Proposal-7843 · structured product concentration at 50% (threshold 40%). Two alternatives: reduce the amount, or document a deliberate exception with manager sign-off.
Investment proposals go out only when they are consistent with the MiFID II profile.
MiFID II Disclosure Gate checks every investment product proposal before it reaches the customer. It verifies consistency with the customer's MiFID II suitability profile, the product's target market, and the applicable suitability rules. Pass, flag, or block — with a structured rationale. Integration with the bank's CRM and portfolio management system is delivered during the project.
MiFID II Disclosure Gate at work.
I'm proposing the customer reduce the amount. The customer accepts. Amount updated below threshold.
Proposal-7843 rechecked · pass. Event traced for Consob audit.
Why it exists.
MiFID II requires investment service providers to check that every proposal fits the customer's profile. Each product — securities, funds, ETFs, structured products, unit-linked and index-linked policies with a financial component — must be consistent with the customer's suitability profile, the POG-declared target market, the concentration limits. The adviser works across many customers and many products; misalignment errors are recurring.
How it checks the proposal.
The agent activates pre-send. For each proposal: it reads the product (risk characteristics, POG target market, concentration), reads the customer's suitability profile from the bank's CRM, applies the declarative suitability rules. Pass for consistent proposals, flag for borderline with alternatives, block for incompatible proposals with a structured rationale and article references.
The decision stays with the adviser.
The agent does not replace the adviser or the sales operator. It sets the table for the pre-send check. The final decision to send, to document a deliberate exception with the customer's rationale, to choose between the proposed alternatives stays with the adviser, following the bank's procedures.
Who it serves and where it applies.
Financial adviser or sales operator
Gets real-time feedback on proposal consistency. Recovers the time spent on manual checks, reduces misalignment errors, keeps the commercial flow moving without compliance consultation pauses. The check is set up before sending, not after the damage.
MiFID II compliance officer
Sees the verified proposal flow in a structured way. Pass, flag, and block decisions are inspectable for Consob and IVASS audit. The network's error patterns become visible for targeted training — not through subjective perception, case by case.
Sales network head
Sees the network's error patterns in aggregate. Training becomes targeted to the actual misalignment vectors — concentration, target market, profile — not generic scenarios. The network improves on the vectors the data signals.
A concrete example.
The adviser prepares a structured product proposal.
A financial adviser at a bank prepares a proposal to purchase a structured product for a customer. The agent activates pre-send on the configured channel. It reads the customer's MiFID II profile: balanced profile, 5-10 year horizon, intermediate market experience. It reads the proposed product: 90% capital-protected structured product with six-year barriers, target market "balanced customer".
Profile, target market, concentration.
Suitability check in three steps. Profile consistency: consistent. Target market: consistent. Concentration limits: the customer already holds 35% of their portfolio in similar structured products; the new proposal would bring concentration to 50%, above the 40% threshold the bank has set for balanced profiles. Outcome: flag with a structured rationale and two alternatives — reduce the amount, or document a deliberate exception.
The adviser chooses. The agent records.
The adviser sees the flag in the work channel, proposes reducing the amount to the customer, the customer accepts. The final proposal falls within the threshold, the adviser confirms, the email goes out. The full event — proposal, check outcome, chosen alternative — remains in the runtime audit registry for Consob and IVASS inspection.
Configuration and technical resources.
The rules are declarative. The bank's MiFID II compliance team and risk officer define, in a readable format, the suitability rules per product category, concentration thresholds, suitability profile rules. The rules live in the bank's repository, versioned, validated at agent startup. The three standard checks cover the most frequent non-compliance vectors: profile consistency, POG target market, concentration limits.
- Language
- TypeScript (Node.js)
- LLM model
- customer's choice: Anthropic, OpenAI, Mistral, open source models hosted internally, AWS Bedrock for a private model
- Built-in controls used
- pii-detector, credential-detector, topic-guardrail, tool-param-validator
- Native delivery channels
- Slack, Telegram, HTTP OpenAI-compatible (pre-send proposal webhook)
- Bank CRM and portfolio management system integration
- dedicated adapter built during delivery
- POG registry integration
- dedicated adapter built during delivery
- MiFID II rules
- declarative, versioned, written by the compliance and risk team
- Memory
- persistent per instance, pgvector + PostgreSQL FTS
- Registry
- append-only, queryable with a standard SQL client (Consob/IVASS audit inspectable)
Frequently asked questions about the agent.
For proposals clearly incompatible with the customer's suitability profile — a high-risk product for a conservative profile, a product outside the declared target market — the agent generates a block with a structured rationale. For borderline proposals, a flag with alternatives. The rule that distinguishes a block from a flag is declarative and written by the bank's compliance team.
When a customer knowingly subscribes to a proposal that exceeds the concentration threshold or their own profile, the agent records the exception with the customer's rationale and the bank manager's sign-off. The trace remains in the runtime audit registry for Consob and IVASS inspection.
The typical pattern for MiFID II Disclosure Gate is 12-18 weeks. Discovery 2-3 weeks, writing MiFID II rules with compliance and risk 4-6 weeks, bank CRM and portfolio management system integration 4-6 weeks, hand-off 2-3 weeks. The effective duration is defined in discovery on the real case.
From a 30-minute conversation to the squad in production.
A 30-45 minute conversation to understand how MiFID II Disclosure Gate would configure to the bank's case. Which bank CRM, which product categories are in scope, which existing suitability rules to transfer into declarative format.