AI-powered fact-find filling
Prospo can automatically fill in the client fact-find from a meeting transcript or any block of text. Instead of manually entering data from your notes, the AI reads the transcript and maps the information to the correct fields.
How to use AI fact-find filling
- Open the client record and go to the Fact-Find tab.
- Click Auto-fill from transcript (or the AI fill button).
- Prospo displays a preview of the fields the AI has populated from the most recent meeting transcript.
- Review each field. You can accept, edit, or clear any value.
- You can also use the voice-to-form ability to apply edits.
- Click Save to apply the changes to the fact-find.
What the AI extracts
The AI reads the transcript and extracts financial facts, mapping them to the correct sections of the fact-find. This includes:
- Personal and contact details
- Employment and income information
- Assets and liabilities
- Superannuation details
- Insurance coverage
- Investment holdings
- Cash flow and expenses
- Estate planning details (wills, POAs, beneficiaries)
- Life goals and objectives
- Risk profile
You can also paste in text from any source - meeting notes, a scanned document, or any other input - and the AI will extract relevant information from that text.
Manual override
AI filling is always a starting point - you have full control. You can edit any AI-filled value, add information the AI missed, or clear fields that are incorrect. The fact-find saves a version history so you can see what changed and when.
How the knowledge graph supports this
In the background, Prospo builds a knowledge graph for each client - a structured store of everything known about them across all meetings, documents, and communications.
- When the AI fills the fact-find, it does not just read the most recent transcript. It can draw on the full history of what the client has said across all their meetings.
- Client entities, contacts, assets, liabilities, goals, and advice events are all connected in the knowledge graph.
- This means the AI can answer questions like "What did this client say about their risk appetite across all meetings?" rather than only looking at the latest conversation.
The knowledge graph is powered by GraphDB (a graph database) and is modelled on a formal ontology covering all entities in the financial advice domain.
Semantic search across meetings
Meeting transcripts are chunked and converted into embeddings - numerical representations of meaning. These are stored in MongoDB Atlas and used for semantic search.
- Semantic search finds relevant content based on meaning, not just keywords.
- For example, searching for "retirement planning" will surface transcript excerpts where the concept was discussed, even if those exact words were not used.
- The AI Agent uses this same search capability when you ask it questions about a client's history.