Alexa, Again, a Shopping List Is Not a Shopping Cart

On February 26, the Federal Circuit issued its opinion in the battle between Freshub, Inc. and Amazon.com, Inc. regarding the Alexa device. The case came down to a distinction between an “item” and a “word,” which was hotly contested at the oral argument we reported on back in January.

Freshub contended that Amazon’s shopping-list feature of Alexa-supported devices infringed patent claims directed to a voice processing system that “identif[ies] an item corresponding to the text” and “add[s] the identified item to a list.” The jury concluded that these limitations were not met, and the Federal Circuit found that substantial evidence supported the jury’s verdict. 

In hindsight, Freshub’s biggest misstep might have been its failure to seek claim construction of the limitations in dispute. That left the question of infringement squarely in the hands of the experts. Amazon’s expert explained that the shopping list feature is programmed to add words to a list, whether or not those words correspond to purchasable “items.” This was demonstrated by evidence that words such as “sad” and “unicorns in a can” could be added to the shopping list, when, clearly, those words do not correspond to purchasable items. What’s more, Amazon contrasted the accused shopping-list feature with Alexa’s non-accused shopping-cart feature, which does undertake the additional step of identifying an “item” that corresponds to the text. 

But perhaps Freshub’s claim was doomed regardless of whether it sought claim construction. The Federal Circuit made clear that the jury’s verdict was consistent with how the claim term “item” was used in the intrinsic evidence. The claim language itself alludes to the phrase “identify an item” to mean a “specific, purchasable item.” In any event, without any claim construction, the jury was free to use the evidence provided at trial to conclude that the shopping-list feature did not infringe Freshub’s patent.

[T]here is no claim construction of “item.” Nor is there a claim construction indicating that a system comes within the claim as long as the actions taken upon execution of the system instructions even sometimes amount to performance of the claim-required operations.

files.passle.net/...