AI engine

How every client question is handled.

Every client question goes through the same pipeline before an answer is delivered. This is the engine behind Personalised Conversations.

See Personalised Conversations

How it works

Four steps, every time.

Client question

Should I be worried about my tech exposure given current market conditions?

Enters the platform

Step one

Retrieval

The question is grounded in the client’s actual portfolio and the sources your institution approves, scoped to your house. Generic AI answers from the open web are excluded.

Client portfolioApproved sourcesMarket data

Step two

Policy check

Before any response is generated, the question is classified against your compliance rules. Is this a suitability trigger? A regulated advice question? Your policy layer determines what can be answered, and how.

Suitability rulesDisclosure requirementsProduct guardrailsTier classification

Step three

Response generation

The answer is generated using your approved sources, your tone of voice, and your disclosure framework. The client receives a response that sounds like your institution, not a generic AI.

Your tone of voiceDisclosure frameworkApproved sources only

Step four

Logging

Every interaction is logged with full context. Your compliance team has a complete, exportable audit trail for every client conversation, searchable by client, date, classification, or advisor.

Full audit trailExportable logsCompliance review

Delivered to client

Compliant response

Nextvestment · Classified Tier 2 · Compliant

Based on your current portfolio allocation, your tech exposure is 34%, above your stated risk threshold of 25%. Given current market conditions, your house view suggests a cautious position on growth tech. Would you like to review rebalancing options with your advisor?

Timestamp logged

Classified Tier 2

3 sources retrieved

Policy checks passed

Disclosure applied

Audit trail written