AI-based Investment Management Platform

Next-gen solution for informed investment decision-making.

AI-based investment management platform

Challenge

Build an AI-powered investment assistant for a banking and finance institution.

Solution

Development of a web platform that enables customized investment strategy building using predictive analytics and NLP.

Tech stack

.NET, Azure, AI, React

Client

A well-established Central-European private investment firm with a half-century-long operating history reached Modsen team in search of an innovative AI-powered software solution. The institution deals with servicing both physical and legal persons in terms of providing advice and practical guidance to investors, managing their portfolios, and building successful investment strategies.

AI-based platform interface

Challenge

The ever-expanding trend for digital transformation in finance has got our client thinking about ways of leveraging the latest and most powerful technologies to add their firm a solid competitive edge. The investment industry has been always sitting on terabytes of data that can and should be analyzed and used for the benefit of investors. By incorporating AI into a web platform, our partner hoped to:

  • Provide a larger number of investors with high-quality services;
  • Attract new segments of customers;
  • Streamline the efficiency and accuracy of the advisory and portfolio management services;
  • Reduce overheads.

The AI-powered digital investment tool was required to combine all the necessary features and capabilities to ensure high-accuracy analysis and forecasting of trends and low-risk investments. The implementation of such cutting-edge software requires the completion of the following tasks:

  • Data collection;
  • Forecasting model training;
  • Cognitive mining;
  • Incorporation of admin and client interfaces.

Modsen team has been improving our AI-taming abilities since the inception of the technology mass use and over the years, we’ve managed to reach a level of creative proficiency that allows us to provide our partners with more sophisticated and efficient digital solutions like the one we are describing in this case study.

Team

1

Project manager

1

Business Analyst

1

Software architect

1

Team lead

5

Back-end developers

3

React developers

3

QA testers

2

UX/UI designers

Modsen AI developer

A brief overview of the AI-for-investment solution

The increasing attention to AI as an unparalleled investment tool is gaining momentum, however, despite the obvious unlimited potential of the technology in the financial industry, there are surprisingly few entities leveraging the power of AI which makes those who are now incorporating the technology far ahead of the curve. The essence of AI-based investment lies in using predictive analytics and NLP, the combination of which ensures in-depth analysis of historical data to identify trends and processing of the current multi-source data to build accurate predictions on safe investment moves. Apart from comprehensive data analysis, AI algorithms are deprived of the loss aversion factor that often prevents people from making beneficial portfolio additions.

Development process

Requirements gathering and processing

In the development of successful efficient software products, it's all about the details which are collected and carefully documented during online client consultations. The complexity of this particular project required conducting 2 in-depth meetings with the client and their in-house investment specialists to bring in valuable insights and point out key business goals and challenges the software was projected to solve.

Discovery Phase

Apart from receiving the basic general data about features and capacities of the future AI platform during the requirement gathering phase, our team also delved deep into the aspect of internal business processes to identify obstacles and growth points, set correct project KPIs, and draft up a structured and detailed project development strategy. We started off by analyzing the data available on the web and then got in touch with the respective company employees who could expand our understanding of certain processes the software would have to automate and enhance. The quality of the discovery phase affects the whole upcoming development process by setting the right trajectory for the work of our expert team. To conclude this step and make sure we were on the right track, our PM with the help of BA, and other project team specialists prepared a presentation to be submitted to the client. If any misunderstandings or corrections should have appeared, they could be quickly and easily discussed and solved.

Planning 

Before getting down to code cracking, we drafted up a well-thought-through project development plan outlining the deadline dates, sprints duration, and set aside some time for unexpected challenges. At this point, when the essence of the project, its scale, complexity, tech stack, and team size was settled, we clarified the final cost of the project.

Team assembly

The development of an AI-driven investment management platform required the input of a 15-specialist team, namely, 1 PM, 1 BA, 1 software architect, 1 team lead, 5 back-end developers, 2 UI/UX designers, 3 react developers, 3 QA testers. The employed tech stack included Azure AI tooling as the backbone of highly performative and responsive AI-powered software which implied the work could be managed only by senior-level engineers with a solid background experience in using the technology. To ensure the task-appropriate composition of the team, Modsen CTO conducted the process of additional competence check for the project team candidates and provided the client with a team of unique dedicated experts destined to make the project shine.

Design

The work of the 2 UI/UX Modsen designers with previous investment app interface-building experience was divided into several stages:
  • UX research
  • UX strategy building
  • Wireframing
  • Visual design
  • Testing
It’s worth noting that our partner had a well-established visual brand that had to fall in line with the software design under development. Modsen Design Studio specialists did a wonderful job creating a solid business-looking front-end that didn’t overcomplicate the software functionality for the end user and allowed them to quickly grasp the navigation logic.

Coding

The development of AI-powered software was split into two separate processes which were merged together when the AI-part was ready for deployment. Here is how Modsen team approached the task.

Software development

Among the variety of technologies, our engineers’ team considered .NET Azure, React, and React Native as the most efficient and appropriate ones for the development of a high-quality robust AI-driven investment management web platform.

AI-module development

To create the AI part of the software, our team relied on Azure AI, capable of equipping engineers with all the necessary tools and frameworks for quick and hustle-free building, training, and deployment of AI models as well as cognitive search services.

The steps taken by Modsen team to develop a performative AI app module included:

Data exploration

Data exploration

Data is the heart of any AI algorithm, so its careful preparation, validation, and analysis are crucial for the success of the whole project.

AI model development and training

AI model development and training

The high-quality data generated during the previous phase serves as the basis for training algorithms to access financial assets and forecast their performance short and long-term performance.

Model testing

Model testing

At this point, all the developed models are evaluated on the basis of accuracy, robustness, and scenario-handling performance, and the best ones are approved for live usage within the application.

Comprehensive testing

When it comes to AI-powered software, quality assurance testing becomes even more significant. To make sure our client gets a sleek, robust, accurately-working investment platform, Modsen testers completed a series of manual and automated test scenarios to prepare a detailed report highlighting the identified inconsistencies waiting to be corrected by the engineers’ team.

Integration

To finish off this challenging yet highly satisfying project development process, we conducted a smooth transfer of the code, documentation, and usability guides to our partner, ensuring trouble-free integration of the new software into the company’s internal routine.

Servicing

Our team regularly mentions the necessity of high-quality post-launch software servicing and the perceiving of the app development process as an ongoing one. In the case of the AI-powered investment management platform, the importance of support and servicing became particularly crucial as the AI component has to be monitored and retrained on a regular basis to preserve forecasting accuracy.

Result

The web platform was deployed in May 2022, and now, after a year of successful operating in the Central-European financial market, the client was ready to share the ultimately positive outcomes of the AI-powered digital investment management system incorporation.

34%

client coverage expansion

59%

increase in the accuracy of the investment decision-making

65%

yearly income growth compared to the pre-app-launch period

Let’s calculate the accurate cost and resources required for your project