Morningstar: portfolio accounts checked against the target allocation, proposals with reasoning.
At a glanceStarting point
Everyone talks about AI. Hardly anyone sees what a use case looks like in everyday work.
This is one way to summarise an observation that Morningstar makes in many client conversations. The clients, mostly financial institutions on the investor side, are often uncertain about what AI can really do for them.
Morningstar supplies the data that such solutions need, but deliberately does not build software or finished AI applications itself. Precisely with clients who want to build their own, independently working AI on top of this, the questions are considerable: who has to be involved, and how do things stand with compliance and IT security? These are projects that many firms cannot manage on their own internally.
On top of this comes the everyday work of the wealth managers: the care of existing clients regularly falls behind in day-to-day business, because acquiring new clients takes priority. At the same time, a huge transfer of wealth is imminent: poorly looked-after existing clients switch provider at exactly that point. Added to this are laborious mandatory documentation and a data situation scattered across many providers: any number of manual tasks that cost time. Time that is then missing for the actual advice.
Solution
A three-way collaboration: data, industry knowledge and the AI employee.
Out of this situation, a three-way collaboration emerged: scoreprise.AI builds the AI employees, EY brings its experience from the finance industry, and Morningstar supplies the data. For Morningstar's AI bootcamp in Frankfurt, scoreprise.AI and EY developed a working prototype in just one week: the AI employee for wealth management (named Lucy in the project), entirely based on the Morningstar data.
She works through a clear interface in the Morningstar look and through the GDPR-compliant platform Langdock, not a black box: identified deviations in the portfolio accounts, upcoming appointments with prepared talking points, a market overview, mandatory documents to be approved, and a complete record of all steps.
A central element is the reallocation proposal: the AI employee reads a client's positions live from the Morningstar data, identifies a deviation from the target allocation and proposes concrete purchases and sales, with reasoning, a new allocation and a forecast over twelve months. All recommendations deliberately remain proposals: she prepares, the wealth manager decides. For asset management, the AI employee for asset management applies the same way of working to competitive and comparative analyses. The models run on dedicated servers in Germany, and Langdock too works on infrastructure located in Germany. Sensitive client data does not leave the secured environment.
Overnight run
Portfolio accounts checked against the target allocation
Every night, the AI employee for wealth management reads the positions of the client book from the Morningstar data and identifies deviations from the target allocation.
Proposal
Purchases and sales with reasoning
For each deviation, a proposal is created with a new allocation, reasoning and a forecast over twelve months. A proposal is made, the decision is taken by the adviser.
Documents
Mandatory documents in draft
The documents required by regulation, morning briefings and email drafts are ready for approval, and every step remains recorded.
The aha moment
„Eine reine Grenz- und Compliance-Prüfung kann heute jede Software. Echte KI muss mehr leisten: eine eigene, begründete Einschätzung, an der man merkt, dass wirklich Intelligenz dahintersteckt.“
After one week there was no concept on slides, but a tangible digital colleague working live with the Morningstar data. The reasoned reallocation proposal became the most convincing part of the use case.
Result
From the bootcamp into the client conversations in the DACH region.
The AI employee was presented at Morningstar's AI bootcamp in Frankfurt, and the prototype has continued to have an effect since then. Morningstar's sales and advisory teams in the DACH region and Switzerland use the functioning demo in client conversations to make tangible what is possible with the data.
In the wider group, the ideal target audience became apparent: small to medium-sized wealth managers and private banks, above all in Switzerland, who cannot build such a solution on their own and for exactly that reason need a partner. Concrete improvements flowed directly from the feedback: a more detailed statement of evidence and sources for every data query, as well as a separate workspace per client. The form of the collaboration also took shape: a shared path in which both sides win clients, rather than a pure data feed.
The use case shows how data providers, advisers and AI specialists together turn existing data into an AI employee that can be put to use, and thereby close the gap between AI promises and concrete benefit for financial institutions.
„Was scoreprise.AI und EY hier in kürzester Zeit gebaut haben, ist genau das, was viele unserer Kunden suchen: ein konkretes, greifbares Beispiel statt abstrakter KI-Versprechen. Ein fertig vorzeigbarer digitaler Kollege auf unseren Daten hilft gerade Häusern, die beim Thema KI noch unsicher sind, den nächsten Schritt zu gehen.“
Next step
Your use case, shown on a comparable one.
Cancellable monthly · Operated in Germany