It explains investment concepts and your own products from the documents you approve, gives a verified client a read-only view of their own portfolio, books time with their adviser, and brings a real person in when it should — on your website or client portal.
14-day free trial. No card required. We build your assistant for you.
What it does
It looks up the answer in your approved knowledge, shows where it came from, and knows exactly where its job ends and your adviser's begins.
Send us your factsheets, fee schedules and product guides and we turn them into its knowledge base. Product facts are quoted with the document name, version and date — never from the model's memory.
General education — what an ETF is, how a risk rating works, what a fee schedule says — answered clearly, without turning into a recommendation.
After the client verifies their identity, it gives a read-only summary of their own holdings, allocation and performance. Every figure is calculated by code and carries an as-of date.
Requests a meeting or callback with the client's own adviser and reads the details back to confirm. It never claims something is booked until the system confirms it.
Complaints, suspected scams and clients in distress go straight to your team as a confidential case with a warm, fixed reply — not a product pitch.
Questions answered from documents versus escalated, topics clients ask about, and the knowledge gaps to fix — aggregated, with small groups hidden and confidential conversations excluded.
How it works
Assembling a knowledge base, writing agent instructions and wiring up integrations is our job, not yours. Our team does all of it inside your account — every action logged and visible to you — and hands it over for you to test and approve.
Factsheets, fee schedules, product guides, disclaimer wording and your adviser list. Client and holdings data comes through a secure import — never pasted into a form.
Our team structures the knowledge base, writes the assistant's instructions and connects your tools. You review it in a test conversation before anything goes live.
Nothing reaches your clients until you sign off. We add it to your website or client portal and it starts working.
Trust
In investing, an assistant that improvises a return or nudges a client toward a product is a liability. These limits are enforced in code, not left to a prompt.
It will not recommend what to buy, sell or hold, predict returns, or place a trade. Those requests get a fixed, honest reply and an offer to reach your adviser. There is no trade or transfer capability to misuse.
It is built to decline rather than invent a price, a return, a holding or a fee it can't find in your approved documents. That's a design constraint, not a setting.
Access to account information is decided in code from a verified identity and the adviser assignment — never from what someone claims. Every firm's records are isolated at the database level.
Your required disclosure wording is appended by code to every reply that contains figures or product facts. It is not left to the model.
Your agent runs on multiple frontier AI models with automatic failover, so a single provider having a bad day doesn't take your assistant offline. Which model answered is never exposed to your customers.
Pricing
Setup and configuration by our team is part of the service. Plans are being finalised, so we don't show numbers we may change — your 14-day free trial starts when your assistant goes live, and we confirm plan and price with you before the trial ends.
Send us your documents and we'll build your assistant during your 14-day free trial. You review it, you approve it, and only then does it talk to a client.