We support the use of the latest AI technology to promote reforms in our customers' business models.
At UNAIIT, we develop custom-made solutions to solve your business problems from scratch. ERPAI solution that utilizes business / core data, accounting data, and financial data. "Image recognition" and "abnormality detection" that utilize image data and voice data. Fits customer issues such as "structuring and visualization of unstructured data" using various natural language data, "demand forecasting" and "risk forecasting" using sleeping data, marketing data, etc. We will help you accelerate your business by building a highly accurate AI algorithm from scratch.
Flow of AI development
Understanding the current situation, defining requirements, setting goals
First of all, we will interview the customer and understand the current problem that the customer is having. Based on a common understanding of the gap between the ideal and the current situation, we will sort out issues (things to be solved) regarding the tasks and priorities required for that. After organizing, set goals to be solved by tools such as AI / IT / optimization algorithms.
Data collection, confirmation and data cleansing
If you do not have the data you need to solve your problem, we will help you collect data, support you, and build big data. In addition, since the format and meaning of data differ depending on the customer, we perform exploratory data analysis to check the data and understand its contents. Then, in order to build a more accurate AI algorithm, cleanse the data (delete / correct / correct / complement data, etc.), analyze the characteristics and trends of the data, and consider the approach method. I will do it.
Scheduling, cost scaling
We will formulate a specific policy based on the trends and characteristics of the grasped data. We divide the phase of the project, propose the overall flow, rough schedule, and share the recognition with the customer about the work required for actual construction. At that time, we also perform cost analysis when building AI and perform cost scaling according to your budget.
AI / IT environment construction
After completing the assignment settings, we will build the environment for operating AI / IT / algorithms. Whether you use the public cloud or on-premises, we design and build a secure, cost-effective, highly available AI / IT environment. * Environmental costs will be discussed separately.
Development of prototype AI
We will develop prototype AI by utilizing the prior information obtained in the assignment setting phase and our know-how.
Demo UI creation, accuracy evaluation of prototype AI
We will scrutinize the accuracy and performance of the prototype AI, determine with the customer whether a solution that can solve the customer's problem is feasible, and consider whether to continue the project or collect more data.
We will use prototype AI to test whether AI is available before actual operation. The logs obtained at this time and the user's actions are forged back to AI, and tuning is performed to make it better AI.
Regardless of whether it is equipped with AI, you can design the system according to your existing system environment / requirements. For example, when embedding AI in a core system, design an API and link the core system with AI. In addition, it is also possible to design as a complete system as a WEB application / native application / mobile application / edge device application. We also design large-scale AI so that it can operate hardware resources efficiently.
Code the system based on the system design. Even during development, we will work closely with you to develop better systems in order to reflect the needs of our customers.
Assuming actual operation, check whether the system operates normally, including AI performance tests. Perform system tests to improve UI / UX and tune AI.
System migration / system deployment
After system development, deploy the AI / system to the actual production environment. It monitors the load during deployment and whether it is operating normally. It also monitors AI logs and creates reports on how AI is used by customers.
AI auto tuning
Creating AI and systems is not the end. It only makes sense if you operate and grow it. We will automatically extract the data accumulated after operation, automatically analyze the customer's data, and build a mechanism to be smarter by ourselves.
Performance / auto-tuning monitoring
It monitors how the AI / system is operating normally, whether auto-tuning is functioning normally, how accurate it is, and reports the information that customers want to know. We will share information on how much we are contributing to our customers' issues.
We also maintain the AI / system. We will sincerely respond to questions and inquiries about your system, including operation monitoring, and will tune and update your system according to your needs, including solving questions and solving problems.