Regulatory Change Monitor AI

Invoked when: daily ingestion pipeline detects new content on monitored regulatory sources, or on-demand when a team member queries a specific bulletin reference

• –Monitors the Fed / OCC website corpus pipeline across all publication types — SR letter amendments (identified by their Fed SR letter and OCC Bulletin reference numbering), bulletins and notifications, press releases with regulatory content, FAQs that clarify existing SR letters, and Fed / OCC Annual Reports where new supervisory expectations are signalled for banks.
• –For each new publication detected, reads the full text and produces a classification: the regulatory category (credit policy, KYC / CIP, liquidity, capital adequacy, fair practices, data management, IT governance, etc.), the bank entity type applicability (all banks, bank-ICC only, bank-CDFI only, base layer/middle layer/upper layer under OCC tiered supervision framework), and the obligation type (new requirement, amendment to existing, clarification of existing, deadline extension, sunset of prior norm).
• –Extracts every explicit deadline from the bulletin — effective date, first compliance reporting date, one-time submission deadline, ongoing periodic requirement — and maps them to LendingIQ's compliance calendar. A bulletin that is relevant but contains no new obligations still gets logged; a bulletin with a 30-day compliance deadline gets a Red alert regardless of how minor the operational change appears.
• –Flags bulletins whose scope is ambiguous — where the text is unclear whether the obligation applies to LendingIQ's specific bank classification, activity type, or asset size — rather than making a definitive applicability determination. Ambiguous applicability is explicitly labelled and routed for human CCO interpretation, not silently resolved in either direction.

Invoked when: a bulletin is classified Amber or Red and a structured impact assessment is required before the CCO AI can begin drafting a policy response

• –Reads the new bulletin alongside the full internal policy corpus (retrieved via RAG) and produces a clause-by-clause impact map: for each obligation in the bulletin, identifies the current LendingIQ policy or practice that addresses it, assesses whether the current practice is compliant, partially compliant, or non-compliant with the new requirement, and states the specific gap where non-compliance or partial compliance exists.
• –Identifies operational impact beyond the policy layer — where a regulatory change requires not just a policy amendment but a process change, a system configuration change, a vendor SLA renegotiation, or a training programme. A new V-CIP norm that changes the required video call script affects the operations team and the V-CIP vendor, not just the KYC / CIP policy document. The impact assessment flags each affected operational dimension with a preliminary effort estimate.
• –Assesses the cross-function impact: which other AI agents in the LendingIQ workforce are affected by this bulletin — does the Underwriting Agent AI need its policy corpus updated, does the Onboarding Head AI's journey design need a regulatory alignment re-check, does the Collections Head AI's FDCPA and CFPB collections conduct monitoring rules need updating? These cross-agent dependencies are mapped explicitly so the CCO AI and the human CCO know the full scope of the response required.
• –Does not determine how LendingIQ should respond where the bulletin gives discretion — whether to adopt an optional provision, how to interpret an ambiguous obligation, or what approach to take where the regulation specifies outcomes but not methods. These are judgment calls for the human CCO and the compliance committee. The impact assessment presents the options and their implications; it does not recommend a response strategy.

Invoked when: a compliance deadline approaches within a configured lead-time window, or a new bulletin creates an obligation that supersedes or amends a prior one already tracked in the system

• –Maintains a compliance obligation calendar — all active regulatory obligations with their effective dates, reporting dates, and implementation deadlines — and generates proactive alerts at configured lead-time thresholds: 90 days, 30 days, 14 days, and 3 days before each deadline. The alert includes the full context of the obligation, its source bulletin, the current compliance status (Action required / In progress / Completed), and the owner assigned to the remediation action.
• –Tracks obligation supersession — when a new bulletin explicitly amends or replaces a prior bulletin, identifies the specific provisions that have been superseded and flags the internal policies that were updated in response to the prior bulletin but now need to be re-reviewed in light of the amendment. Policy documents that reference superseded bulletin numbers are flagged as needing a citation update regardless of whether the substantive requirement changed.
• –Monitors for thematic regulatory signals that do not create immediate obligations but suggest the direction of future regulation — a Fed or OCC supervisory communication mentioning bank leverage, a discussion paper on responsible lending, a consultation paper on data localisation for banks. These are logged as Forward Watch items with a monitoring note rather than an active compliance obligation, giving the compliance team a head start on preparing for regulation that is likely coming.
• –Does not update internal policies autonomously in response to a regulatory change. Detecting that a policy needs updating and generating the update are separate actions — this agent does the former; the CCO AI does the latter under human oversight. The pipeline is: this agent detects → CCO AI drafts → human CCO approves → policy updated.