Using multiple steps to solve a problem

Although programming solutions step by step can be very helpful, one of the best abilities that Semantic Kernel gives users is allowing them to make requests using natural language. This will require using planners, which we will use in Chapter 5, to break down a user request into multiple steps and then automatically call each step in the appropriate order.

In this section, we will solve problems by telling Semantic Kernel which functions to call. This is helpful for making sure that the solutions we make available to the planner work, and it is also helpful when we want to explicitly control how things are executed.

To illustrate the manual approach, we will see how to give Semantic Kernel clues about an animal, guess it with a semantic function, and then generate an image of the animal using the native function we created in the previous section.

Generating an image from a clue

In the following code, we have two steps. In the first...