Ingesting signals from chat/voice assistants in AAM

Avatar

Avatar

jigneshb7072369

Avatar

jigneshb7072369

jigneshb7072369

16-07-2018

With the increased usage and adoption of chat and voice assistants (e.g. Amazon Alexa, Google Home, Microsoft Cortana, Apple Siri) by consumers, it has become imperative for marketers to catch consumer intent and insights from these new age channels and use it for personalized and relevant targeting. For example, a consumer can interact with Alexa by stating in simple English or any other regional language

   "I am looking for buying a high end camera. Can you suggest some good options?"

If we try to break this statement into intent and entities (basis of natural language processing) structure, we would know that this consumer's intent is to 'explore' or 'buy' and the entity here is a 'high end camera' or simply put 'SLR'. In similar way, the consumer might express something like this -

   "I am looking for buying a high end camera with Carl Zeiss Lens and price range somewhere between $300-$350. Can you suggest some good options?"

Here there are more entities that need to be captured in terms of lens brand and the pricing range. The intent and entities can be captured by building data models using bot and voice technologies such as Amazon's Lex, Microsoft's LUIS/Bot Framework or any AI/NLP technology.

Furthermore, when it comes to building consumer segments, the signals (intent and entities) coming from these channels have to make their way in a DMP platform (AAM) so that consumer traits can be built on top of them and finally the audience segments can be constructed. These segments can then be used to activate personalized media and content. Also, the device IDs for these devices (for example, Amazon's Echo (/Firestick) devices has its own advertising ID for unique identification, similary Google Home devices have GAIDs, ditto for Apple) have to be ingested into AAM to target consumers based on authenticated or unauthenticated states. These additional devices can also be made part of consumer's device graph in the system so that progressive messaging and targeting can be used for effective consumer engagement and conversion.

Comment