Digitized AI Product Manager and Business Analyst Platform

The platform is designed to provide business users with the most accurate predictive analysis of sales, cost-effectiveness of product functionality, user behavior by gathering the original company’s uploaded data using artificial intelligence and machine learning algorithms.

System structure

Customer web version
Admin web panel

Services

Backend development
Front end development
Product design
Project management
QA

Industry

Business management

Timeline

1.5 year

Team composition:

Dmitry, Project Manager

Sasha, Software Engineer

Nikolay, Software Engineer

Dasha, UX/UI Designer

Yanina, QA

Alex, CEO

Task from the Client

We were approached by a medium-sized company that craved to grow, but could not figure out the suitable product line and struggled with finding a good product manager. During this hard process of growing, they understood that there are numerous of start-ups, small, and medium-sized businesses that can’t grow up because can’t afford to hire business analysts or find a convenient product for detecting bottlenecks, best and worst products and features, and CJM.

Therefore, our partner decided to develop such a tool to optimize their own expenses and product line, work out their customer journey map, improve the user experience of their products and help other companies in assessing of UX and product’s marketing effectiveness based on the company’s data and with the help of the artificial intelligence analyzation abilities.

The Block Scheme of Digitized BA

After an active discussion with the customer, we have formed a maximum ergonomic product operation scheme that allows to cover all functions without loss of quality and performance of the system.

Technology stack used:

Design

Dasha, UX/UI Designer

Colours

Typography

Dasha, UX/UI Designer

Challenges We Faced With

As soon as we had to include artificial intelligence in the forecasting, assessing the current state, and constructing analyses, we had to implement strong validation parameters, so the uploaded data will be interpreted in the best shape.

Companies’ data has to be in the right format and includes all required fields.
Once the AI model is trained, it has to be stored for further generating of the synthetic data and analytics.
The safeness of business user’s data has to be the priority for the high product reliability.

Parameters in the uploaded datasets may not be divided into B2B and B2C since there are businesses that do both.
Accuracy of the predictions is controversial, because the amount and content of the data being uploaded can be different from the data the model was trained.

Functions We Implemented

Virtual BA Assistant

The core feature of this project. Each product company-user has its own CRM in which they upload data about the clients and sales. The algorithm studies the data and on their basis creates a report that shows at which stage the different processes take place: where the customer leaves, where it is halted, at what stage it moves to the selling point of the product. Then, the system forecasts possible further conversions of current leads and provides recommendations for improvement and income growth. Each CRM uses machine learning and in-depth data analysis based on the company’s information.

Virtual PM Assistant

For the companies who want to create new products or integrate new features in the existing ones, our system can be a virtual project manager, that will analyze the ROI of the product, detects its possible bottlenecks, highlight the core pros and cons of this solution, and provide some advices for growth, based on the uploaded data about the project or about the company as a whole.

Start-Ups ROI Auto-Evaluation Tool

For start-up companies only, we have created a tool to evaluate the profitability of investments and develop different functionalities for the weak project. As with KPI assessment and user behavior analysis, the client should dump all available data into their unique CRM, and an AI-based analyzer will predict the success of a certain function or project as a whole, as well as provide advice on improving CJM and UX.

Data Preprocessing Microservice

To solve the problem of the possible client’s data variety, we developed a data preprocessing microservice. This microservice accepts raw data from clients, either directly or from another service, and transforms it to a suitable format for the model. It cleans the data and handles missing values as necessary.

Unique CRM for Each of the Client

Once the processed data is stored in a database, it has to be saved in security. We realized the system of auto CRM generation. Each client will have its own data storage, so the confidential information won't be leaked or be available to other clients and their CRMs.

Main Selling Points for Each Type of Business Users

For start-ups

Since this project is intended to be used by different sizes of the companies, our algorithm’s filters are built with each of them in mind. For start-ups, we created the deep ROI evaluation filters that will analyze the full company as soon as it's only on the starting point of its existence.

For small businesses

For small-sized companies, our CRM will find the bottlenecks and possible points for fast and effective growing, also providing the advises and instructions for improving.

For medium businesses

For medium-sized companies, our project will be useful for improving the existed functions and evaluating the perspectives of current growing processes.

Received Feedback from the Client

“Celadon has built a suitable infrastructure to get the data required for the project. Despite the language barrier, the team still communicates effectively with the client. Overall, their expertise and dedication to their work have been valuable to the client.”