A lot of different resources are needed for AI, including hardware, software, data, and people.
Hardware is needed to run the algorithms that make up AI applications. This can be anything from a simple desktop computer to a powerful server farm. The more powerful the hardware, the faster the computations can be performed.
Software is needed to implement the algorithms on the hardware. This can be off-the-shelf software such as TensorFlow or Microsoft Cognitive Toolkit, or it can be custom software written by AI experts.
Data is needed to train and test the AI application. This data can come from internal sources such as transaction records or clickstream data, or it can come from external sources such as weather data or social media data. The more data that is available, the better able the AI application will be to learn and improve over time.
People are needed to design and build AI applications. These people need both technical skills (in areas such as mathematics, statistics, computer science, and engineering) and domain expertise (in areas such as finance, healthcare, retail).