You should consider specifying the consequent items to include those products you want to get recommended, as shown below. You can generate up to 500 associations for each specific product.
If you have a large number of products, you should do it programmatically using BigML's API. For that, you will need to find a strategy (metric) to trigger or select certain rules. In any case, to build a "modern recommender" you might want to use a mix of classifiers and Association Discovery rules that consider context as well as other knowledge you might be learning from your customers' behavior.