Skip to content
GUIDANCE OUTLOOKFinancial Services

Piramal Finance issues fresh guidance

Piramal FinancePIRAMALFIN

TL;DR

The cited Investor Day 2026 material does not provide a numeric endpoint for either the Cost-to-Income ratio or CAC for the retail lending book. The quantified target disclosed is instead the retail operating-expense-to-AUM ratio, which had been reduced by 25 bps from the earlier expectation; Q2 FY27 was at 3.48%, below the lower end of that earlier range.

With the Investor Day 2026 presentation highlighting an 'AI-Driven Growth Strategy,' what specific targets have been set for the Cost-to-Income ratio and customer acquisition costs (CAC) for the retail lending book, and how do these projections reconcile with the operational expenditure trends disclosed in the most recent quarterly filings?

The cited Investor Day 2026 material does not provide a numeric endpoint for either the Cost-to-Income ratio or CAC for the retail lending book. The quantified target disclosed is instead the retail operating-expense-to-AUM ratio, which had been reduced by 25 bps from the earlier expectation; Q2 FY27 was at 3.48%, below the lower end of that earlier range [1].

The operating trend is supportive of the strategy. In Q1 FY27, AUM grew approximately 25% YoY, while operating expenses grew only about 10%, driving the Cost-to-Income ratio down to 52.5% [5]. By Q2 FY27, retail opex-to-AUM had reached 3.48% [1]. This suggests that scale, digitisation and employee productivity are generating operating leverage.

However, the reconciliation is incomplete:

  • Cost-to-Income and opex-to-AUM have different denominators. The latter can improve through AUM growth even if income conversion or product mix changes differently.
  • CAC cannot be inferred from opex-to-AUM. The sources do not identify acquisition spend, customers acquired, or CAC per acquired customer.
  • The evidence therefore validates the direction of the AI efficiency thesis, but not a specific future Cost-to-Income or CAC target. Any precise target for either metric would require the Investor Day slide or quarterly disclosure that explicitly states it.
MetricInvestor Day positionLatest operating evidenceReconciliation
Cost-to-Income ratioNo specific future endpoint could be confirmed from the cited presentationCompany-wide Cost-to-Income was 52.5% in Q1 FY27, versus 65.6% in Q1 FY26 [2]Directionally consistent with operating leverage, but the Investor Day material does not establish a formal target such as “below X%”
Retail opex-to-AUM3.48% in Q2 FY27, below the lower end of the earlier guided range [1]Q2 FY27 retail opex-to-AUM was also described as having declined from roughly 5.5% to 3.5% as AI and technology were deployed across lending processes [3]Strong evidence of improving cost intensity, although opex-to-AUM is not the same metric as Cost-to-Income
CACNo numerical CAC-per-customer or CAC-as-a-share-of-disbursement target could be confirmedAI-linked productivity gains are visible: monthly disbursements per employee rose 33% YoY from September 2024 to September 2026, with AI contributing half the uplift [4]Supports lower acquisition and servicing intensity, but does not quantify CAC or prove that CAC has declined

Regarding the AI-driven growth outlook, what is the confirmed timeline for the deployment of these AI tools across the retail loan origination process, and how does this strategy specifically alter the projected AUM mix between the retail and wholesale segments compared to the guidance provided in the previous annual report?

The confirmed timeline is “already deployed and expanding,” not a dated future rollout. At the November 2026 Investor Day, Piramal Finance said AI was already embedded from sales through collections, with the deployment period described as the preceding 12 months. The retail-origination stack includes sales and distribution intelligence, document intelligence, AI credit decisioning, bank-statement analysis and direct-to-disbursement tools. [6] [6] The company has not provided a date by which every retail product or branch-originated loan will be fully AI-enabled; mortgages and personal loans were still described as areas where AI-enabled processes were being developed. [3]

AUM mix implication

There is no confirmed numeric revision to the retail-versus-wholesale AUM target in the cited Investor Day material. The company reiterated a total AUM growth ambition of 20–25% CAGR, but did not provide a before-and-after mix bridge against the previous annual report. [7]

The latest portfolio chart provides a directional indication rather than a new projection. Classifying housing, LAP, used-car, LAS, salaried personal, unsecured business, rural micro, digital and other retail loans as retail, and corporate mid-market and real estate as wholesale, gives an approximate September 2026 mix of:

The categories sum to 101% because of rounding. This is an estimated current product mix, not management’s revised forward guidance.

Analytical read: AI is intended to make retail expansion more scalable by reducing manual underwriting and processing time, improving productivity and supporting direct-to-disbursement. That points to a more retail-led growth model and relatively lower dependence on wholesale growth, but the exact change versus the previous annual report’s projected mix could not be confirmed from the cited disclosures.

SegmentApprox. share of AUMBasis
Retail88%Derived from disclosed product categories [8]
Wholesale13%Derived from corporate mid-market and real estate categories [8]

In the context of the AI-driven credit underwriting models discussed at the Investor Day, how do the projected credit cost improvements compare to the historical credit cost ratios of the retail book, and what specific data points from the existing loan book are being utilized to validate these AI-driven risk assessment models against industry peers?

The Investor Day did not provide a forward retail credit-cost target. It reported an approximately 50% reduction in credit risk over the preceding 12 months, which is not the same metric as a 50% reduction in credit cost [9]. The disclosed credit-cost series was 1.5% in FY25, 1.6% in FY26, 1.5% in H1 FY26 and 1.5% in H1 FY27 [10].

A mechanical 50% reduction applied to the historical 1.5%-1.6% credit-cost range would produce 0.75%-0.80%, but that is only an arithmetic illustration—not management guidance—because “credit risk” and “credit cost” are not established as identical measures in the presentation.

The model-validation framework therefore appears to rely primarily on internal loan-book evidence: quarterly 90+ DPD trends, vintage delinquency at 6MOB, and borrower/customer-friction indicators, supported by granular document, bank-statement, bureau and fraud signals. The presentation also cites 20,000+ features analysed and 400+ AI/ML/agentic models, although it does not identify which features specifically drive the loan-book back-testing [11].

Peer comparison remains unproven. Comparable FY26-FY27 retail credit-cost ratios, AI-underwriting metrics or model-validation results were not reported for LIC Housing Finance, Bajaj Housing Finance, PNB Housing Finance, Aadhar Housing Finance or Sammaan Capital. Accordingly, the evidence supports improvement versus Piramal Finance’s own historical retail-risk indicators, but not a quantified advantage versus those industry peers.

EvidenceReported measureInterpretation
Historical credit cost1.5%-1.6% across FY25-H1 FY27 [10]Stable baseline; the slide does not identify this as a retail-only ratio
AI impact claim~50% reduction in credit risk over the past 12 months [9]Risk improvement claim, not a directly comparable credit-cost forecast
Retail unsecured outcome90+ DPD declined from 1.65% in Q2 FY25 to 0.86% in Q2 FY27 [9]A 0.79 percentage-point, or 47.9%, reduction derived from the reported endpoints; this is the clearest retail-book validation signal
Retail cohort monitoring30+ DPD at six months-on-book, alongside retail customer-complaint tracking [2]Early-vintage performance and borrower-friction indicators used to monitor underwriting and servicing quality
Underwriting inputsDocument reading, bank-statement analysis, commercial bureau data and FI/fraud detection through AI tools [6]Core data inputs for automated risk assessment, rather than reported peer-benchmark outcomes

Sources

  1. [1]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.37
  2. [2]15 September 2026 BSE Limited National Stock Exchange of India Limited 1 Floor, New Trading Wing, Exchange Plaza, 5 Floor, — BSE India, 2026-09-15T00:00:00
  3. [3]Piramal Finance Scales AI-Powered Hands-Free Collections 2.5x to ₹754 Crore — Financialexpressb2B, 2026-11-06T00:00:00
  4. [4]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.40
  5. [5]9 October 2026 Company Update | Sector: Financials - NBFC ... — Sainathinvestment, 2026-10-09T00:00:00
  6. [6]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.25
  7. [7]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.45
  8. [8]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.5
  9. [9]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.41
  10. [10]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.39
  11. [11]Piramal Finance Investor Day 2026: AI-Driven Growth Strategy and Financial Outlook — 2026-11-19T09:27:20.050000, p.22

Keep digging

With the Investor Day 2026 presentation highlighting an 'AI-Driven Growth Strategy,' what specific targets have been set for the Cost-to-Income ratio and customer acquisition costs (CAC) for the retail lending book, and how do these projections reconcile with the operational expenditure trends disclosed in the most recent quarterly filings?

Ask Copilot
Logo

Unlock financial AI for your firm