Building an translation layer connecting which can connect outside API’s to our datawarehouse in combination with a treatment layer which can pretreat the incoming data.
Using machine learning which can find common identifiers on each platform in order to match user data strings to each other. Also developing an IP-based layer of deduplication with san associated probabilistic model that takes into account user behaviour.
Developing a production machine learning pipeline with data preparation recipes, data transformation algorithms and applied relevant machine learning models.
Developing an automated ad creation and visualization engine based on client visual, textual input and a channel management layer as an API connection that takes the treated and validated input of the client and exports it into the relevant marketing platform.
Developing an execution layer that takes the data insights exported from the machine learning pipeline, treats it and connects it to the API of relevant marketing channels.
This execution layer should be able to recognize which types of insights produce specific actions such as narrowing a campaign targeting, turning on and off specific ads, and increasing or reducing budget, regardless of client, platform or sector. Also developing a data feedback layer that optimizes the machine learnings models with the data of the campaigns changed by this execution layer.