← Agent catalogue

AI Agent Profile · LendingIQ · Bengaluru

Data Protection Officer AI

Invoked via: internal orchestration APIRuntime: AWS Bedrock · ap-south-1Model: Claude Sonnet 4Context window: 200K tokens

DivisionCompliance

Resume

What this agent does

The DPO AI monitors LendingIQ's data processing activities for compliance with India's Digital Personal Data Protection Act 2023, handles the operational workload of data rights requests, drafts the organisation's response to data breach incidents, and runs DPDP compliance audits across systems, vendors, and consent records. It is the intelligence and drafting engine of the data protection function. The named human DPO holds statutory accountability and signs all regulatory communications.

Primary functions

Consent Architecture Design & Monitoring

Triggered at product launch or periodic audit

Invoked when: new product or data processing purpose introduced, consent audit due, or DPDP regulatory update received

  • Maps every personal data element LendingIQ collects against the processing purposes declared at collection — credit assessment, KYC, collections, marketing, analytics — and checks whether valid consent or a legitimate use basis under the DPDP Act exists for each purpose-data combination.
  • For new products or features: reads the product specification and data flow, identifies every personal data element the product will process, and drafts the consent notice language — specific, granular, plain-language, purpose-linked — as required under Section 6 of the DPDP Act. Does not produce one-size-fits-all consent blankets.
  • Monitors the consent management platform for withdrawals, expired consents, and purpose drift — where data is being used for a purpose the borrower consented to at origination but which has since expanded without fresh consent. Flags these proactively before they become violations.
  • Cannot validate whether the technical consent capture mechanism on the product UI actually works as specified. It audits the consent records and the legal architecture; a separate technical QA process must validate the UI implementation.
Output: Consent architecture map — data elements by purpose, consent basis for each, gap flags where consent is absent or expired, and draft consent notice language for new processing purposes.

Data Rights Request Handling

Triggered on incoming request from data principal

Invoked when: borrower submits access, correction, erasure, grievance, or nomination request via the prescribed channel

  • Classifies the incoming request by right type under the DPDP Act — right to access personal data (Section 11), right to correction and erasure (Section 12), right to grievance redressal (Section 13), right to nominate (Section 14) — and identifies the applicable response timeline and obligations.
  • For access requests: reads the consent records and processing log for that data principal, compiles the data inventory the borrower is entitled to receive, and drafts the response in the prescribed format. Flags categories of data that may be withheld under lawful exemptions — e.g., data held for legal proceedings or regulatory compliance purposes — with the specific statutory basis cited.
  • For erasure requests: checks whether erasure is permissible under the Act given the borrower's current relationship with LendingIQ — an active loan account creates legal and regulatory retention obligations that override erasure rights. Drafts a response that explains the retention basis clearly, not a blanket refusal.
  • For correction requests: identifies what data elements are in scope, whether the correction affects downstream regulatory data (credit bureau reporting, ITR cross-check records), and flags to the human DPO where a correction has compliance implications before the response is sent.
Output: Classified request record, draft response to the data principal, internal processing note on exemptions or downstream implications applied, and a flag to the human DPO for any request outside standard handling parameters.

Breach Response

Triggered on confirmed or suspected breach incident

Invoked when: CISO or security team raises an incident that may constitute a personal data breach under DPDP Act Section 8(6)

  • Reads the incident report from the security team — what data was accessed or exfiltrated, which data principals are affected, how the breach occurred — and applies the DPDP Act's breach notification criteria: does this constitute a "personal data breach" requiring notification to the Data Protection Board and to affected data principals?
  • Drafts the breach notification to the Data Protection Board — what happened, the categories and approximate number of data principals affected, the likely consequences, and the measures taken — in the format and timeline the Act requires. Clearly marked as a draft for the human DPO to review, approve, and submit.
  • Drafts the communication to affected data principals — plain-language, specific about what data was involved, what they should do, and what LendingIQ is doing — for the human DPO and communications team to approve before despatch.
  • Does not perform forensic investigation of the breach. It cannot access systems, review logs, or determine root cause — that is the CISO and security team's function. It works from the incident report provided to it and flags where the report lacks information needed for a complete notification.
Output: Breach classification assessment (notifiable or not, with reasoning), draft Data Protection Board notification, draft data principal communication, and a checklist of information gaps in the incident report that must be resolved before notification.

DPDP Compliance Audit

Triggered on annual cycle or regulatory change

Invoked when: annual DPDP audit due, new Rules notified, or post-breach review required

  • Reads the DPDP Act and Rules (via RAG), the full internal privacy policy and data retention schedule, all vendor Data Processing Agreements, and the consent management platform audit logs — and produces a structured gap analysis: every obligation under the Act mapped to LendingIQ's current practice, with a pass/fail/partial verdict and the specific gap described.
  • Covers all eight domains of DPDP compliance: consent management, notice adequacy, data principal rights operationalisation, data fiduciary obligations, processing restriction (children's data, sensitive data), security safeguards, breach response readiness, and Data Protection Officer designation and access.
  • Audits vendor Data Processing Agreements against the Act's requirements for Data Processors — does the DPA require the vendor to implement adequate security measures, restrict sub-processing, delete data on termination, and notify LendingIQ of breaches? Flags DPAs that are non-compliant or silent on material obligations.
  • Does not test technical security controls, penetration-test systems, or validate whether data is actually being deleted on schedule. The audit covers the legal and policy architecture; a separate technical audit must validate operational implementation.
Output: DPDP compliance audit report — obligation-by-obligation gap analysis across all eight domains, vendor DPA compliance assessment, prioritised remediation actions with statutory deadline references, and an executive summary for the board privacy committee.

Knowledge base

DPDP Act 2023 & Draft Rules (RAG)

Digital Personal Data Protection Act 2023 as enacted, MeitY's released draft Rules, official clarifications, and related MeitY notifications. Pipeline updated as Rules are finalised.

Consent Management Platform Records

Borrower consent records by purpose, consent capture timestamps, withdrawal log, and purpose-processing activity map. Injected at invocation — not stored between sessions.

Internal Privacy Policy & Retention Schedule

LendingIQ's privacy notice, data retention and deletion policy, data flow diagrams, and data inventory — retrieved via RAG, always current version.

Vendor Data Processing Agreements

DPAs with all data processors — bureau partners, cloud providers, fintech integrations, collection agencies. Audited for DPDP Act compliance in each audit cycle.

RBI Data Localisation & Privacy Circulars

RBI's data localisation requirements for payment data, storage norms for financial data, and KYC data handling guidelines. Applied in consent and audit functions.

General Data Protection Knowledge

Pre-training knowledge of global privacy frameworks (GDPR, CCPA), data protection principles, and privacy-by-design standards — used where DPDP Rules are silent or pending.

Hard guardrails

Will notSubmit breach notifications to the Data Protection Board or any regulatory body. All notifications are drafted by the agent and submitted by the named human DPO under their statutory authority.
Will notRespond directly to data principals. Draft responses are prepared for human review and approval before being sent. The agent has no channel to communicate with borrowers.
Will notExecute data erasure or correction in any system. It identifies what should be erased or corrected and by whom — the technical execution requires authorised human action in the relevant system.
Will notAccess live borrower personal data directly. It reads consent records, incident reports, and audit logs that are passed to it in structured form. It does not query production databases containing personal data.
Will notPerform forensic breach investigation. Root-cause analysis of a breach requires system log access, network forensics, and security tooling — the CISO team's domain. This agent works from the incident report the security team produces.
Will notProvide a definitive legal opinion on DPDP compliance. Its analysis is compliance intelligence for the DPO function, not legal advice. Where a matter carries material legal risk — enforcement action, litigation — qualified legal counsel must be engaged.

Known limitations

The DPDP Rules are not fully notified. Large portions of the Act's operational requirements — the specific form of consent notices, the Data Protection Board's complaint mechanism, the exact breach notification format and timeline — depend on Rules that MeitY has not yet finalised. The agent works from the Act and draft Rules but flags every pending Rule that affects its output.Maintain a tracked register of pending Rules and their expected impact. Each time MeitY notifies a Rule, update the RAG corpus and re-run the compliance gap audit against the new Rule.
Consent audit quality depends entirely on the completeness of the consent management platform records. If consent captures are incomplete, timestamps are missing, or purpose codes are inconsistently applied, the audit will reflect those gaps rather than fill them.Audit the consent management platform's data quality before invoking the consent architecture function. Define mandatory fields — data element, purpose code, consent timestamp, channel, withdrawal status — and enforce them at capture.
Erasure request handling is the most operationally complex function. An erasure request from a borrower with an active loan triggers a conflict between the DPDP right to erasure and RBI's regulatory data retention requirements. The agent identifies and flags this conflict; the human DPO must make the judgment call on how to respond.Build a clear internal decision matrix for erasure requests by account status — active loan, closed loan within retention period, closed loan past retention period, NPA under litigation. The agent applies the matrix; humans define it.
The agent cannot validate whether vendor DPAs have been operationalised. It audits the DPA text against the Act's requirements. Whether the vendor is actually implementing what the DPA says — deleting data on schedule, notifying breaches — requires vendor audits that go beyond document review.Supplement the agent's DPA text audit with periodic vendor compliance questionnaires and, for high-risk data processors, on-site or remote operational audits.
Children's data obligations under the DPDP Act are particularly sensitive and operationally complex. Verifiable parental consent, the prohibition on tracking and behavioural monitoring of children, and the definition of "child" in the lending context require careful human judgement that goes beyond rule application.Any product or feature that may collect data from individuals who could be minors must be reviewed by the human DPO and legal counsel before launch — do not rely solely on agent output for children's data decisions.
Agent Profile · Data Protection Officer AI · LendingIQ · BengaluruLast updated April 2026 · For internal use

Important Reads

Learn more about how to deploy Data Protection Officer AI to your lending workflow.