against the declared income; bank statement header details confirm the account owner. Mismatches trigger a re-upload request with specific guidance on what was wrong.

Loan qualification is where the biggest inefficiency in retail and SME lending lies — most leads generate 5–10 phone tag cycles between the customer and the DSA or branch officer before a complete application is submitted. Kallix's AI qualification call compresses this to a single structured conversation.

Conversion benchmarks by loan type: new car loans show the highest qualification rate (70–78%) because the customer's intent is well-formed and the qualification criteria are relatively simple. Used car loans achieve 62–70% due to the added complexity of vehicle age and RC transfer conditions. Education loans achieve 58–66% — lower due to the co-applicant availability constraint (the parent needs to be available during the same call). SME loans achieve the lowest first-call qualification rate (55–65%) because financial data pull via Account Aggregator often requires multiple SMS OTP verifications that some SME borrowers abandon mid-call.

Cost per qualified applicant: Kallix produces a qualified, document-submitted application at Rs 220–380 per applicant (including telephony, API costs, and platform fee). DSA-originated applications cost Rs 1,200–2,000 for the same qualification quality (DSA commission is the dominant cost). Branch-originated applications cost Rs 800–1,400 when bank staff time is factored in.

Time-to-complete: a Kallix-qualified applicant's application moves from lead to ready-for-underwriting in under 24 hours (qualification call + document upload via WhatsApp). The branch origination equivalent takes 3–5 days (multiple visits, physical document submission, data entry). This time reduction translates to higher applicant retention — applications that stall more than 3 days have a 35–40% higher drop-off rate.

For SME specifically: the Account Aggregator integration reduces the physical bank statement submission cycle from 5–7 days to under 3 minutes, which is the single largest origination timeline reducer in the SME segment.

RBI's Priority Sector Lending (PSL) guidelines require scheduled commercial banks to direct 40% of Adjusted Net Bank Credit (ANBC) to priority sectors — agriculture (18%), MSME (7.5% sub-target for micro enterprises), export credit, education (up to Rs 20 lakh per borrower), housing, renewable energy, and weaker sections. Banks that fall short of PSL targets must purchase PSLC (Priority Sector Lending Certificates) from surplus banks or contribute to RIDF (Rural Infrastructure Development Fund) — making PSL classification a financially significant attribute of each loan application.

For SME loans: the MSME classification (micro, small, or medium) and the Udyam registration category determine PSL eligibility. Micro enterprises (investment up to Rs 1 crore, turnover up to Rs 5 crore) qualify for the priority micro-enterprise sub-target — lenders track this separately from the general MSME target. The agent flags the Udyam category in the LOS application metadata.

For education loans: loans up to Rs 20 lakh to Indian students for studying in India or abroad qualify as PSL under the 'Education' category. Loans above Rs 20 lakh don't qualify for PSL. The agent confirms the loan amount at qualification and flags PSL eligibility in the LOS.

For agriculture equipment financing (tractors, power tillers, irrigation equipment): if a CV loan is for a tractor or agricultural machinery, it may qualify under the 'Agriculture — Allied Activities' PSL category, which is part of the 18% agriculture target. The agent identifies the vehicle type and flags accordingly.

PSL certificate purchases (PSLCs): for banks that originate excess PSL loans (particularly small finance banks and microfinance institutions specialising in MSME and agriculture), PSLCs are tradeable instruments allowing them to earn fee income from surplus PSL. Kallix's loan origination data, structured for PSL metadata, feeds directly into the bank's PSLC reporting system.

The goal of AI pre-qualification is to make the human loan officer's job easier, not to replace their credit judgement. Kallix routes qualified applications to the LOS with all data pre-filled, so the officer reviews rather than re-collects.

LOS integration: Kallix supports pre-built API connectors for major LOS platforms — Newgen Loan Origination, Nucleus FinnOne, Intellect Design Arena, Salesforce Financial Services Cloud, and LeadSquared. Post-qualification, the agent submits a structured application payload: applicant identity, income and FOIR data, bureau score and key derogatory flags, loan amount, LTV, product type, PSL flag, and document submission status. The LOS assigns a case ID which is sent to the applicant's WhatsApp for tracking.

For straightforward cases (salaried applicant, CIBIL above threshold, vehicle loan within standard LTV): the LOS workflow takes the case directly to the underwriting queue — no loan officer call required until the physical document verification stage.

For complex cases requiring human judgement: (1) CIBIL below threshold (loan officer reviews case holistically); (2) irregular income (self-employed with significant cash income, seasonal income patterns); (3) property-backed SME loan (property valuation, title verification, and CERSAI search require human officer involvement); (4) first-time commercial vehicle operator (income proxy documents need judgment). For these cases, the agent schedules a callback: 'A loan specialist will call you within 24 hours. Your pre-qualification reference is [case ID].' The case is assigned to the officer in the LOS with the AI qualification summary attached.

DSA routing: if the bank uses an external DSA network for specific loan types (common for used car and two-wheeler loans), the agent routes qualified applicants to the nearest DSA partner using a geo-based allocation from the bank's DSA management system — sending the DSA the applicant profile and contact details for follow-up.

Used car loans require a vehicle assessment layer on top of the standard borrower credit assessment — both the buyer and the asset must qualify. Kallix's used car qualification call covers the vehicle check in parallel with the borrower FOIR and CIBIL review.

Vahan API integration provides real-time RC data: registered owner name (confirms the seller is the registered owner), hypothecation status (if the vehicle is hypothecated to a bank, the existing loan must be cleared before transfer), fitness certificate validity, and pollution under control (PUC) certificate status. A vehicle with an active hypothecation triggers an immediate flag: 'This vehicle has an existing loan from [Bank Name]. The seller must obtain an NOC before we can process your loan.'

Market value calculation: the AI uses make/model/year/mileage inputs to calculate current market value from the lender's internal valuation model (or third-party tools like CarWale/Spinny valuation API where integrated). LTV is calculated against this market value: a vehicle worth Rs 4 lakh can support a loan of Rs 2.8–3.2 lakh at 70–80% LTV. If the applicant's requested loan amount exceeds the LTV ceiling, the AI advises on the gap the applicant must fund from own sources.

Vehicle age and odometer limits vary by lender: some banks (Kotak, HDFC) cap at vehicles under 10 years old from manufacture date; others allow up to 12 years. Mileage caps of 80,000–1,00,000 km are standard. The AI captures the odometer reading from the RC inspection report or self-declaration and confirms eligibility before proceeding with the full qualification.

Tractor loans are fundamentally different from urban vehicle loans: the borrower is a farmer whose income is seasonal (kharif + rabi harvest cycles), the asset is productive (the tractor generates income), and the lender's security includes both the vehicle hypothecation and land title. Kallix AI's tractor qualification script is calibrated for this context.

Income assessment for farmers: the standard FOIR metric (fixed obligation to income ratio) does not apply directly to agricultural income because farmers have no fixed monthly income. Kallix AI uses a 12-month income reconstruction: total harvest income (two crop cycles) minus input costs (seeds, fertiliser, labour) equals net annual agricultural income. The AI asks the farmer to state last year's crop yield, the selling price per quintal, and the acreage under cultivation — cross-verified against the stated land holding.

KCC (Kisan Credit Card) credit history: many tractor loan applicants have thin CIBIL files because their borrowing history is limited to KCC credit limits from cooperative banks, which may not be reported to credit bureaus. Kallix AI flags this scenario and confirms that the lender's policy allows KCC statement substitution for CIBIL thin-file applicants.

Manufacturer tie-up pricing: tractor loan rates are significantly lower under manufacturer tie-ups (Mahindra Finance with Mahindra tractors; HDFC Bank with John Deere, etc.) — typically 0.5–1.5% lower than open-market rates. The AI checks whether the applicant's preferred tractor brand has a tie-up with the lender and presents the subsidised rate. This rate advantage drives 28–34% higher conversion on manufacturer-aligned calls.

Overseas education loans are one of the highest-value retail lending products in India — and one of the most complex to qualify. The qualification covers the student's academic profile, the institution's ranking, the borrower's family financial capacity, and the collateral assessment — all in a single AI call.

University ranking and approved institution list: most lenders (SBI, Axis, HDFC Credila) maintain a list of approved overseas institutions ranked by QS World University Rankings or THE (Times Higher Education). The AI verifies whether the applicant's target institution is on the approved list and flags if it is not — preventing the applicant from discovering eligibility issues only after formal application. Top-100 QS-ranked institutions typically qualify for the full loan amount; lower-ranked institutions may be capped at a lower loan-to-cost percentage.

Cost of attendance (CoA) validation: the AI collects the I-20 (US) or CAS (UK) or offer letter CoA figure and cross-validates against the lender's CoA database for the specific institution. Artificially inflated CoA claims are a fraud risk that lenders monitor closely. For scholarship-awarded students: the AI deducts confirmed scholarship amounts from the loan need calculation — overstating the loan need is a common applicant error that delays processing.

Collateral structure for large overseas loans: loans above Rs 40 lakh require collateral (residential property, FD, NSC/LIC). The AI collects property location, owner, market value estimate, and whether it is already mortgaged (existing mortgage deducts from available collateral value). The AI explains the CERSAI registration requirement for property collateral — the lender registers a charge on the property, which cannot be sold without the lender's NOC.

GIC (Guaranteed Investment Certificate) for Canada / blocked account for Germany: some countries require students to deposit Rs 8–12 lakh in a blocked account as part of the visa process. Kallix AI confirms whether the loan can fund this blocked account amount and the documentation required.

The traditional SME loan qualification dilemma — self-declared financials are unreliable but formal audited accounts are unavailable for many small businesses — is partially resolved by GST filing data. A business that has consistently filed GSTR-3B for 24 months with stable or growing turnover is a significantly lower credit risk than one that provides selective bank statements.

Kallix's GST-based qualification workflow: the applicant provides their GSTIN, the AI generates a consent link via the Account Aggregator framework for GSTN data access, and on receipt of consent, the AI downloads 12–24 months of GSTR-3B filings. The qualification calculation: average monthly turnover → annualised revenue → eligible loan amount (typically 20–30% of annual turnover for working capital loans; up to 50% for term loans secured by property).

GST filing gaps and red flags: irregular GSTR-3B filings (months with nil returns during stated busy periods), sudden turnover spikes in the application month, or high input tax credit relative to output tax (suggesting purchase inflation) are flagged to the loan officer with a specific observation note. These flags do not auto-reject the application but trigger enhanced documentation.

GSTR-1 vs GSTR-3B consistency: the AI cross-checks GSTR-1 (sales data) with GSTR-3B (consolidated return). Material inconsistencies between the two filings — which can occur due to data entry errors or deliberate manipulation — are flagged as discrepancies for underwriter review.

For businesses below the GST threshold (annual turnover < Rs 40 lakh — exempted from GST): the AI transitions to bank statement analysis (12 months) and Udyam registration confirmation for MSME classification. Below-threshold businesses are still eligible for MUDRA and CGTMSE-backed loans.

LAP is the primary secured borrowing instrument for SMEs who cannot access collateral-free credit at the scale they need. A Rs 50 lakh+ LAP against a self-owned commercial property provides working capital that would otherwise be unavailable as an unsecured loan.

Kallix's LAP qualification covers two tracks simultaneously: (1) Business repayment capacity — for LAP amounts below Rs 50 lakh, FOIR against business income is used; for amounts above Rs 50 lakh, DSCR (Debt Service Coverage Ratio) is calculated: Net Annual Income / Annual Debt Obligations. A DSCR of 1.25 or above is the standard threshold. The AI collects the business's gross annual income, existing EMI obligations, and target LAP EMI at the quoted rate to calculate DSCR on the call.

(2) Property track — the AI collects: property type (residential/commercial/industrial), property location (metro/tier 1/tier 2 — affects LTV eligibility), registered ownership (self-owned vs joint vs family), any existing encumbrance (existing mortgage, legal dispute), and approximate market value. The AI flags properties that are typically difficult to mortgage: properties in unauthorized colonies, agricultural land, properties with unclear title, or inherited properties without registered partition.

CERSAI encumbrance check: the AI checks whether an existing mortgage is registered on the CERSAI portal (Central Registry of Securitisation Asset Reconstruction and Security Interest). An existing CERSAI registration means the property is already mortgaged — the existing lender's NOC and partial/full discharge is required before the new LAP can be registered.

Balance transfer LAP: for applicants with an existing LAP from another lender, the AI calculates the balance transfer benefit (rate differential + processing fee net) and the top-up eligibility (if the property has appreciated and there is unutilised LTV).

Invoice discounting is the fastest-growing SME credit product in India — driven by the GST-linked digital invoice infrastructure and TReDS (Trade Receivables Discounting System) platforms mandated for MSME buyers by RBI. Kallix AI qualifies SME sellers for invoice discounting by assessing both the seller's operating profile and the quality of the specific invoices to be discounted.

Buyer credit quality is the primary determinant: an invoice from a PSU (government buyer) or a large listed corporation is almost always discountable at 80–90% of face value with minimal documentation. An invoice from a small private buyer carries higher risk and may be discounted at a lower percentage or require additional security. The AI categorises the buyer and quotes the applicable discount rate and LTV.

TReDS platform navigation: for MSMEs with PSU buyers, TReDS (M1xchange, RXIL, Invoicemart) offers the lowest-cost receivable financing in India — rates of 6–9% p.a. vs 12–18% from NBFCs. The AI confirms whether the PSU buyer is registered on TReDS, guides the MSME through the platform onboarding process, and connects them to a registered factor/financier on the platform.

GST invoice verification: the AI cross-checks whether the invoices to be discounted are reflected in the GSTR-1 filing of the seller — undeclared invoices are a fraud risk and disqualify the specific invoice from discounting. This cross-check is done via the Account Aggregator-linked GSTN API.

Repayment risk for recourse vs non-recourse discounting: in recourse discounting (most NBFC products), if the buyer does not pay, the seller must repay the finance amount. In non-recourse discounting (typically PSU-backed TReDS), the risk sits with the financier. The AI explains this distinction and factors the recourse obligation into the seller's FOIR calculation.

Working capital limit renewal is a retention-critical event in SME banking: an SME that loses its CC/OD limit migrates to a competitor. Yet most banks manage renewals through a manual, document-heavy process that creates anxiety and delays — exactly the conditions that drive switching.

Kallix's renewal qualification call begins with a positive framing: 'Your working capital limit of Rs [X] is due for annual review. I see your account has been active for [N] months — I will check your renewal eligibility and any scope for a limit increase in this call.' This opener signals that the bank is evaluating both maintenance and enhancement — not just deciding whether to cut the limit.

Utilisation analysis: the AI calculates the average CC/OD utilisation over 12 months from the account statement. Under-utilisation (below 30% average) signals the limit is too large for current needs — the bank may offer a lower limit with better pricing. High utilisation (above 85% consistently) signals the business needs more — the AI presents a limit increase as a proactive suggestion rather than waiting for the customer to ask.

Adverse event check: bounced cheques in the CC/OD account, overdue interest, and any CIBIL flags during the year are pre-loaded from the CRM. The AI addresses these proactively: 'I see your account had a temporary overdrawing in [month] — can you briefly explain the context? This helps us process your renewal with accurate information.' Proactive explanation opportunities reduce technical rejections by 22–30%.

For limit increase requests: the AI collects the current year's turnover (GST-linked), the specific business reason for needing additional limit (new order, new contract, seasonal peak), and confirms FOIR eligibility for the incremental limit. Renewal calls with an embedded limit increase offer achieve 28–36% higher revenue per interaction than plain renewal calls.

Co-applicant and guarantor calls are time-sensitive: the primary applicant has qualified and is waiting for co-applicant confirmation to advance to documentation. Delays in reaching the co-applicant are a leading cause of lead leakage — the applicant applies elsewhere while waiting.

Kallix's co-applicant call begins with identity and relationship confirmation: 'I am calling from [Bank] regarding [Applicant Name]'s application for a [loan type]. [Applicant Name] has listed you as the co-applicant/guarantor. Are you aware of this application and do you consent to be included?' The consent is recorded and logged. Without explicit consent, the AI does not proceed with data collection.

Co-applicant income qualification: for home and education loans, co-applicant income is typically added to the primary applicant's income to increase eligibility. The AI runs the combined FOIR check: total EMI obligations of both applicants / combined monthly income. For education loans: co-applicant (usually parent) must demonstrate sufficient income to service the EMI during the moratorium period when the student is studying.

Guarantor obligations explained: many guarantors do not understand the extent of their liability. The AI explains in plain language: 'As a guarantor, if [Applicant Name] is unable to repay the loan, the bank has the right to recover the full outstanding amount from you — including from your property if necessary. Please ensure you are comfortable with this obligation before confirming.' This transparent explanation reduces future guarantor disputes significantly.

For guarantors with existing loan liabilities: the AI calculates the contingent liability impact — the guarantor's credit bureau may show the guaranteed loan as a contingent liability, which affects their own future borrowing capacity. The AI flags this: 'Being a guarantor on this loan will appear as a contingent liability on your CIBIL report — this may affect your eligibility for a personal loan in the future if you need one.'

Franchise and dealer loans are a growing SME credit category in India — driven by organised retail, QSR (quick service restaurant) chains, automotive dealerships, and pharma distribution franchises. Lenders who specialise in this segment (Kotak, IndusInd for auto dealers; HDFC Bank for QSR chains) offer preferential rates because the borrower's revenue predictability is higher than a standalone business.

Kallix's franchise loan qualification covers the franchise agreement assessment: remaining tenure of the franchise agreement is a key underwriting criterion — a loan with a 5-year term should have a franchise agreement tenure of at least 5 years remaining, or the franchisee should demonstrate a renewal likelihood. Lenders typically require the franchise agreement tenure to exceed the loan tenure.

Royalty obligation in FOIR: the AI includes the monthly royalty payment in the applicant's FOIR calculation. For a QSR franchisee paying 6% royalty on Rs 12 lakh monthly revenue = Rs 72,000/month in royalty, which counts as a fixed obligation alongside EMIs.

Franchisor performance benchmarking: if the applicant's sales performance versus brand targets is available (from the franchise disclosure document or FDD), the AI uses the target achievement rate as a proxy for credit quality: franchisees achieving 80%+ of brand targets are significantly lower credit risks than those consistently below 60%.

Dealer stock finance: for automotive dealerships, stock finance (floor plan credit) against the inventory of vehicles is a separate facility from the dealer's working capital loan. Kallix AI identifies whether the applicant needs stock finance (which is typically provided by the OEM's captive finance arm at preferential rates) or general business credit.

Green lending is the fastest-growing SME and retail lending sub-category in India — driven by the government's solar rooftop subsidy scheme (PM Surya Ghar, targeting 1 crore rooftop installations), FAME-II EV subsidies, and RBI's Priority Sector Lending (PSL) green finance classification.

Rooftop solar loan qualification: the AI collects the system size (kW), DISCOM approval status (required before installation), whether the applicant is applying under PM Surya Ghar (which offers central subsidy of Rs 30,000–78,000 per kW depending on system size), and the electricity bill reduction projection. The payback calculation: if an Rs 1.8 lakh 3kW system reduces the electricity bill by Rs 2,500/month, the payback is 72 months — and the AI presents this as the EMI comparison: 'Your monthly EMI at Rs 3,200 is offset by Rs 2,500 in electricity savings — net additional outflow is Rs 700/month.'

EV fleet loan qualification: for commercial fleet operators converting to EVs, the AI qualifies both the vehicles (FAME-II subsidy eligibility, battery capacity, range) and the operator's repayment capacity. The AI calculates total cost of ownership (TCO) advantage: EV per-km fuel cost of Rs 0.8–1.2 vs CNG Rs 3.5–4.5 or diesel Rs 7–9. TCO advantage over 5 years quantified in rupees is a strong conversion lever.

Energy efficiency loans: for SMEs installing energy-efficient equipment (LED lighting, VFD motors, compressed air system upgrades), the Bureau of Energy Efficiency (BEE) star rating certification of the equipment is the qualification trigger. Loans under the National Mission for Enhanced Energy Efficiency (NMEEE) qualify for SIDBI concessional refinancing.

Loan rejection handling is one of the highest-empathy scenarios in financial services — the applicant is financially stressed and may be turned away from a second lender by a poorly handled call. Kallix's rejection rehabilitation approach converts a dead end into a structured remediation pathway.

The AI begins by asking the applicant to share the reason they were given for rejection: 'To help you explore the right options, could you share what the previous lender mentioned about their decision?' The three most common rejection reasons and Kallix's routing:

(1) Low CIBIL score (below 650): the AI does not attempt to route to another bank — most banks have similar CIBIL floors. Instead, it routes to an NBFC with more flexible scoring (Bajaj Finance, HDB Financial, Fullerton India) that accepts applicants with CIBIL 600–650, or to a secured loan option (LAP, gold loan) where the asset quality compensates for the credit score.

(2) High FOIR (existing debt burden too high): the AI checks whether adding a co-applicant's income could reduce the FOIR to an eligible level, or whether refinancing an existing high-EMI loan at a lower rate could reduce the monthly obligation and improve FOIR.

(3) Insufficient income documentation: the AI confirms which alternate income proofs are available (GSTR, bank statements, ITR) and routes to lenders with lower documentation thresholds or NBFC-fintech partnerships that use alternative data.

Important: the AI does not make another CIBIL inquiry on a recently rejected applicant until 60 days have elapsed — multiple inquiries in a short period depress the score further. The AI explains this to the applicant: 'Every loan inquiry reduces your CIBIL score slightly — let us build the right application before pulling your bureau again.'

Multi-product loan qualification deployments are more complex than single-product deployments because each loan type has its own eligibility rule set, document checklist, bureau threshold, KFS template, and LOS submission format. Kallix manages this through a modular product configuration layer — each loan type is configured independently within the same platform.

Weeks 1–2: Infrastructure integration. LOS API connection (Newgen, FinnOne, or Salesforce FSC), bureau API setup (CIBIL, Experian, and/or CRIF High Mark — each requires a separate API key and data usage agreement), and Account Aggregator integration (Finvu, OneMoney, CAMS Finserv, or PhonePe AA — each bank has different AA partnerships).

Weeks 2–6: Product configuration per loan type. Each loan type requires: eligibility rule configuration (CIBIL threshold, LTV policy, FOIR cap, income type handling), KFS template (different fee structure for vehicle vs education vs SME), document checklist (loan-type-specific with conditional items), and script variation (vehicle conversations are quick and transactional; SME conversations are longer and more qualitative). Rule configuration is done via Kallix's no-code admin console by the lender's product team — Kallix's team validates and tests.

Weeks 4–5: Account Aggregator integration testing. AA consent flows are tested across all borrower types: salaried (bank statement pull), self-employed (bank statement + GSTR pull), and SME (business current account + GSTR + ITR). AA data parsing is calibrated against the bank's income calculation methodology.

Weeks 6–8: UAT across all loan types (200+ test cases per loan type) and phased go-live — vehicle loans typically go first (highest volume, simplest rules), then education, then SME. A 30-day hyper-care period follows.