Every year, India’s capital markets platforms, from discount brokers to full-service wealth managers, spend heavily to bring investors on board: digital campaigns, referral programs, influencer partnerships and AMFI-backed financial literacy drives. And every year, a large share of those prospects walk away before they ever fund an account.

The losses add up to a structural problem that most platforms have accepted as a cost of operating in India’s regulated, document-heavy financial system. India’s demat account base had crossed 21 crore by late 2025, according to SEBI’s monthly bulletin, up from less than 4 crore at the end of 2019. The market is expanding fast. Yet industry practitioners consistently observe that a meaningful share of account-opening journeys begun each year are never completed, an abandonment problem that rarely shows up in the headline growth numbers.

The economics are unforgiving. Customer acquisition costs in Indian wealth management and broking run high, and they climb further for HNI segments. Every investor who abandons an incomplete account turns that spending into a sunk cost. The acquisition budget worked as intended; the onboarding process is where the money leaks.

The acquisition budget worked as intended. The onboarding process is where the money leaks.

What makes this costly is that investors who abandon rarely come back. They open an account with a competitor whose process was a little less painful. The industry’s answer for years has been better rules-based automation: pre-filled fields, save-and-resume features, SMS reminders to re-engage. These helped at the margins, and they left the underlying problem in place.

The real question is whether AI can solve what rules-based automation cannot. Across a growing number of Indian capital markets platforms, the answer is taking shape, and the gap between early adopters and everyone else is getting harder to ignore. In my work with platforms that have rebuilt onboarding around integrated AI infrastructure, I have seen drop-off fall by as much as 60 percent.

Investor onboarding abandonment is the rate at which prospective clients begin a registration or account-opening process and exit before completing it. In India, this plays out across a distinctive regulatory stack: SEBI’s KYC Registration Agency framework, Aadhaar-based eKYC, DigiLocker document verification and in-person verification requirements for certain investor categories. Drop-off can happen at any stage: Aadhaar OTP verification, PAN validation, bank-account linking through a penny-drop, nominee registration or while waiting for NSDL or CDSL activation. Each exit represents wasted onboarding cost, a weaker CAC-to-LTV ratio and lost ground in a market where platforms like Zerodha, Groww and Upstox keep raising the bar.

India’s onboarding breaks down in three predictable places. Each one sheds a different kind of applicant. Switch between them below.

AI-powered document intelligence combines optical character recognition, liveness detection and real-time validation against Protean PAN verification APIs and UIDAI’s Aadhaar verification gateway, turning identity verification from a bottleneck into a background step. Manual verification once meant human review queues measured in hours. AI document processing can complete the same checks in seconds, and for standard identity documents its accuracy can match or exceed manual review. In India, this includes AI-assisted reconciliation of name mismatches between Aadhaar and PAN records, one of the most common technical failure points in domestic KYC flows.

Form sequencing changes next. AI reorders and suppresses fields in real time, based on investor segment, jurisdiction and product. A retail investor in Jaipur opening a basic demat account should not see the same form architecture as an NRI in Dubai setting up a PIS account under RBI’s Portfolio Investment Scheme. Conditional field suppression drops questions that are irrelevant to a given investor’s profile based on answers already provided. Form completion gets measurably faster, with no loss of data quality or regulatory completeness.

The approval wait is the third fix. AI-powered decision engines can generate live status updates and indicative approval timelines inside the session, calibrated to CDSL and NSDL processing queues, KRA verification timelines and document completeness. An investor who can see where the application sits experiences the wait as a managed process, with a clear sense of what is pending and when it should clear. For India’s HNI segments, where the expectation of visibility is highest, this has an outsized effect on retention.

The 60 percent drop-off reduction comes from stacking improvements across several failure points at once, built as integrated infrastructure instead of isolated tools. No single feature gets a platform there. Platforms that fix one or two failure points, usually document verification or form simplification on their own, make real gains and still capture only part of what is available.

Platforms that integrate the full stack, where document intelligence, adaptive sequencing and real-time decision transparency share data and operate as one system, are reaching 60 percent and beyond. In India, that integration also has to span the multi-registry architecture of SEBI’s KRA system, pulling KYC status from CVL, NDML, CAMS, Karvy and DotEx into a single investor intelligence layer.

Document checks, smarter forms and a visible approval clock each help on their own. The 60 percent only appears when they run as one system.

The practical takeaway for platform leaders comes down to one test: is the AI built as infrastructure, with unified data flows, feedback-loop learning and cross-stage optimization, or as a feature that improves one stage on its own? That design choice, more than any single tool, determines the result.

Three questions separate a durable build from a quick patch.

On the first point, the regulation is now explicit. SEBI’s Regulation 16C, in force from February 10, 2025, makes every regulated entity using AI tools solely responsible for the accuracy of those outputs and for compliance with applicable laws. Auditability stops being a feature and becomes a condition of doing business.

India is still adding investors at a pace few markets can match, and the demat base has more than quintupled since 2019 as equity culture spreads well beyond the largest cities. The investor joining today, a first-time participant in Patna or Coimbatore whose baseline for digital experience is set by the ease of UPI payments and Aadhaar-linked services, has close to zero tolerance for opaque, multi-day onboarding.

The platforms that have crossed the 60 percent threshold are not done. They are training their models on richer behavioral data, widening decision transparency and pushing AI deeper into the post-onboarding relationship. In a market adding crores of potential investors a year, with acquisition costs rising and the margin for onboarding failure narrowing, the platforms moving now are setting the conversion standard their competitors will be measured against. That is a durable strategic asset, and the time to build it is before the gap grows too wide to close.