Enabling AI Services Through Operationalisation & Self-Service Analytics - A White Paper by Dataiku

You are the CXO of a company that serves as a sales platform for thousands of different clients. You and your management team have identified a list of processes that could be improved via better use of your data and advanced analytics. For instance, in order to help your clients increase sales (and thus increase the stickiness of the platform), you decide that the development team should surface a custom recommendation module offering three product recommendations per client.
Requirements include that:

• The recommendations be available in real time and
•  The recommendation engine can be updated without causing platform downtime

At the same time, there are internal requests from different teams (like sales and marketing) who want to make data-driven decisions (for example, key trending products with positive reviews from social networks or how transformation rates are influenced by the historical browsing behavior of a visitor), but their old dashboards are static and don’t address their needs. Even though the data exists internally, they can’t get insights for themselves because they don’t have direct, regular, monitored access to data that can help them do their jobs.

Which need should be prioritized? And how do you even begin to tackle these projects with an approach that will be sustainable and reproducible for other projects and requests down the road?