AI in business decision-making processes

The present is AI

new solutions

Experience in business management using advanced methods

We have combined the results of our research on the application of artificial intelligence methods with our long-term experience in the theory and practice of business management. The result is new tools for management and decision-making that would be unfeasible without the use of cutting-edge AI.

Speeding up objective decision-making by managers

A decision model based on enterprise data allows managers to objectify their decision-making in the broad context of the enterprise organism expressed in quantified parameters.
Most decisions are generated and evaluated
in real time.

Managerial decision-making in the current conditions of the corporate organism

Managerial decision-making in an industrial enterprise at any level (strategic, tactical, or operational) is becoming progressively more complex as the environment in which corporate managers are forced to make these decisions becomes more complex.

The main problems of managerial decision making for enterprise development today include:

  1. Turbulent business environment - dynamic and unexpected changes in the internal and external business environment, which often take place in short intervals, make it necessary to make decisions immediately, often without sufficient, systematic and objective information, usually only intuitively. 
  2. Lack of information/information overload for decision-making - inadequate responses to provide quality information sources for decision-making, capturing the dynamics of developments, lead to blind decision-making or, on the contrary, to slowing down decision-making processes. 
  3. Complex data analysis - traditional data analysis to prepare the ground for objective management decisions requires knowledge, systematicity, time, capacity and multiple inputs. They often fail due to lack of data and therefore advanced analysis methods (using artificial intelligence) cannot be used.
  4. Complex decision methods - systematic methods and procedures that enable decision making in the broader context of the business organism with parameterisation of objectives, which entail continuous work of experienced teams, are often replaced by intuitive decision making at the operational level at the expense of quantitative methods. 
  5. Real-time decision making - requires rapid analysis of often incomplete data with a high risk of bad decisions, the ability to be flexible and change quickly, which is very difficult without the use of advanced artificial intelligence methods. 
  6. Parameters and decision limits - management decisions often do not have objectively set target parameters or limiting constraints for a particular entity because their determination and verification is usually difficult. This also applies to the determination of conditions and costs for the implementation of decisions.
  7. Decision impact monitoring and change management - unlike operational decisions, tactical and strategic decisions are difficult to monitor and evaluate. Due to the turbulent business situation, the need for further corrective interventions, changes and decisions by managers is essentially continuous.

The potential to overcome these challenges and obstacles is largely enabled by tools using artificial intelligence methods.

 

We bring you the following solutions:

A tool for generation decisions for business managers

ADAM aix

1. Assessing the starting position of your business, RATING

Without requiring any of your data, ADAM evaluates the position of your company based on economic data obtained mainly from your financial statements, published by law in the commercial register. Our system evaluates your current position in the following areas:

Finance

Calculated Rating. Overview of the economic activity for a minimum of 5 years at the level of the financial statements: result for the accounting period, Turnover. Tables, Charts.

Assets

Overview in the range of min 5 years: current assets structure, Assets x Liabilities. Liquidity analysis. Tables, Charts.

Profitability

Overview of the indicators in years: ROE Return on equity, ROA - Return on assets, turnover of liabilities and receivables. Tables, charts.

2. Determining the growth potentials of your business

The growth potentials are determined by comparing your data on the management and provision of business processes and the corresponding performance data of your company against the results of comparable companies in the so-called dataset of more than 700 companies, which is created over a long period of time by our cooperating organization VŠTE České Budějovice. The dataset is a completely unique dataset (probably unparalleled in research practice in the Czech Republic), which contains data on the way and conditions of how a selected part of enterprises (of different sizes and industries) provide/prefer their processes and their organization, how they make strategic decisions and in which areas they invest. There are currently 700 enterprises in the system for which (soft) data are available on their behaviour in the main areas of strategic management: Organisation, Strategy, Enterprise Environment, Resources and Processes - Investment, for a total of 320 soft data for each enterprise. In addition to this, economic data from financial statements were added, from a total of about 35 datasets for each company in time series of the last 10 years, i.e. about 350 additional economic data/entity. This created a set for the application of AI methods (Deep learning) with about half a million data, which is continuously updated every year and which ADAM aix uses for comparison.
The comparison procedure for assessing growth potentials is as follows:

Completing a checklist about your business

For your company, you will fill out a checklist (anonymous questionnaire) in our system to evaluate the state of management and assurance of your business processes.

Search for similar business

Based on the desired growth potential represented by the selected economic parameters, comparable entities to your business are selected from the dataset using AI. The searched entities are anonymized.

Determining the growth potentials of your business

The growth potentials of your business are determined using AI methods based on the growth potentials of similar businesses - the nearest neighbours - of the dataset's subjects, which have been searched for as a sub-category.

3. Creating Enterprise Digital Data Enterprise Data - ED

The digital enterprise database is a phenomenon using various labels, e.g. DATA LAKE, without which it would be impossible to deploy generative AI for strategic analysis in the enterprise. The ED Enterprise data system includes the organization of Data Lake type in the form of structured data - tabular, numerical data from management systems e.g. ERP, CRM, HR, from production process management systems, which are contained in relational SQL databases (Microsoft, Oracle) and also unstructured data, which are mainly various text and other files in original, native formats, which are produced across enterprise systems (processes) and form a unique and extensive knowledge base (about) the enterprise. In designing an ED organization, it is used with an object-organized repository that stores data according to its metadata with a unique identifier, which facilitates retrieval across the repository and improves performance. Unlike generic Data Lakes, the enterprise ED organization is tailored to the enterprise's information and knowledge acquisition needs. This ED organization defines Enterprise data implementation practices in a specific enterprise.

The implementation of Enterprise data completes the process of digitizing enterprise information. In the ADAM aix system, it proceeds as follows:

Integration with DMS

An enterprise digital data system for advanced search generation and for the application of AI can be created on top of existing DMS - Document Management System systems in your enterprise in the form of data interfaces.

ED Organization

The data handling system in the Enterprise Data ED (Digital Enterprise Information System) is ready for extended searches in structured (data) and unstructured (documents) information. It contains catalogues and metadata about the ED organisation.

RAG search methods

ED significantly improves the effectiveness of RAG - Retrieval-Augmented Generation methods for deploying large-scale LLM and AI machine learning language models. Therefore, we develop custom tools for applying this method.

4. Decision-making model for business managers

The decision model of the ADAM aix system is based on the description of the corporate organism, which it defines using strategic decision frameworks - SDFs in the basic areas of the company: finance, business, processes, potentials. It evaluates the selected user query in terms of its impact on each part of the SDF. Solution design = management decision based on the analysis of the capabilities of the internal (Enterprise Data) and external business environment. Each managerial query is parameterized by growth opportunities (potentials) and constrained by limits (costs). The system also proposes a way to implement the management decision in the form of an Implementation Project.

1. Choice of questions

The manager enters a query into the system by selecting from options that include major improvement opportunities, management changes, process optimization and solutions to important situations in the management of the company.

2. Definition of parameters

The basic query is further refined in the SDF options by means of system communication with the user, defining numerical, usually economic parameters of the decision.

3. Design assessment

The system will suggest a path to achieve the query goals within the given constraints using activities at different levels of execution. The user has the possibility to make corrections by selecting and adjusting the effectiveness of the proposed activities.

4. Clarification of the decision

The system will propose a solution according to the specified query. Procedure points 3 and 4 can be repeated until a match is reached with the manager's original query idea.

5. Parameters, limits

The resulting solution design is defined for each activity required for implementation by the target parameters of the activities and the overall solution. At the same time, the requirements and costs (limits) for the implementation of the decision are calculated.

6. Projekt implementation 

Once the results of point 5 have been agreed, the system produces a Decision Implementation Project. Depending on the complexity and impact, the decision may be strategic, tactical or operational. The project includes monitoring, risks and changes.

Solution principles:

For a given managerial query, the system first uses LLM to analyze the current position of the affected object in the Enterprise data (ED) using attributes, determine its relationships to relevant SDF objects, and propose preferences and limits for quantifying the query. Subsequently, according to the quantification chosen by the user, the SDF system determines for the query object the expected new values of the object attributes, the links and the impacts/requirements on the relevant concerned objects in parameters and time horizon. The core methods used here will be LLM analysis (RAG) and machine learning methods over data organized on the principle of neural networks.

Large Language Models (LLM)

Large Language Models (LLMs) are a new class of Natural Language Processing (NLP) models that have significantly outperformed their predecessors in performance and capability on a number of tasks such as answering open-ended questions, summarizing content, executing almost arbitrary instructions, as well as generating content and code. LLMs are trained on massive datasets using machine learning algorithms to learn the patterns and structures of human language.

Retrieval Augmented Generation (RAG)

Extended RAG loading generation is an architectural approach that improves the efficiency of large language model (LLM) applications by leveraging custom data. This is done by retrieving data/documents relevant to a query or task and providing them as context for the LLM. RAG has demonstrated success in supporting systems that access specific knowledge.

DATA LAKEs

A data lake is an environment for storing and processing large amounts of raw data in any format, including structured, semi-structured, and unstructured data . Most data lakes use cloud-based object storage, which also serves to use RAG and LLM technologies.
VUSTE-APIS is creating a custom environment for RAG in enterprise decision making - Enterprise Data.

Compare DATASET

The dataset is created in cooperation with the VŠTE České Budějovice. It is a unique dataset that contains data on how and under what conditions a selected number of companies (of different sizes and industries) ensure their processes and their organisation, how they make strategic decisions and in which areas they invest. The dataset contains half a million data on 700 enterprises. It is used for ML enterprise decision models.


Cooperations

Vysoká škola technická a ekonomická v Českých Budějovicích
Institute of Corporate Strategy, Department of Management. 

Leave a message

Don't hesitate to contact us

We are ready to work with you

Phone: +420 220 188 401
Email: vuste-apis@vuste-apis.cz

  • Introduction
  • Contacts
  • VAMAXis
  • Careers

© Copyright 2025 VUSTE-APIS, s.r.o.