One book, several rulebooks, one standard of evidence.
An external asset manager carries a book across custodians and borders, and answers to more than one supervisor for the same client. Luscent generates the evidence once, as the work happens, with the jurisdiction attached.
Luscent is a System of Intelligence for external asset managers: it turns the communications a firm already holds into a Health Score for every client relationship and the MiFID II suitability evidence behind every recommendation, across every custodian the book sits on.
The pilot, in four factsCheckable on this page
- Pilot90 days, free
- Seats3 to 5
- DataEU only
- MigrationNone
Instruction PROP-0113 · blocked
One instruction across both booking sleeves, judged once against both regimes.
Two rulebooks on one instruction, resolved by the stricter of the two.Synthetic client. Rebuilt in the product's own idiom, not a screenshot.
PROP-0113 spans both booking sleeves and cannot be split, so it is judged once, against both regimes rather than either one.
EU MiFID II passes. The Brazilian product class, CVM-30-A3, does not. Two supervisors, two answers, the same instruction.
Most restrictive wins: blocked, with the path to comply attached. A rule decided it, so the reason is on the record rather than in someone's head.
Whose morning
do you want to see?
“… the proceeds, roughly 9.200.000 EUR, will land next month.”
Products matched · held for the suitability checkOpen the briefNew money noticed at the next review, if someone rereads the thread.
What changedThe sentence becomes a Signal, products are matched to what the Client said, and nothing is proposed before the suitability check.
Join the pilot programmeWhat this book looks like from here.
Three patterns that hold across firms like yours, and one number from outside that frames them.
of wealth and asset managers have scaled generative AI to more than one use case; 27 percent report substantial business impact.
EY GenAI in Wealth and Asset Management Survey 2025 (100 firms)
The book sits on platforms you do not control.
Positions live in someone else's format on someone else's schedule. What you control is the relationship, and nobody reads it back.
Two supervisors, one client, two standards of proof.
Keeping one set of evidence and hoping it satisfies both frameworks is the common approach and the fragile one.
Small teams carry large books.
More clients per person than a bank, and no supervision layer to catch what gets missed.
The figure is somebody else's number about the market, with its source. It is not a Luscent result.
Who asks what.
The seats that have to be satisfied, and the concern each one carries. Every question below is the first thing that seat asks, with the answer that holds today.
How long before this is worth anything?
The pilot runs ninety days on synthetic data first, with success metrics agreed before day one. A relationship representation accumulates, so month six is worth more than month one and we say so rather than promising a fast win.
What happens when a client moves custodian?
The understanding survives the move. A relationship representation is built from your own communications rather than from the platform holding the assets, so the positions change and what you know about the client does not.
What is the integration burden?
Microsoft 365 and Salesforce, through their own consent flows. Nothing is migrated.
Which regime does the evidence answer to?
MiFID II, out of a deterministic rules core. FinSA and FinIA are on the roadmap, not in the product today, so a Swiss-booked mandate produces no Swiss evidence yet. A compliance outcome comes from the rules engine and never from a model.
What do I have to do differently?
Read a list that is already in priority order. The evidence is produced from the work, not in addition to it.
What it looks like on the day.
Three pieces of the product, in the order a relationship manager meets them: a Signal, the Health Score it moves, and the evidence a rule writes. Every name and figure is invented.
Mandate scope questioned in two separate custodian threads.
Microsoft 365 · 5 messages · 9 daysA custodian statement query is still open.
Microsoft 365 · 1 message · 11 daysSignal
One prioritised feed across the whole book, which for a small team carrying a large one is the product.
Berger & Cie · Health Score
Two unanswered questions about fees carried over from May.
Health Score
A Health Score built from the relationship, so it does not break when a client changes custodian.
Evidence
The rule that fired is recorded with the result, so a cross-border file shows which framework produced which conclusion.
Every trade checked before the advice leaves.
Before a relationship manager sends a rebalance, Luscent checks each trade against the mandate and the strictest jurisdiction, and names the smallest fix.
Read the use case →Use case 02 · Comms to ComplianceAn email changes the client. The advice waits.
When a client email says their circumstances changed, Luscent reads the change, holds the advice that relied on the old picture, and opens a reassessment.
Read the use case →Use case 03 · Cross-BorderTwo rulebooks. One household. One defensible answer.
A household booked in Luxembourg and São Paulo sits under MiFID II and CVM at once. Luscent evaluates one action under both and shows a path to proceed.
Read the use case →Use case 04 · Relationship DriftRelationships rarely end loudly. They drift.
When a client relationship slips, Luscent shows it in the Health Score, links the messages behind it and lists the next conversation for the relationship manager.
Read the use case →Use case 05 · Liquidity EventNew money is announced in passing. Plan it before it lands.
When a client mentions a sale, an inheritance or a payout, Luscent raises it as a Signal with its source, matches products to what they said and holds them until suitability is checked.
Read the use case →Use case 06 · SupervisionSupervise the files, not the people.
One view of every Client file and firm obligation for the compliance function: what is open, overdue or overridden, with the record behind each item. No Relationship Manager score is displayed.
Read the use case →Where this sits against the rules.
Each row names its article, its method and what it puts on the record. A row that is not shipped says so, and is drawn that way.
This describes what Luscent does, not what your firm is obliged to do. It is not legal or compliance advice. The platform generates evidence and surfaces Signals; decisions remain with the firm and its professionals at all times.
Whether this is a fit.
Said plainly, both ways, so nobody spends ninety days finding out what one page could have told them.
Built for
- Managers running Microsoft 365 in the Swiss and Luxembourg corridor
- Books held across more than one custodian
- Firms answering to more than one supervisor for the same client
Not for
- Firms looking for custodian aggregation or performance reporting
- Firms wanting the system to place or approve a trade
- Firms whose client contact happens by phone and stays there
The same for every firm.
Five things that do not change with the kind of firm you are. Each one opens in place.
06.1Where Luscent sits in your operating model.
An external asset manager runs its Clients across front, middle and back office on systems it already owns, and Luscent replaces none of them. The numbers mark the three places relationships are most often lost. Select a tile.
Luscent today
Health Score for every Client, 0 to 100
Every Client carries a Health Score from 0 to 100, updated from the conversations Luscent reads. A falling score is the early sign that a relationship is cooling, before it shows in the assets.
See it on the day, section 03 →- 1A relationship cools between reviews, and the first sign anyone sees is the transfer request.
- 2Advice is given on a call and written up later, from memory.
- 3Supervision reads a sample of files, weeks after the conversation.
Luscent reads only the systems it is connected to, listed under What is connected below. In build means designed and in development; no date is promised on this page.
06.2Two engines, and a rules core that never guesses.
The three things you saw in section 03, and where each of them comes from.
A Signal carries a confidence because a model produced it. A compliance outcome carries the rule that fired and no confidence at all, because a regulatory result is not a guess.
Inference is EU-hosted, by EU providers only, with no US model API in the client-data path. Tabular models are on the roadmap and are drawn that way.
06.3What is connected, and what is not.
2 of 4 live
Microsoft 365 and Salesforce are live today. The two drawn dashed are on the roadmap and not shipped. We would rather tell you that than let you assume otherwise. Luscent reads them and replaces none of them.
06.4Where this gets it wrong.
Every firm evaluating this has been shown a confident wrong answer by something before. So the limits are on the page rather than in the room, and each one names the control that catches it.
Confidence is a number, not a feeling.
Every probabilistic Signal carries a numeric confidence on its face. Deterministic outcomes carry none, because they have none to state: the same input produces the same result every time.
It reads what it is connected to, and nothing else.
A relationship conducted by telephone, in person, or on a channel Luscent does not read is invisible to it. The systems drawing above is the whole of what it can see.
It surfaces, it does not decide.
No Signal approves advice and no score authorises an action. Every consequential decision is taken by a named person, and the record says who.
It will be wrong, and the pilot counts it.
False positives are reviewed weekly with your compliance function during the pilot, and the rate is one of the success metrics agreed before day one rather than something discovered at the end.
A thin month is a thin month.
The representation of a relationship accumulates from your own communications. Month six is worth considerably more than month one, and no amount of capital shortens that for anyone.
06.5What your outsourcing officer needs.
Whether a deployment is a critical or important function is your firm's classification to make, and your outsourcing officer owns it. What we can do is put the facts that classification turns on in one place.
Luscent reads it and does not write to it.
Every consequential decision is taken by a named person.
Stored on Scaleway, Paris; services on Hetzner, Germany.
Fact means true today and checkable, here or in the Trust Center. Position means the stance Luscent takes into a negotiation; it binds nobody until it is in a signed agreement.