Catchlight

A predictive conversion model trained on more than 170,000 client conversions carries the proposition, telling an adviser which prospect to call next. Without it an enriched but unordered list remains, which is not what anyone buys this for.

Each lead carries a confidence mark showing where attributes are missing, so a user sees which scores rest on a thin profile before acting. The mark reflects profile completeness rather than the model's confidence in its prediction, and no approval step or audit record follows a low score.

Over 100,000 adviser and client interactions describe what the model learned from, but no accuracy rate, calibration statement or holdout result says how well it predicts. Describing the data as institutional grade puts an adjective where a figure belongs.

An executive at a named advisory firm is quoted by full name and title, and trade reporting put adoption near 300 firms in late 2024. The two to three times conversion lift, the figure this product lives on, is self reported, tied to no firm and measured against nothing stated.

Scoring is said to get smarter with adviser use, which in one shared model means a firm's conversion results improve predictions sold to its competitors. Whether uploaded leads and outcomes train that model, and on what terms, is never stated.

Profiled prospects have no relationship with the searching firm and never supplied the two thousand attributes assembled about them, and nothing indicates they learn a wealth estimate exists in their name. Data a person gave the parent group for marketing is reused to estimate their value to someone else, and no consumer notice, deletion route or state privacy position appears.

Product pages, support material and third party listings show no named attestation, audit report or trust page. For a firm uploading prospect lists, that leaves the vendor's controls to be established in diligence.

A marketing technology subsidiary inside a regulated asset management group needs no license, and the group's authorizations do not extend to it. Its disclaimer is unusually precise, disavowing any endorsement of advisers or recommendation to prospective clients, a line drawn because ranking consumers by value to an adviser sits close to steering.

A model trained on who became wealth management clients learns advisers' historical choices, shaped by inherited wealth, geography, race and gender, so people resembling existing clients rank higher and the rest stop getting calls. That loop feeds the next model across every subscribing firm at once, and no fairness evaluation or score distribution test is published.

The vendor disclaims responsibility for the accuracy of its information and for outcomes of interactions relying on it, allocating all risk to the buying firm. The profiled individual has no published way to learn a score exists, see its attributes or correct an error, and meets the consequence only as an adviser who never called.

Calling the models proprietary discloses that no external model provider does the scoring. The data brokers feeding a wealth estimate about named individuals matter as much here, and no source, aggregator or record type is named.

A customer relationship platform application, a programmatic interface and bulk loading of up to 2,500 leads cover the routes into a firm's estate, and a large custodian's marketplace adds distribution. No connection is named to the client relationship systems independent firms typically run.

The hosted application publishes nothing on hosting provider, region or residency. A single domestic region is the expected answer, but it is not a published one.

Support material mentions pricing plans in the plural, yet no rate, tier or unit appears. Since the product enriches leads in batches, whether a firm pays per seat or per lead enriched changes the economics for high volume acquisition.

Independent advice firms in the United States are the only buyers, and only for acquisition, with nothing in planning, servicing or compliance. A separate program for small firms spreads it across the size range within that one buyer type.