Umbrella IT’s system analysts were the first to join the project. They conducted competitive intelligence of existing virtual assistants: Google Assistant (Google), LubeChat (Shell), Alisa (Yandex), Marusya (Mail.ru). The best practices were taken into account when building the Voice Assistant interface. Then the team of analysts, together with the UI/UX designers, conducted UX testing of the application mockups on the company’s employees to cover the core use cases. After the use cases validation, the development team started the MVP development. The team built 34 requests which the employees could make: for example, upon the voice command, the Assistant could display the information on problems at the well and send a detailed report by e-mail. The developers included the most frequently used terms from the oil and gas industry in the Assistant’s vocabulary frequently used terms from the oil and gas industry. Whenever it was necessary to add a new request, the project manager and stakeholders determined its priority, and the new request either replaced one of the original or was added in the backlog for the next app version. To ensure that the employees could use the Assistant offline, the UIT team implemented the speech recognition and generation feature within a closed local network that streams telemetry without the third-party services.