We recommend getting started by ordering a kit from Fluent.Pet! We've developed specialized buttons and matching tiles to make setting everything up as easy as possible, and to maximize the probability of success with your learner.
(Please note, however, that you do NOT need to use FluentPet buttons to participate in the research) We have found that using food rewards to teach buttons can be counter-productive. Early results seem to suggest that even though a "treat" button is easy to teach, it ends up slowing down learning of other buttons. Good first non-treat words are "outside", "play", or potentially "water" or a favorite toy. What makes all of this possible is that dogs pay us very close attention, can recognize the words we use, can press buttons, and (it seems) can recognize the sounds that the buttons produce.
What you're trying to do is create a strong association in your dog's mind between two things:
particular activities or things that matter to your learner and
particular buttons that, when pressed, make the sound of a word that corresponds to those particular activities or things
When you're modeling a word, you want to use it repeatedly, in context, without adding different conjugations.
So if you're trying to model PLAY, then when you're about to play with your learner, say "Do you want to play? Okay let's play! Play play play!"
You want the phrases you use to be simple, clear, and easy to understand. Avoid speaking quickly. You will want to repeat your learner's words frequently immediately before and during the action that they correspond to.
2. Pay close attention to the words you're using
The best words to start with are those that refer to things your learner cares about.
Watch the way you speak about things, and notice if there are particular ways that you are already referring to objects or activities that matter to them. Do you say "treat" or "cookie"? Do you say "eat" or "hungry"? Do you say "play" or "have fun"?
Figure out which it is that you use more, and that your dog seems to recognize, and use that one exclusively. This will be the word you end up recording into one of your buttons.
3. Model: use their buttons!
Dogs can learn by watching, also known as mimicry. When you are working to model a new word or concept for your learner, say the word, then press the button it corresponds to.
This routine of saying a word or phrase, pressing the corresponding button or buttons, and then either directing your learner's attention to an object ("ball!") or engaging in the action you've just described, is called modeling.